Managing tech

transformations
Managing transformational and organizational change in a
rapidly changing technological landscape

May 2021
Contents
Introduction 3

1. Transformation that works

How to become ‘tech forward’: A technology-transformation approach 5

that works

Seven lessons on how technology transformations can deliver value 12

2. Reimagining the role of technology

The CIO challenge: Modern business needs a new kind of tech leader 22

Products and platforms: Is your technology operating model ready? 30

3. Reinventing technology delivery

How companies can win in the seven tech-talent battlegrounds 39

CIOs are redefining what a successful relationship with their IT providers 47

looks like

Unlocking value: Four lessons in cloud sourcing and consumption 52

Three actions CEOs can take to get value from cloud computing 58

Breaking through data-architecture gridlock to scale AI 63

4. Future-proofing the foundation

Managing large technology programs in the digital era 71

evin Goering
Reed Haley
Rob Huefner
Mark Huntington
Sarah Johnson
Mithun Kamat
Richard Kelly
 Introduction
With COVID-19 increasing the need for technology to be a competitive advantage, technology transformations
have become increasingly important. This sentiment was echoed in last year's annual McKinsey Global Survey
on technology strategy, where almost all respondents had conducted some type of technology transformation
in the past two years, with more planned on the horizon.

This emphasis on transformation seems to be paying off. In this year's technology strategy survey, most
companies report some or significant impact from their companies’ technology transformations. For all that
progress, however, huge value from technology is still untapped at most large incumbent companies. As
technology moves front and center in the business, successful transformations require changes across three
dimensions, or vectors, to become truly “tech forward”:

— Reimagining the role of IT

— Reinventing tech delivery

— Future-proofing the foundation

Our latest survey shows that there are no silver bullets. Tech leaders have to frame and orchestrate the full
transformation across all these vectors—in a constantly shifting and dynamic landscape.

To help bring clarity to this reality, this collection explores the structural shifts needed on each of these
vectors to drive sustainable change and business value.

On behalf of the tech transformation and management community at McKinsey:

Abhi Bhatnagar Leorizio D’Aversa Anusha Dhasarathy Rahil Jogani

Partner Senior Partner Partner Partner
Atlanta Milan Chicago New York

Naufal Khan Suman Thareja Steve Van Kuiken Hrishika Vuppala

Senior Partner Partner Senior Partner Partner
Chicago New Jersey New Jersey San Francisco

Managing tech transformations 3

1. Transformation
that works
 How to become ‘tech forward’:
A technology-transformation
approach that works
Getting value from tech relies on tackling multiple interdependent programs at once.

This article was a collaborative effort by Anusha Dhasarathy, Isha Gill, Naufal Khan, Sriram Sekar, and Steve Van Kuiken,
representing views from McKinsey Technology.

© Getty Images

November 2020

5
 For executives looking for lessons in the wake of important pitfalls are—we spoke with nearly 700
COVID-19, one has emerged clearly: every company CIOs at some of the largest companies across the
needs to become a tech company. world. These conversations illuminated a number
of consistent factors that most consistently kill off
Whether it’s been the shift to online working, the even the most promising tech transformations and
spike in online demand, or the increase in cyber revealed antidotes to address them. Following are
assaults, technology has emerged as a critical three of the most common failure modes.
business capability. That reality has injected
a renewed importance and new urgency into Piecemeal activity and limited scope
modernizing the technology function. Companies There is no shortage of technology-transformation
can no longer afford the long timelines and often- initiatives, all of them with good intentions and
disappointing business returns that have hampered promising payoffs. In fact, our latest analysis shows
many of the large tech-transformation projects of that companies are expanding the range of tech-
the past. related transformations (Exhibit 1).

Instead, some technology leaders have pursued But too often companies focus on a series
a new approach that is comprehensive enough to of initiatives without accounting for crucial
account for the myriad interlinkages of modern dependencies that need to be in place to enable
technology joined at the hip with the business so the change. Simply migrating systems to the cloud
that change delivers value, and self-funded so that without also thoughtfully implementing cyber
the scope of the change can continue to expand. strategy, agile, and DevOps, for example, would
We think of this comprehensive approach as “tech leave a company unable to take advantage of
forward.” the automation, scale, and flexibility that cloud-
based systems offer. The other side of the coin
is that activities in one area can have unintended
Counteracting the most devastating consequences in another, often breaking or
tech-transformation failure modes disabling tangential systems. Modernizing
Some companies are starting to see real impact the architecture, for example, changes how
from their tech transformations. In a recent development teams deploy to it; using old methods
McKinsey study, some 50 percent of surveyed results in errors and delays. Successful CIOs,
companies reported moderate to significant in contrast, are explicit in identifying system
impact on realizing new revenue streams, almost dependencies and deliberate in managing them
70 percent reported impact on increasing existing so that the full scope of potential benefits can be
revenue streams, and 76 percent reported impact captured.
on reducing costs.¹
No link to business value
Tech transformations, nonetheless, remain New technologies continually hit the market, many
notoriously difficult and complex. Though many with tempting promises to solve many of tech’s
companies are transforming their tech organizations, ills. Unfortunately, many of these “shiny objects”
about 50 percent of them report that they’re still in which technology functions invest have limited
in the pilot phase (small tech teams working with value to the business due to limited partnering
advanced technologies but isolated from the rest of between technology and the business, the inability
the technology function).² of technology to communicate the value of tech
to the business, and an often unclear sense of the
To understand better what successful tech business value at stake.
transformations look like—as well as what the most

1
McKinsey Global Survey on IT and the Business, August 2020.
2
Ibid.

6 How to become ‘tech forward’: A technology-transformation approach that works

Exhibit 1
Companies are
Companies are pursuing
pursuing a larger
a larger range
range of transformations
of transformations andbeyond
and moving moving beyond
modernizing infrastructure.
modernizing infrastructure.
Over the past three years, infrastructure transformations have led the way as the most-cited type of transformation

60%
pursued in the past two years by

of respondents¹
… but this year, respondents are planning a wider spread of
transformations, with more focus upward along the tech stack

Types of transformations most likely to be pursued over next two years,² %

Digitizing of end-user experience 37

Scaling data and analytics 36
Enhancing IT architecture 31
Modernizing infrastructure 30
Transforming cybersecurity practices 26
Redesigning tech organization 23
Redesigning the IT operating model 22
Changing IT’s delivery model 17
Transforming talent strategy 15
Transforming vendor management 12

Companies that modernized their infrastructure have a more modern, stable, and flexible infrastructure in place—and are
ready to pursue new types of transformations.

¹Q: Of the following types of IT transformations (large-scale change efforts that are more comprehensive than short-term improvement programs), which, if any,
has your organization pursued in the past two years? (n = >450). Data for “Scaling data and analytics” not available for 2017 and 2016.
²Q: Which of the following IT transformations, if any, is your organization most likely to pursue in the next two years? (n = 487).

Top organizations instead are deliberate in Too expensive to sustain

developing a governance program tied to the Tech transformations are expensive. When their
business, grounding each initiative in an explicit ROI lies too far in the future (or is disappointing, as
P&L result and building in specific metrics to track has happened in the past), critical investment is too
progress against business targets. This becomes often pulled back. That doesn’t need to happen.
even more critical in a post-COVID-19 world
in which budgets are tightening and return on Successful transformations, in contrast, frontload
investment (ROI) is essential. activities that unlock value quickly. Those activities

How to become ‘tech forward’: A technology-transformation approach that works 7

 can include agile sourcing strategies, clean- It is important to call out that the identification of
sheeting the portfolio, or optimizing the balance of the activities themselves isn’t the main reveal—
engineering and non-engineering roles—changes CIOs will be familiar with most if not all of them.
that often unlock millions of dollars. This tech-forward model has proven most useful,
however, in helping organizations understand
the scale of needed change and think through
What a ‘tech forward’ transformation interdependencies across vectors and plays.
looks like
Detailed conversations with CIOs as well as our own Vector #1: A reimagined role for technology
experience helping businesses execute complex that’s focused on the business
technology transformations yielded a broad array Effective technology functions maintain
of insights, best practices, and guidelines. We’ve close ties with other business functions, but
synthesized them into a “tech forward” model that best-in-class CIOs take this a step further,
highlights three interconnected vectors, within with technology driving the business. That
which are ten specific “plays,” or domains of activity requires reimagining technology’s role through
(Exhibit 2). technology-led business models (play #1), a

Exhibit 2
Successful technologytransformations
Successful technology transformationsspan
spanthree
threevectors
vectorsofof activity,
activity, each
each
consisting of aa specific
consisting of specificset
setofofplays.
plays.

Vector 1: Reimagine role of technology Vector 2: Reinvent technology delivery

1 4
Tech-forward business strategy V1 Agile@scale software delivery
(new tech-enabled business models
or customer-facing products) 5
2 Next-generation infrastructure services
(cloud, end-to-end automation/
Integrated business and technology NoOps, platform as a service)
management (no silos, product/
platform orientation) with 6
strategic spend allocation
V2 Engineering excellence with top
3 V3 talent (both internal and external);
Steward of digital user do more with less
experience (design
Vector 3: Future-proof the foundation
7
thinking, user-centricity,
seamless integration Flexible technology partner-
with analog) 8 ships (capability-focused,
outcome-based)
Flexible, business-backed architecture rehaul
delivered iteratively (open architecture,
microservices, application programming interfaces)
9
Data ubiquity and advanced
analytics enablement

10
Defenses that preempt evolving
threats (cyber, data privacy)

8 How to become ‘tech forward’: A technology-transformation approach that works

product- and platform-centered operating model performance benefits. Such modern systems
with ingrained strategic funding allocation (play #2), are arranged according to a flexible architecture
and technology functions becoming the steward consisting of self-contained applications connected
of digital user journeys (play #3) given their unique with easy-to-configure application programming
perspective across functions such as marketing, interfaces (APIs) (play #8). A modern technology
sales, and operations. In organizations that have core includes data and analytics systems that
truly reimagined technology’s role, the role of the provide technology teams across the enterprise
CIO is also often elevated. with the high-quality information and powerful
tools they need to gain insights into customer
This vector of change often isn’t the first one we see. and employee preferences, design innovative
Rather, this reimagined role for technology develops applications, and enrich user experiences (play #9).
over time as the other two vectors begin delivering It also enables tech teams to integrate security and
value and the credibility of the technology function privacy protections as they develop solutions, rather
grows. However, this aspiration for technology to than adding them after solutions development
drive business value must be explicitly defined up is complete. This approach greatly accelerates
front or the results may fall short, as they often do. delivery while maintaining or even improving
information security (play #10).
Vector #2: A technology delivery model built for
flexibility and speed The challenge in using this model lies not just in
Modern technology functions set up their delivery coordinating the interdependencies, as challenging
models to keep pace with the fast-evolving needs as that can be; it’s in sequencing the transformation
of customers and employees. Using agile methods, initiatives so that they build value quickly. It is
tech teams prioritize and carry out activities that essential that a tech transformation deliver value
have the greatest potential to help their companies within a year; beyond that, skepticism builds and
realize sought-after performance gains (play #4). support fades. To enable this focus on value, the
Next-generation infrastructure services based in transformation road map should take on a few
the cloud accelerate delivery and stabilize the tech interdependent changes at a time, with a series
environment by automating development, testing, of coordinated efforts, each of which can be
and deployment processes (play #5). To improve the completed within three to six months.
quality and efficiency of their work, modern tech
functions hire highly skilled engineers to deliver Tech forward in action
mission-critical engineering in house (play #6). They
also partner thoughtfully with a variety of vendors, A consumer-services company transforms
ranging from hyperscalers to software-as-a-service its tech function to support better customer
(SaaS) firms to niche engineering organizations experiences
to large-scale systems integrators, for help in A major institution in the consumer-services
building or augmenting capabilities that are more sector was losing business to its rivals, who were
challenging to develop or scale, using contracts that aggressively rolling out new and better digital
reward vendors for producing business outcomes products and channels. Working closely with the
rather than merely augmenting in-house capacity CEO and other business-unit leaders, the CIO
(play #7). determined IT needed to develop a set of digital
channels and products to improve customer
Vector #3: A future-proof foundation of core tech retention, increase share of wallet, and improve
systems that support innovation, collaboration, customer experience (play #1).
and security
Renewing core systems so they support new digital With this clear articulation of how new digital
functionalities, multiple daily production releases, products would deliver business value, the CIO was
and frequent upgrades can provide significant ready to start building them. But he quickly realized

How to become ‘tech forward’: A technology-transformation approach that works 9

 that progress would be difficult unless IT changed of value (play #9). SWAT teams of engineers and
how it developed products that customers actually architects came together to anticipate system-
wanted (play #3), how IT worked with the business reliability issues and their root causes. They
to ensure that the technology products delivered tackled the most urgent ones first and managed
value (play #2), and how teams collaborated to the backlog. They also actively checked that fixes
make better and faster progress. Without these were working and stepped in quickly to address
changes, he knew the company would run into the any that weren’t.
same delays and issues that had dogged its earlier
technology initiatives. At the same time, another team modernized
foundational systems by building out a
Acting on this understanding, the CIO partnered microservices-based architecture for all new
with business leaders to design a new model for development. To enable this shift, more easily
how business and technology would partner. That accommodate new digital solutions, and help
included, for example, creating a single “point of improve time to market, they worked on updated
entry” for any technology requests and frequent cloud-based platforms, which allowed them to
meetings to jointly review and prioritize them. Each use cloud-based data services to rapidly process
month, they reviewed the tech road map against and analyze their data to identify new business
the business strategy. One result was the creation opportunities. Working collaboratively, business
of a fast track for product requests that didn’t and IT teams created almost 50 use cases, such
require significant work, a simple solution to the as improved demand and inventory forecasting,
previous monolithic development process that that have the potential to add as much as $1 billion
every request had to go through. of incremental revenue.

Simultaneously, he implemented a new, agile,

product-engineering model (play #4) where cross- Questions that help orchestrate
functional teams made up of people with design, a successful tech-forward
development, operations, and other expertise transformation
collaborated around a specific user experience To get the sequence of transformation activities
(mobile ordering or setting up an account, for right, executives need to be clear about where
example). To ensure speed and momentum, these they’re going and what their current capabilities
teams were trained to use agile ways of working are. Companies often have an incomplete
together, such as breaking initiatives into two- understanding of these two elements, which
week projects (sprints), piloting new products to creates confusion in the executive suite and will
get user input, and rapidly testing operational derail a tech transformation before it ever gets
effectiveness before scaling. To help focus their started.
work, the teams used design thinking to build clear
pictures of true user needs and pain points. To plot a company’s tech-transformation road
map, we find the following questions particularly
This initial phase of work allowed technology helpful:
teams to roll out the first set of digital offerings
successfully and under budget—and three to — What is your expectation from technology?
five times faster than similar technology projects
undertaken in the past. — Which strategic outcomes are most critical
(for example, speed and quality of delivery)?
With the digital-products workstream well
under way, the CIO focused his attention on — Which are the most urgent pain points and
another cluster of critical dependencies: scaling what causes them?
cloud-based services (play #5), modernizing and
migrating foundational systems to microservices The following questions help executives
(play #8), and leveraging data to find new sources understand the current state of the technology

10 How to become ‘tech forward’: A technology-transformation approach that works

function and its experience with transformation
programs:
The current COVID-19 crisis, of course, is having
— Which, if any, of the ten plays from the tech- a significant impact on how CIOs and businesses
forward approach are in place, and what is their manage tech transformations. Despite the
maturity? pressures it has added to costs, however, the
urgency to get moving and transform has never
— Is transforming your company’s tech one of the been higher, according to many CIOs. But while the
top two priorities in your C-suite? If not, why not? demands placed on the technology function have
grown, so too have the opportunities. Experience
— How well does the technology function support suggests that the most effective transformations
your company’s strategic objectives or digital are not only comprehensive, covering the function’s
ambitions? role, delivery model, and core systems, but also
sequenced to ensure that changes that reinforce
— What tech-transformation efforts has your each other are carried out together. With up-front
company launched to date? What effect have planning focused on business value and careful
they had? What went well, and what didn’t? delivery, a company can bring its technology
function forward and gain the capabilities to thrive in
— What factors might restrict the pace of your challenging digital markets.
tech-transformation efforts? In particular, how
much capital and other resources can the
company devote to tech transformation?

Anusha Dhasarathy is a partner in McKinsey’s Chicago office, where Isha Gill is an associate partner and Naufal Khan is a
senior partner; Sriram Sekar is a senior expert in the New Jersey office, where Steve Van Kuiken is a senior partner.

Copyright © 2020 McKinsey & Company. All rights reserved.

How to become ‘tech forward’: A technology-transformation approach that works 11

 Seven lessons on how
technology transformations
can deliver value
Our annual IT strategy survey shows how technology investments are proving
their worth, especially at companies making more tech-based changes and
bridging more of the technology–business divide.

© Jorg Greuel/Getty Images

March 2021

12
 In the past year, the COVID-19 crisis has made revenue from existing streams, and more than half
clear the business imperative of making technology- cite the creation of new revenue streams: for example,
driven changes, which are more ubiquitous now than a new product line or new business (Exhibit 1).
ever.1 Indeed, our latest McKinsey Global Survey on
technology and the business suggests that even in The results also suggest that these investments
the crisis’s earlier days,2 respondents were reporting aren’t one-off attempts to catch up, with nearly all
progress on their integration of technology and respondents reporting plans to pursue at least one
business—and that these efforts were creating transformation play in the next one to two years.
tangible business value across four measures,
including new revenue streams and lower costs.
The results also suggest that, on average, some Lesson #2: People-focused plays result
transformation activities result in more impact in the most value
than others (namely, those related to talent and With regard to impact, the results suggest that
capabilities). And according to the data, the not all types of transformations are created
companies with top-performing IT organizations3 equal. Across the ten transformation initiatives,
have differentiated themselves from others in their respondents say that changes to their companies’
efforts to create value, adopt new technologies, and people and talent strategies are among the highest-
bring technology and business closer together. value moves to make (Exhibit 2). At companies that
have transformed their approaches to technology
More specifically, the results point to seven key talent—that is, changed practices to attract, retain,
lessons about technology transformations.4 and upskill talent with digital and engineering skills—
respondents report the greatest impact on all four
measures of business impact.
Lesson #1: Technology investments are
creating significant business value Meanwhile, those that pursued changes to their
In the latest survey, companies’ tech-transformation sourcing strategies report a significant impact on
activities appear to be paying off. The survey asked three of the four measures: realizing new revenue
about ten different types of transformation initiatives streams, reducing costs, and improving employee
(for more information on the ten plays in our “tech experience. And according to the results, scaling up
forward” approach,5 see sidebar, “A tech-forward data analytics is a critical enabler of new revenue
transformation”).6 According to respondents, more and increases to existing revenue streams. By the
than three-quarters of the initiatives their companies same token, respondents whose companies saw no
pursued have yielded some or significant cost or negative value across these measures say they
reductions and improvements to employee were least likely to pursue talent transformations or
experience. What’s more, more than two-thirds of the scaling of their data and analytics capabilities.
respondents say these change efforts increased

1
“How COVID-19 has pushed companies over the technology tipping point—and transformed business forever,” October 5, 2020, McKinsey.com.
2
The online survey was in the field from April 14 to April 30, 2020, and garnered responses from 487 participants. Of these, 275 have a
technology focus, and the remaining 212 are C-level executives representing other functions. The participants represent the full range of
regions, industries, company sizes, and tenures. To adjust for differences in response rates, the data are weighted by the contribution of each
respondent’s nation to global GDP.
3
We define “top-performing IT organizations” as those that, according to respondents, had an average effectiveness score in the top 25 percent
of the survey sample, based on ratings of 15 key IT activities that were tested in the survey.
4
We define “technology transformations” as large-scale change efforts—which are more comprehensive than short-term improvement
programs—to modernize the technology function.
5
See “How to become ‘tech forward’: A technology-transformation approach that works,” on p. 5 of this compendium.
6
The ten initiatives the survey asked about were changing IT’s delivery model (for example, lean IT, agile at scale); digitizing of end-user
experience (that is, digitization of end-to-end business processes or end-user/customer journeys across the organization); enhancing IT
architecture (for example, using a flexible, services-based architecture, modernizing legacy applications); modernizing infrastructure (for
example, cloud migration, infrastructure automation); redesigning the IT operating model (for example, establishing a stronger partnership
between the business and IT functions, changing processes such as budgeting and IT demand management, organizing around product-
focused teams); redesigning the technology organization to support new digital products or services; scaling data and analytics (for example,
deploying artificial-intelligence models, building next-generation data platforms); transforming cybersecurity practices (for example,
strengthening defenses against cyberthreats and data-privacy threats, proactively running cyberthreat drills); transforming talent strategy
(for example, changing practices to attract, retain, or upskill talent with digital and engineering skills); and transforming vendor management
(for example, revamping sourcing strategy, consolidating suppliers, entering new types of strategic partnerships).

13 Seven lessons on how technology transformations can deliver value

A tech-forward transformation

Through detailed conversations with 3. Steward of digital user experience 8. Flexible, business-backed
nearly 700 chief information officers at (design thinking, user centricity, architecture rehaul delivered
some of the world’s largest companies, as and seamless integration with iteratively (open architecture,
well as through our own experience helping analog technologies) microservices, application
businesses execute complex technology programming interfaces
transformations, we’ve synthesized our 4. Agile@scale software delivery
findings into a “tech forward” model of 9. Data ubiquity and advanced-
guidelines and best practices. This model 5. Next-generation infrastructure analytics enablement
includes the following ten “plays,” or services (cloud; end-to-end
domains of activity: automation/no operations, or NoOps; 10. Defenses that preempt evolving
platform as a service) threats (cyber, data privacy)
1. Tech-forward business strategy (new
tech-enabled business models or 6. Engineering excellence with top
customer-facing products) talent, both internal and external (do
more with less)
2. Integrated business and technology
management (no silos, and a product/ 7. Flexible technology partnerships
platform orientation with strategic (capability focused, outcome based)
spend allocation)

Web <2021>
<Seven lessons on how technology transformations can deliver value>
Exhibit
Exhibit <1>1 of <6>

Most respondents
Most respondentsreport
reportsome
someororsignificant
significant impact
impact from
from their
their companies’
companies’
technology transformations.
technology transformations.
Impact from technology transformations over the past 2 years, % of respondents,1 n = 487

Realization Increased Reduced Improved

of new revenue from costs employee
revenue existing experience
streams streams

21 31 34
Significant impact 19

Some impact 40 47
45 45

No impact 24 21 16 14

1 5 4
Negative impact 3 3
1
Figures do not sum to 100%, because respondents who answered “don’t know” are not shown.

Seven lessons on how technology transformations can deliver value 14

 Web <2021>
<Seven lessons on how technology transformations can deliver value>
Exhibit 2 of <6>
Exhibit <2>

Talent- and
Talent- and sourcing-related
sourcing-relatedtransformations
transformationstend
tendtoto result
result inin more
more value
valueto
to both the top and bottom line.
both the top and bottom line.
% of respondents reporting significant impact from Top-line measures
technology transformations in the past 2 years Bottom-line measures

To realize new revenue streams

Redesigning technology organization to support new digital offerings 22
Transforming vendor-management and sourcing strategy 22
Scaling data and analytics 24
Transforming talent strategy 30

To increase revenue from existing streams

Digitizing end-user experiences 26
Enhancing IT architecture 26
Scaling data and analytics 33
Transforming talent strategy 33

To reduce costs
Digitizing end-user experiences 35
Transforming vendor-management and sourcing strategy 35
Transforming talent strategy 37

To improve employee experience

Transforming cybersecurity practices 41
Transforming vendor-management and sourcing strategy 42
Transforming talent strategy 45

Yet even though the people-focused initiatives link Lesson #3: Talent remains the holy
most closely with value creation, they are the least grail of technology transformations—
likely ones that companies plan to pursue in the valuable to pursue but difficult
future (Exhibit 3). Instead, the largest shares of to execute
respondents predict their companies will pursue Not only do the transformations focused on
digitization of the end-user experience, scaling of talent strategy stand out in their value potential,
data and analytics, and enhancements of IT but they are also much more commonplace at
architecture. That is a notable shift from our past top-performing companies. Top-quartile
three annual surveys, when infrastructure respondents are more than three times likelier
transformations were the most-cited play that than their bottom-quartile peers (41 percent,
companies pursued. Now, respondents are half as compared with 12 percent) to say they’ve pursued
likely to say their companies will modernize a transformation of their talent strategy in
infrastructure in the next one to two years. recent years.

15 Seven lessons on how technology transformations can deliver value

Web <2021>
<Seven lessons on how technology transformations can deliver value>
Exhibit 3 of <6>
Exhibit <3>

Despitetheir
Despite theirhigh
high value
value potential,
potential,people-focused
people-focused initiatives
initiatives are
are the
the least
least likely
likely to be pursued by companies.
to be pursued by companies.
Types of transformations most likely to be pursued over next 2 years,1 % of respondents

Digitizing end-user experience 37

Scaling data and analytics 36

Enhancing IT architecture 31

Modernizing infrastructure 30

Transforming cybersecurity practices 26

Redesigning the tech organization 23

Redesigning the IT operating model 22

Changing IT’s delivery model 17

Transforming talent strategy 15

Transforming vendor management 12

1
Respondents who answered “other” or “don’t know/not applicable” are not shown. We define “technology transformations” as large-scale change efforts that
are more comprehensive than short-term improvement programs.

Yet the need to address talent is universal and urgent. Lesson #4: The talent challenge has
Respondents believe that more than 40 percent of clear implications for sourcing
their workforce will need to be either replaced or Perhaps because companies have found talent-
fundamentally retrained to make up for their related changes so difficult to pursue, responses
organizations’ skills gaps. But only 15 percent of suggest that they have been using new or different
respondents say their companies plan to pursue a approaches to sourcing to fill some of the gaps. We
talent-strategy transformation in the next two asked technology executives and respondents
years, even though the talent challenge remains about recent changes to their technology-sourcing
considerable (Exhibit 4). At companies that have strategies, and they tend to say that reliance on
pursued recent transformations, the top challenges external providers to support both core IT activities
to doing so continue to revolve around talent as well and digital activities has increased. Among
as culture: namely, skill gaps and cultural differences, respondents reporting changes to their sourcing
the difficulty of changing cultures and ways of strategies, 47 percent say they are relying more on
working, and difficulty finding talent to fill new roles— sourcing partners to supplement internal
which is as challenging for top performers as it is for capabilities. Overall, most respondents say their
everyone else. Talent also appears to impede companies have engaged partners in a range of
progress at the companies that haven’t pursued sourcing models, from traditional time-and-
technology transformations; 42 percent of materials to managed services and joint ventures.7
respondents say they have stuck with the status quo
because it’s difficult to source the talent they need.

7
Other changes to sourcing strategy that the survey asked about: moving work on digital or front-end applications from global providers to niche
sourcing partners; changing commercial models (that is, from time-and-materials or contingent workers to managed service providers); and
pursuing larger sourcing models (for example, joint ventures or build-operate-transfer models where partners develop assets that they hand
over to an organization to operate).

Seven lessons on how technology transformations can deliver value 16

 Web <2021>
<Seven lessons on how technology transformations can deliver value>
Exhibit 4 of <6>
Exhibit <4>

Talent-relatedand
Talent-related andcultural
cultural issues
issuespose
posethe
thegreatest
greatest challenges
challenges to technology
to technology
transformations.
transformations.
Challenges to changing organizations’ technology operating models, % of respondents1

Skill gaps It has been difficult It has been difficult Traditional Integrating
and/or cultural to change culture to find the right talent teams have new technologies
differences have and ways of working to fill new roles struggled to keep into core
arisen as we from being project (eg, scrum masters, up with the pace architecture has
change the focused to product product owners) at which digital been harder
way we work focused on digital teams teams work than expected
54 52 37 35 30

Likelihood of Speed of digital The business’s Newer ways of Newer ways of

cyberthreats delivery and releases needs have working3 have limited working3 have reduced
and/or security has made it hard for outpaced the our ability to manage valuable governance
breaches the business speed of demand and/or track or transparency on
has increased to keep up2 IT delivery delivery teams’ work
30 25 25 21 21

1
Out of 12 challenges that were offered as answer choices. Question was asked only of respondents who said the target state of their organizations’ technology
operating models are digitally integrated or fully digital; n = 334.
2
Eg, not enough time to train end users on the new changes, poor adoption of products by end users.
3
Eg, agile, cross-functional teams.

Lesson #5: No silver bullets—the This result is consistent with our experience
top performers execute more that building capabilities in one area often
transformation plays than others requires the development of others at the same
We looked more closely at the results from a subset time because these capabilities reinforce one
of respondents whose companies are in the top another. For example, companies that work on
quartile of performance on core technology scaling their agile-development capabilities often
activities, or our “top performers.” These companies invest in hiring new talent—and accelerating their
not only have seen more value as a result of their cloud or automation strategies to enable continuous
technology transformations but also have focused integration/continuous delivery (CI/CD) and
on multiple initiatives—and more so than their DevOps—in parallel.
peers. On average, they have run five out of ten
transformation initiatives in recent years, versus
three initiatives at the bottom-quartile companies.

17 Seven lessons on how technology transformations can deliver value

 Lesson #6: The broader use of What’s more, the top-performing respondents
advanced technologies supports report using a slightly larger suite of technologies.
greater value creation Out of the six we asked about,8 nearly one-quarter
The results suggest that overall, advanced of top performers say their companies used four
technologies can generate outsize value in tech to six advanced technologies, compared with
transformations. Forty-four percent of respondents 10 percent of all other respondents. Inversely, the
reporting the use of the Internet of Things (IoT) top performers are half as likely as others to report
or edge-computing technologies in recent using only one advanced technology.
transformations say they saw significant cost
reductions—compared with an average of only
31 percent who saw significantly reduced costs Lesson #7: Bridging the business-
overall. Yet these technologies are relatively technology chasm is critical to
uncommon. Only one-quarter of respondents say outperformance
their companies use IoT in the first place. At the same Beyond their focus on talent, deployment of new
time, 45 percent of respondents at companies using technologies, and a broad transformation agenda,
the cloud to process data at scale report a significant the top performers also follow several practices that
improvement in employee experience from their foster a stronger partnership between technology
transformations, versus an average of 34 percent and the business (Exhibit 5). At top-performing IT
of all respondents. organizations, 57 percent of respondents say their
senior leaders are very involved in strategic planning,
versus 17 percent in the bottom quartile.

At top-quartile organizations, 57 percent

of respondents say their senior tech
leaders are very involved in strategic
planning—versus 17 percent in the
bottom quartile.

8
The survey asked respondents about the following technologies, and which their organizations had deployed at scale in their technology
transformations: automation; advanced analytics (that is, artificial-intelligence and machine-learning-based solutions); large-scale data
processing through the cloud; design thinking (that is, user-centered product development); the Internet of Things or edge computing; and
advanced mobility (for example, use of 5G mobile networks).

Seven lessons on how technology transformations can deliver value 18

 Web <2021>
<Seven lessons on how technology transformations can deliver value>
Exhibit <5> of <6>
Exhibit 5
Top performers bridge the business and technology gap at significantly higher
Top performers bridge the business and technology gap at significantly higher
rates than others.
rates than others.
% of respondents who say their most senior technology leaders are very involved in shaping
enterprise-wide business strategy and agenda
Bottom-quartile Top-quartile
performers performers1
17 57

0 20 40 60 80 100

Companies’ process for setting IT/digital strategy, % of respondents2

Bottom-quartile Top-quartile
performers performers1

Business and IT/digital strategies are cocreated equally by

business and IT and revisited throughout the year in an iterative process 8 4.3× 34

Business and IT/digital strategies are cocreated

equally by business and IT every year 11 22

The business strategy is developed first, and the

IT/digital strategy and priorities derive directly from it 39 34

The business and IT/digital strategies are developed 22 7.3× 3

independently of one another; there is no link between them

Not applicable; we do not set a separate IT/digital strategy 13 7

Respondents who reported an average effectiveness score in top 25% of the sample, based on ratings of 15 key IT activities that were tested in the survey.
1

Respondents who answered “don’t know” or “not applicable; we do not share the IT/digital strategy with the rest of the organization” are not shown. For top-
2

quartile respondents, n = 125; for bottom-quartile respondents, n = 120.

At these organizations, IT and business teams also are Finally, the top performers are much more focused
much more likely to work together to both develop than others on measurement, even for metrics
strategy and deliver technology. Top-quartile that aren’t technology-specific. According to
respondents are nearly three times as likely as their respondents, top-quartile companies are more likely
bottom-quartile peers to say that business and IT to track their technology organizations’ performance
cocreate corporate and technology strategies. And as well as team performance across the company,
they are more than four times likelier than their bottom- using more business-oriented metrics such as user
quartile peers to have a digitally integrated or fully satisfaction, time to market, and financial impact.
digital operating model, in which digital and business-
oriented teams—or cross-functional teams—all deliver
technology across the organization (Exhibit 6).9

9
For more on technology operating models, see Naufal Khan, Gautam Lunawat, and Amit Rahul, “Toward an integrated technology operating
model,” October 2, 2017, McKinsey.com.

19 Seven lessons on how technology transformations can deliver value

Web <2021>
<Seven lessons on how technology transformations can deliver value>
Exhibit 6of <6>
Exhibit <6>

Top performers are more likely

likely than
thanothers
others to
to involve
involveboth
bothdigital
digital and
and business-oriented teams in technology delivery.
business-oriented teams in technology delivery.
Organizations’ current technology operating model, Top-quartile performers2
% of respondents1 Bottom-quartile performers

Predigital Pilot programs Digital factories Digitally integrated Fully digital

Traditional operating Traditional operating Traditional operating Technology delivered All technology-
model for delivering model with digital pilot model with an at-scale at scale by both digital delivery teams
technology solutions programs (eg, small digital factory3 and traditional teams4 operate in a digital
ring-fenced teams manner, using entirely
that work in new ways) modern software,
infrastructure,
and tools

46 29
36
14
18

8 24 10 8 4

1
Respondents who answered “don’t know” are not shown. For top-quartile respondents, n = 125; for bottom-quartile respondents, n = 120.
2
Respondents who reported an average effectiveness score in top 25% of the sample, based on ratings of 15 key IT activities that were tested in the survey.
3
A group of delivery teams that is dedicated to building digital products and is largely separate from the traditional technology organization.
4
Teams are not siloed or incubated and are governed by a single operating model.

The contributors to the development and analysis of this survey include the following members from McKinsey’s Chicago
office: Anusha Dhasarathy, a partner; Ross Frazier, an associate partner; Naufal Khan, a senior partner; and
Kristen Steagall, a consultant.

Copyright © 2021 McKinsey & Company. All rights reserved.

Seven lessons on how technology transformations can deliver value 20

2. Reimagining the
role of technology
 The CIO challenge: Modern
business needs a new kind of
tech leader
As technology becomes increasingly important, an organization’s success
depends on whether the CIO can move from being a functional to a strategic
business leader.

by Anusha Dhasarathy, Isha Gill, and Naufal Khan

© skynesher/ Getty Image

January 2020
22
 “There’s no worse time than now to be an average This reality requires a CIO to first come to terms
CIO.” These words, uttered by an executive at a with the scope of the transformation itself. In
recent conference, neatly capture the intense our experience, it’s been helpful to think about it
pressure on CIOs. For years, executives have along three vectors:
stressed the need for CIOs to move beyond
simply managing IT to leveraging technology to 1. Reimagine the role of technology in the
create value for the business. This priority is now organization. This vector includes establishing the
a requirement. New technologies have been at role of technology as a business and innovation
the center of trends—from mobile-first consumer partner to design a tech-forward business
shopping preferences to the promise of artificial strategy (for example, tech-enabled products and
intelligence in critical decision making—that have business models), integrate tech management
reshaped the competitive landscape and disrupted across organizational silos, and deliver excellent
business models. For this reason, companies need user experiences.
to be tech forward: technology needs to drive the
business. 2. Reinvent technology delivery. IT needs to
change how it functions by embracing agile;
Despite this pressing need, of the organizations improving IT services with next-generation
that have pursued digitization, 79 percent of them capabilities such as end-to-end automation,
are still in the early stages of their technology platform as a service, and cloud; building small
transformation, according to McKinsey’s 2018 IT teams around top engineers; and developing
strategy survey.¹ Legitimate factors are delaying flexible tech partnerships.
progress, from the scale of the change to the mind-
boggling complexity of legacy systems. We believe, 3. Future-proof the foundation. To keep pace with
however, that one of the biggest issues is that many rapid technological advancements, organizations
CIOs have not accepted the degree to which their need to implement a flexible architecture
role needs to expand beyond cost and performance supported by modular platforms, enable data
responsibilities in order to transform IT into a core ubiquity, and protect systems through advanced
driver of business value. cybersecurity.

Three vectors of a holistic Five traits of a transformative CIO

transformation For IT to become a driver of value, the
Before understanding the responsibilities of the new transformative CIO also needs a new set of skills
CIO, it’s important to understand the nature of tech and capabilities that embody a more expansive
transformations themselves. In most cases we’ve role. In working on tech transformations with
observed, tech transformations are implemented hundreds of CIOs, we have identified five CIO
as a set of disjointed initiatives across IT. That leads traits that we believe are markers for success.
promising developments to stall out or underdeliver.
We have found that a tech transformation must 1. Business leader
be holistic to deliver full business value. Creating To help technology generate business value,
powerful customer experiences, for example, the transformative CIO has to understand
requires a data architecture to track and make business strategy. Findings from our 2018 IT
sense of customer behavior. Architecting modular strategy survey reveal that companies with
platforms needs revamped approaches to hiring in top IT organizations are much more likely than
order to get top-flight engineers. others to have the CIO very involved with shaping

1
“Can IT rise to the digital challenge?” October 2018, McKinsey.com.

23 The CIO challenge: Modern business needs a new kind of tech leader
Sidebar

Questions for the CIO

• Can I clearly articulate the business’s goals?

• Are the most important technology initiatives delivering quantifiable business value to the company?

• What percentage of technology resources is focused on work that drives business goals versus maintenance?

the business strategy and agenda, and strong get firsthand feedback on both products and the
performance on core IT tasks enables faster customer’s experience of doing business with the
progress against a company’s digital goals.² CIOs company. He uses these perspectives to inform
who can make this leap tend to take the following his technology decisions.
actions.
Take responsibility for initiatives that generate
Learn the business inside and out revenue
The scope of an IT transformation means that CIOs CIOs can further develop business acumen
must be prepared to interact with the business in by taking responsibility for initiatives that
different ways. We have found, for example, that generate business impact, such as building an
the best CIOs go far beyond meeting with the e-commerce business, or by working with a
C-suite or attending strategy meetings. They invest business-unit leader to launch a digital product
time with functional and business-unit leaders and then measure success by business-
and managers to gain an in-depth understanding impact key performance indicators (KPIs), not
of business realities on the ground and go out technology KPIs. Such efforts allow CIOs to
of their way to develop a nuanced and detailed build a deep understanding of the business
understanding of customer issues. CIOs do this implications of technology, such as customer
by continually reviewing customer-satisfaction abandonment because of slow download times
reports, regularly monitoring customer-care calls, on a site or other poor user experiences.
and participating in user forums to hear direct
feedback. As part of a digital transformation, for instance,
the CIO at a large financial institution committed
As one large financial institution set out to to developing digital products to help the
build its digital products, the business and business scale its presence in a new market.
technology teams jointly led user listening and While the CIO already understood how to build
feedback panels early and often throughout systems to support financial products, he and
the development process. Both technology and his team had limited experience in creating new
business leaders made it a priority to attend these digital products to sell directly to consumers. So
panel discussions so that they could effectively the team created a program built on rapid test-
guide their teams on developing products that and-learn cycles to identify what mattered to
would best address the needs of end customers. customers and meet those needs. Subordinating
The CIO of a B2B technology-services company, tech decisions to customer needs was crucial in
meanwhile, meets customers on a regular basis to allowing the CIO and his team to develop a digital

2
“Can IT rise to the digital challenge?” October 2018, McKinsey.com.

The CIO challenge: Modern business needs a new kind of tech leader 24
 offering that succeeded where it mattered: with based on common goals, mutual responsibility,
consumers. and accountability. According to a McKinsey
survey on business technology, in fact, the
Get on boards companies in which IT plays a partner role
Developing a deeper well of business knowledge in digital initiatives are further along in both
often requires CIOs to extend their networks implementation and achieving business impact.⁴
beyond the organization. One of the best ways to
do that is by joining the board of another company. To kick-start the transformation journey, the
A third of the boards of companies within the CIO of a transportation-and-logistics company
Fortune 500 today include a former CIO or CTO, made it her first priority to meet with every
and that number continues to increase.³ single business leader to understand their goals
and issues and to set expectations on how
2. Change agent they could best work together, by clarifying, for
A full technology transformation is not about example, what the business side could expect
moving to the cloud or embracing new IT solutions. to get from IT in a consultant role versus IT
It also involves infusing technology into every as a service provider or partner. This effort
strategy discussion and process throughout to understand what mattered to each leader
the organization. Driving a transformation established trust, and from each of these
around the three vectors we laid out earlier discussions it became clear that the business
(reimagining the role of technology, reinventing wanted a true partnership with technology
technology delivery, and future-proofing the and understood what it meant. The CIO further
foundation) starts with a CIO mind-set that both built on the relationship with the business by
acknowledges the need for transformative prioritizing initiatives in the tech transformation
change and commits to a multiyear journey. that addressed business needs and working
closely with business leaders to drive progress.
Partner with business leaders This active collaboration ensured that the
Generating support for a transformation among products and services IT developed were
business leaders across the organization requires adopted.
creating true partnering relationships with them

3
“The digital CIO has arrived,” MIT Sloan CIO Symposium, 2016.
4
“Partnering to shape the future—IT’s new imperative,” May 2016, McKinsey.com.

Sidebar

Questions for the CIO

• Do leaders in the C-suite have a clear understanding of why a tech transformation is important?

• Do you have partner-level relationships with people in the C-suite in developing the vision and plan for both
business and IT?

• Is your tech transformation actively incorporating each of the three vectors of change?

• Do you have a “war room” to manage the transformation that can solve problems as well as track progress?

25 The CIO challenge: Modern business needs a new kind of tech leader
Sidebar

The CEO’s role in making the CIO successful

The stage is set for CIOs both to lead a successful technology transformation and to influence business strategy. They
can’t do it alone, however. The CEO must create an environment where the CIO can thrive. Here are a few things CEOs
can do:

• establish a strategic role for the technology function

• elevate the CIO to report directly to the CEO

• rebalance technology investments and track their business value

• prioritize the development of a world-class tech workforce

For more on this topic, read “The CEO’s new technology agenda,” on McKinsey.com.

Articulate the ‘why’ which allows transformation teams to better

Gaining support for a transformation requires that remove roadblocks and to allocate people and
stakeholders understand that true change will spend when and where they are needed. To
come only from tackling all three transformation actively manage this process, such CIOs also
vectors in a strategic, interlinked manner. That put in place a “war room,” a dedicated team
means not just explaining how this three-pronged that ensures transformation initiatives are
approach is better for IT but also clarifying delivering value by actively tracking progress
how it drives business goals and how it can be and helping to break through root-cause
implemented. When considering a shift to cloud, issues.
for example, executives tend to understand it
first as a cost-saving opportunity. But in helping This was the approach taken in a large global
executives understand the full range of cloud retailer’s digital and technology transformation.
benefits—improved speed to market, better The CIO set up a transformation war-room
developer productivity, and improved resiliency team that worked jointly from the beginning
and disaster recovery—CIOs can help them see with leaders outside the IT function, including
how the cloud can unlock new revenue models marketing, operations, sales, and e-commerce.
and services tied to business priorities. Together, they created detailed work plans.
This detailed early planning revealed which
Have an integrated plan that highlights risks and systems needed to be upgraded and when.
dependencies beyond IT The war-room team actively tracked progress
Large IT initiatives have always required detailed and quickly escalated issues for speedy
planning, but business-oriented CIOs ensure that resolution. The results were clear: a fivefold
transformation plans account for dependencies jump in digital sales, and project delivery four
outside of IT, such as marketing campaigns or times faster than projects of similar scope had
legal implications. They approach planning as a previously taken.
dynamic process rather than something static,

The CIO challenge: Modern business needs a new kind of tech leader 26
 Sidebar

Questions for the CIO

• Are the top people in IT really stars in their field?

• Do you rely exclusively on HR to find your talent?

• Do you have a clear view of the talent you need in the next three years—and a plan to develop it?

• What percentage of the best people you hire are still with you two years later?

3. Talent scout The head of technology and analytics at a large

Nearly half of respondents to McKinsey’s 2018 retail organization set up a talent war room to
IT strategy survey cite skill gaps on traditional hire data scientists and engineers. As part of this
teams as the top obstacle to a successful digital effort, the war-room team revamped recruitment
transformation.⁵ So CIOs need to focus not just on and onboarding processes by using different
recruiting top people but also on retaining them. talent sources, such as HackerRank and General
Two solutions have proven effective. Assembly, and by updating candidate screenings
and interviews with appropriate assessments
Reimagine how to attract tech stars of technical and other skills, such as coding
Companies can reap tremendous benefits and collaboration. In addition, they led weekly
from outsourcing. In the oil and gas industry, check-ins to track the talent funnel and adjust the
for example, the outsourcing of application process as needed.
development grew 50 percent between 2014
and 2018.⁶ But that needs to change, especially Build up internal talent
around the most crucial capabilities. CIOs who Getting good people doesn’t matter if you can’t
want to reinvent tech’s role need tech stars, keep them. Top CIOs, therefore, develop diverse
particularly the best engineers. By hiring the best career paths so that top talent can advance in
tech people, we’ve seen companies reduce their their own areas of strength—for example, by
technology costs by as much as 30 percent while letting a top-notch software engineer advance
maintaining or improving their productivity.⁷ CIOs while continuing to code design software rather
need to move quickly. In just 18 months, one CIO at than forcing her to manage others in order to
a transportation-and-logistics company radically succeed.
reshaped its talent profile. All the direct reports
and approximately 50 percent of tech employees Retraining the existing tech workforce also
were new, and 80 percent had transitioned to needs to be an important element of this platform.
different roles. The CIO of a large consumer company made

5
“Can IT rise to the digital challenge?” October 2018, McKinsey.com.
6
Dhingra, Sverre Fjeldstad, Natalya Katsap, and Richard Ward, “A new mandate for the oil and gas chief information officer,” November 2019,
McKinsey.com.
7
Klemens Hjartar, Peter Jacobs, Eric Lamarre, and Lars Vinter, “It’s time to reset the IT talent model,” March 5, 2020, Sloan Management
Review, sloanreview.mit.edu.

27 The CIO challenge: Modern business needs a new kind of tech leader
Sidebar

Questions for the CIO

• Do you meet or speak with IT employees who are on the front lines at least once a week?

• Do you have a way to accurately measure and track people’s attitudes across the IT department?

• Are your top engineers happy with their work?

• How often do you publicly celebrate success and support noble failures?

digital and analytics upskilling one of the established various community-building and
company’s key strategic priorities, launching knowledge-sharing efforts—hackathons,
an enterprise-wide program, in tandem with “dev days,” tech spotlights, brown-bag
HR’s learning team. The program invested in lunches—where product managers,
an online learning portal to create personalized developers, data engineers, and architects
online learning experiences based on an could meet on a weekly basis to share details
employee’s goals and learning needs. These about their projects and bring up ideas or
were supplemented by other programs, issues for discussion. The CIO attended and
including in-person training, top management actively participated.
immersion sessions, and the cultivation of an
in-house expert network that people could tap Model and support true collaboration
on specific topics. Promoting collaboration across technology
teams and between the business and
4. Culture revolutionary technology is one of the most crucial
An effective talent strategy requires a culture prerequisites for a successful transformation.
that supports talent. Top-quartile IT organizations are more
likely to have an integrated or fully digital
Build a true engineering community operating model, according to McKinsey’s
Pay matters, of course, but top people want to 2018 IT strategy survey.⁸
go where they’re valued. One way to create that
kind of environment is to provide engineers In practice, CIOs can enable collaboration
with more autonomy by reducing the number of if they’re willing to relinquish some control.
managers and often-bureaucratic processes, One CIO at a financial-services firm realized
such as time-consuming reports and multiple that for his people to increase their impact,
rounds of approval. they had to be more closely tied to business
teams. So he embedded them into cross-
Creating ways for cohorts of similar skill sets functional teams aligned around specific
to get together can be a powerful way to products, relying on informal networks of
share best practices and foster a sense of guilds and chapters to provide guidance
community. The CIO of a software company and light oversight. The most effective

8
“Can IT rise to the digital challenge?” October 2018, McKinsey.com.

The CIO challenge: Modern business needs a new kind of tech leader 28
 Sidebar

Questions for the CIO

• Do the questions about technology that leaders in the C-suite ask reflect a true understanding of the impact of
tech decisions?

• When you explain the ramifications of tech decisions, do leaders really understand you?

• How often do company leaders reach out to you for substantive guidance about how tech can improve their
business?

CIOs ensure this level of collaboration is the of a B2B technology-services company found this
norm within IT itself as well. This is particularly level of insight so important that he asked the CIO
important around cybersecurity. IT can radically to present periodically to the board on technology-
reduce cycle times and maintain effective led business models.
security by incorporating security early into
development and working closely with the This role was particularly important when a retail
cybersecurity team on an ongoing basis. giant was looking to acquire an analytics company.
The CIO and his leadership team were involved
5. Tech translator from the very beginning in determining the data
In the past, IT transformations have often proven and analytics capabilities needed to fulfill the
expensive, time consuming, and short on value, company’s business strategy. They performed
and this has made some companies leery of deep-dive technical assessments, system and
undertaking them again. To address this issue data-platform compatibility reviews, and tests
and build trust, the best CIOs play an active role of vendor capabilities. The CIO ran a pilot with
in educating leaders about technologies and a business unit and operations team for three
their applications for the business. months to determine whether the final vendor
could deliver on its capabilities. At the end of
Make the business implications of tech decisions the process, the business was able to make an
clear informed decision.
Many tech decisions don’t get sufficient business
scrutiny beyond cost and high-level strategy
discussions. Transformative CIOs don’t settle
for that kind of interaction, articulating instead These skills are the tools that enable a CIO’s
how a proposed solution solves the underlying ability to transform IT. And in an increasingly tech-
business problem, what alternative approaches driven business landscape, they position CIOs as
exist, and the pros and cons of each. The CEO legitimate contenders to lead businesses as well.

Anusha Dhasarathy is a partner in McKinsey’s Chicago office, where Isha Gill is an associate partner and Naufal Khan is a
senior partner.

Copyright © 2020 McKinsey & Company. All rights reserved.

29 The CIO challenge: Modern business needs a new kind of tech leader
 Products and platforms:
Is your technology
operating model ready?
Forming an integrated technology model creates value but can be
hard to do. The right approach centers teams on tech products and
platforms, focuses them on business goals, and helps them prioritize
technology work.
by Ross Frazier, Naufal Khan, Gautam Lunawat, and Amit Rahul

© Getty Images

February 2020
30
 More and more companies are choosing to operate it in the manner of a digital native. This
remove the divide between their digital programs needn’t be the case. To achieve the technical
and their traditional IT delivery by bringing them agility of a digital native, it is often enough to
into a single technology operating model. Two form integrated, cross-functional technology
years ago, just 18 percent of respondents to teams, which define forward-thinking
McKinsey’s global survey of technology leaders technology organizations.⁴
said their companies had either converted their
digital and IT teams to a single operating model or Moving to an integrated technology operating
developed a fully digital model.¹ In our most recent model does require significant change.
survey, that percentage nearly doubled: one-third Companies can ease the transition by
of respondents say their technology organizations taking three actions: organizing technology
are integrated or fully digital. Sixty-seven percent teams around user-facing products and
say their companies aspire to make the shift to an the underlying platforms that enable them,
integrated or fully digital technology organization.² creating a governance structure to keep
the technology organization focused on the
Combining digital operations (in which cross- business’s strategic priorities, and establishing
functional teams apply new technologies and a rigorous system for prioritizing and delivering
ways of working, such as agile, to improve user technology work. In this article, we offer a
experiences) and traditional IT delivery (in which closer look at what these actions involve and
technical specialists develop and maintain how to get them right.
core business systems according to traditional
methods) has evident benefits. According
to survey respondents, companies with an Reorganizing technology teams
integrated or fully digital technology model are To begin integrating digital and IT operations,
30 percent less likely than other companies to technology and business leaders should
face digital-transformation challenges and less map their companies’ technology activities
than half as likely to face issues in integrating new and assets with respect to two categories:
digital efforts with their core architecture. Notably, products and platforms (Exhibit 1). Products
respondents at companies with integrated digital are the technology-enabled offerings used by
and IT operations are 60 percent more likely customers and employees. Their immediate
than respondents at other companies to say their and primary purpose is to enable users to
companies’ investments in technology create perform activities that create value, in line
business value.³ with a business’s objectives. For example, a
retailer’s search product contributes business
Combining digital programs and IT operations value by making it easy for customers to
is also more practical than other changes that find items on a website or mobile app. Its
business leaders might consider as part of a effectiveness might be measured with
digital transformation. Executives sometimes conversion-to-sale metrics and enhanced
assume that once their digital activities are by improvements to search algorithms. An
running well, the next step in their digital evolution industrial-equipment maker might equip its
should be a far-reaching, potentially disruptive salespeople with a configuration product
effort to reorganize the entire company and

1
In a digitally converted technology organization, digital and traditional IT teams are not siloed or incubated and are governed by a single
operating model. In a fully digital technology organization, all technology teams operate in a digital manner, using modern software,
infrastructure, and tools. For more, see “Can IT rise to the digital challenge?,” October 2018, McKinsey.com.
2
“Managing the fallout from technology transformations,” February 2020, McKinsey.com.
3
Ibid.
4
Klemens Hjartar, Krish Krishnakanthan, Pablo Prieto-Munoz, Gayatri Shenai, and Steve Van Kuiken, “The CEO’s new technology agenda,”
November 2019, McKinsey.com.

31 Products and platforms: Is your technology operating model ready?

Exhibit 1
An integrated technology model centers on products and platforms, which
An integrated technology model centers on products and platforms, which
differ
differ in
in important
important ways.
ways.

Products Platforms
Purpose Create business value by enhancing end users’ Provide capabilities to products and the enterprise
experiences

Primary users Customers and employees Digital-product developers, along with functional
employees who use platforms directly
Responsible Business-minded teams of tech specialists, Technology-minded teams of digital and IT
personnel designers, product managers, and functional specialists
employees
Pace of Rapid: upgrades happen as quickly as possible to Variable: changes to support products and
innovation keep up with users’ needs modernize underlying systems are made as
priorities dictate

Examples B2C: website/online search B2C: inventory management

B2B: order configuration B2B: pricing

that lets them assemble and price orders during However, it is important that these digital units
customer meetings. and IT departments are closely integrated, with
thoughtful coordination and planning between
Platforms are the back-end technology the organizations to prevent any bottlenecks.
capabilities, whether provided by individual For instance, as digital efforts expand to cover
systems or by assemblies of multiple systems, more customer and employee experiences
that power products, as well as the enterprise and incorporate new technologies, integration
more broadly. The retail search product previously between digital and traditional technology
described, for example, might rely on an inventory solutions requires more extensive collaboration.
platform that includes databases and integrations Differences in culture and ways of working
with suppliers. Typical platforms found at large can make it harder for digital and IT groups to
companies include those for enterprise resource integrate new digital offerings with core systems.
planning, customer relationship management, Teams from other business functions can also
inventory management, and field operations. get confused about which technology groups to
work with—and how.
Many companies launch their digital efforts with a
focus on creating and improving products through An integrated operating model helps resolve
a stand-alone organization that is separate these differences by bringing IT and digital
or siloed from company IT. These separately organizations into a single model for planning,
funded digital units deliver user-experience delivering, and managing technology, reinforced
innovations quickly by employing a mix of design by a shared culture and talent-management
and engineering talent, using cloud technologies, approach. In this model, digital and IT specialists
following agile delivery practices, and, often, work together on unified teams, each centered
fostering a different working culture and norms— on an individual product or platform (Exhibit 2).
an approach unlike that of a traditional IT function.

Products and platforms: Is your technology operating model ready? 32

 Exhibit 2
Integrated
Integratedtechnology
technology organizations
organizationscomprise teams
comprise that support
teams products,
that support products,
platforms, and infrastructure, plus a digital-delivery office.
platforms, and infrastructure, plus a digital-delivery office.

B2C example: customer journey, B2B example: salesperson

partial journey, partial

Product teams Creating digitally Browsing items Searching for offerings

enabled user experiences and
Searching for items Developing price quotes
bringing them out
Checking out orders Placing orders
Platform teams Developing features
to meet needs of product teams and API/service layer API/service layer
modernizing systems Customer Transaction Fulfillment Contracts
Infrastructure teams Increasing Profiles Pricing Inventory Pricing
efficiency, consistency, stability, Promos Product info EDI1 integration T&Cs2
and security
Campaigns Timing Returns Memberships
Digital-delivery office Aligning tech
priorities, managing dependencies, Infrastructure Infrastructure
and allocating funds and staff Digital-delivery office Digital-delivery office

1
Electronic data interchange.
2
Terms and conditions.

Product teams focus on serving the needs of on available inventory, then the inventory-
end users in ways that generate revenue, lift platform team might build an API that allows
productivity, or otherwise directly create value product teams to query inventory data for
for a company. They operate like minibusinesses, in-stock items. A platform team will normally
responsible for go-to-market planning, user adjust its ways of working to match the state
experience, and adoption in addition to of the underlying systems and the needs of
technology delivery. To carry out this approach, product teams and external partners. On
product teams include not only engineers but most platform teams, the leader will have a
also designers, analysts, and experts from other technology background, and staff will mainly
business functions, such as operations, marketing, consist of technology specialists.
and compliance. Typically, they use agile methods
to develop products, iterating rapidly to make An integrated technology organization has two
improvements. Most product teams will have additional components. One is a centralized
a leader who is more business oriented than infrastructure-services team, responsible for
technology oriented. provisioning and managing the underlying
technology infrastructure in ways that make it
Platform teams focus on making an organization’s efficient, easy to use, reliable, and consistent.
core systems accessible, reusable, and By automating activities and promoting
modern so that they better enable products. standard development, operations, and
This collaborative approach sometimes calls engineering practices across product and
for platform experts to join product teams platform teams, the infrastructure-services
temporarily. For example, if the team working team continuously streamlines its own work
on the search product described earlier decides and that of the wider technology organization.
that users should be able to filter their searches

33 Products and platforms: Is your technology operating model ready?

 The other is a digital-delivery office, which Next, executives grouped most of the company’s
performs coordinating functions that we describe technology capabilities into platforms. They
later in the article. asked each platform team to prepare a road map
for making improvements that would support
Of course, the differences between digital products while also modernizing the platform’s
programs and IT operations don’t necessarily technology. (Not all platform teams immediately
resolve themselves when specialists from each began working in this new way. At first, the retailer
group combine to form product and platform focused on those platform teams that provided
teams. Indeed, our research suggests that when the most essential functions to high-priority
companies integrate digital and IT specialists, products.) Finally, executives reorganized the
they often find that cultural differences and talent leadership of the technology organization to
gaps get in the way of effective teamwork.⁵ In our reflect its new structure, in a move we will explore
experience, establishing new operating norms further in the next section. As a result of these
for the integrated IT organization can help align changes, the company increased its capacity
digital and IT employees who have diverging ideas for delivering the technologies that its strategic
about how their teams should work. For example, priorities called for, and it gained the ability to
technology leaders should lay out a common set of reallocate people quickly to urgent tasks.
engineering practices as well as standardized ways
of working based on agile principles and design
thinking. Creating a governance model
Each team in an integrated technology
To address talent gaps, companies often find it organization ordinarily has its own leader.
necessary to reskill existing employees or hire Companies might also appoint a head of products,
additional talent. Product teams, for example, a head of platforms, and a head of infrastructure
handle all aspects of product development, from services. In addition, integrated technology
design to user adoption, so their leaders must be organizations need a central team, or digital-
able to understand users, translate their needs delivery office, to balance the priorities and
into technical requirements, manage product resource requirements of product and platform
road maps, guide engineering teams, and oversee teams and to coordinate any overlapping or
releases. Few organizations employ enough related activities, particularly when it comes to
technology specialists with experience across prioritizing changes to platforms.⁷ Here are some
these disciplines, so they must either train the actions that digital-delivery offices can take to
people they have or bring in new people who govern integrated technology organizations well:
have experience as product managers or product
owners.⁶ — Define value-focused performance metrics.
Defining a few performance metrics for each
At one omnichannel retailer, the executive team team, in partnership with business leaders, will
planned the shift to an integrated technology help a digital-delivery office tell whether teams
operating model by first identifying potential are on track. Such metrics should measure
product teams based on the organization’s digital both value creation (“Is the team contributing to
priorities. The team determined that select the business?”) and the delivery of work (“Is the
personnel from across the company should team meeting its commitments?”). For example,
join these product teams to make them cross- the team managing the search product
functional. It also set out performance indicators to previously described might be assigned a
hold the teams accountable. target for increasing the proportion of online

5
“Managing the fallout from technology transformations,” February 2020, McKinsey.com.
6
Chandra Gnanasambandam, Martin Harrysson, Shivam Srivastava, and Yun Wu, “Product managers for the digital world,” May 2017,
McKinsey.com.
7
Oliver Bossert and Driek Desmet, “The platform play: How to operate like a tech company,” February 2019, McKinsey.com.

Products and platforms: Is your technology operating model ready? 34

 searches that convert to sales, in support of a that too many products had fallen behind
broader strategic goal to lift e-commerce sales. schedule. At the outset, the company’s
executive team set goals that defined what
— Create processes for tracking and reporting. the success of these products would look
Digital-delivery offices should have consistent like, including specific changes to customer
workflows for collecting performance updates experience, geographic coverage, and
from product and platform teams (with value-creation metrics. The digital-delivery
particular interest in whether business-value office tracked the technology organization’s
and other targets are being met and whether progress against these goals and held monthly
any interdependencies are causing delays) and reviews with executive leaders. It also tracked
synthesizing those into reports that can inform leading indicators and surfaced early signs
a leadership team’s strategic decisions. of problems in areas such as software quality
and reliability. To resolve those problems, the
— Establish a mechanism for quickly removing office established a weekly leadership forum
bottlenecks and resolving conflicts. Since and series of meetings to escalate issues
unexpected difficulties (for example, a dispute spotted by engineering teams. As a result of
between product teams that want to bring these efforts, the company accelerated the
in the same platform personnel) can slow development of several products that had
down work on technology projects, some been delayed and launched them sooner than
organizations empower their digital-delivery it had planned.
offices to come up with quick solutions that
minimize delays. Others choose to have their
digital-delivery offices surface conflicts to Establishing a system to prioritize
leadership, which resolves them. and deliver technology work
Even though technology teams and roles
— Gradually adopt an agile funding model. It might be well defined, coordinating their many
is common for technology organizations to activities can be difficult. Leaders should
allocate funds for products and platforms on therefore create a system for maintaining order
an annual—or even less frequent—basis. By in how product and platform teams respond
contrast, an agile funding model ties new to the business’s demand for technology
releases of funding to the achievement of services.
development or performance objectives. This
approach keeps teams focused on delivering One feature of this system should be
value and prevents companies from backing processes by which a technology organization
struggling initiatives merely because funds aligns product and platform road maps with
were already allocated to them.⁸ With respect the business’s priorities. The head of each
to the search product mentioned before, the business function should have a single
team might receive initial funding to create and contact within the tech organization, likely a
test a new function but only receive full funding product manager or user-journey lead, who is
to develop and launch the function if user tests responsible for understanding the function’s
show that the new function increases the priorities and translating them into a set of
likelihood of sales conversions. desired technology features, recorded on
product backlogs. Product managers then
One global consumer company set up a digital- work with their platform-team counterparts
delivery office to facilitate the development of on updating the platforms’ development road
new digital products for customers after noticing maps to incorporate the features that products

8
Santiago Comella-Dorda, Khushpreet Kaur, and Ahmad Zaidi, “Planning in an agile organization,” February 2019, McKinsey.com.

35 Products and platforms: Is your technology operating model ready?

 require. (Functional leaders should also assign coordinating development efforts across teams
their personnel to product teams when necessary can get complicated. To minimize conflicts
to speed development.) The digital-delivery office and surprises, every dependency should be
then helps sort out project, staffing, and other addressed on at least one team’s release plan.
priorities across multiple teams, consulting with
the organization’s leadership when conflicts have The digital-delivery office should also make sure
strategic implications (Exhibit 3). that product teams depend on platform teams
but that platform teams never depend on product
The digital-delivery office also helps manage teams. The idea is to prevent any one product
the dependencies among teams, particularly team from restricting the work of a platform
product teams’ dependencies on platform teams. team because that can interfere with what other
Product and platform teams can do some of this product teams want to do. The one-way nature of
on their own. Earlier in the article, we noted that a dependencies between products and platforms
retailer’s online search product might depend on might require product teams to postpone the
an underlying inventory platform. In such a case, release of certain features while platforms are
the search-product team and inventory-platform being updated and to fast-track other features in
team would ideally work together to ensure that the meantime.
changes to the inventory platform are in sync
with the release plan for the search product. But The digital-delivery office helps manage
because each platform supports multiple products, dependencies by reviewing each team’s road
and each product depends on multiple platforms, map to identify the impacts of any planned

Product
Exhibit 3 and platform teams and the digital-delivery office handle and
prioritize
Product technology
and requests
platform teams from
and the business functions.
digital-delivery office handle and prioritize
technology requests from business functions.
Example system for managing technology demands, illustrative

3 Digital-delivery office

Business Product Cross-functional Platform Tech-led

functions leaders product teams leaders platform teams

1 2 4

Infrastructure services

1 Leaders generate 2 Product leaders capture 3 The digital-delivery office 4 Platform teams build
technology demand from demand in their road prioritizes work, allocates modular, reusable
business functions and maps, own the execution, resources, tracks activity technology capabilities
units or from the and liaise with platform and value creation, and that support end-user
company’s strategy teams; a business leader provides a unified view demand
should have one product of what is being
leader as a main tech delivered and when
contact

Products and platforms: Is your technology operating model ready? 36

 changes. It also holds regular meetings with Platform teams should also reserve capacity
product and platform leaders to go over the to ensure that any changes are exposed in
value and feasibility of the features they want to as-a-service offerings for all product teams
add, identify new dependencies, and negotiate to use (through an API endpoint, for example).
teams’ competing interests (such as product This approach not only ensures that changes
teams’ needs to borrow the same personnel from to platforms support multiple products without
platform teams and product teams’ simultaneous adding duplication or complexity but also
requests for additional platform features). Some allows product teams to build new capabilities
digital-delivery offices establish a common without always engaging platform teams.
release cadence so that they can more easily
coordinate assignments of resources. Other
practices, such as feature flagging, can prevent
dependencies from delaying releases. Increasingly, companies are choosing an IT
operating model that integrates customer-
Last, integrated technology organizations should and employee-facing digital efforts with the
look for and seize opportunities to modernize activities of conventional IT departments and
IT products and platforms as they are working promotes uniform approaches to technology
on requests from the business for new features work. By forming integrated teams to support
and solutions. For platform teams, especially, products and platforms, instituting consistent
modernizing core systems in the course of their governance processes, and establishing
normal work ensures that the platforms remain a common system for prioritizing work, a
capable of meeting increasingly sophisticated technology organization can increase the value
technical requirements from product teams. it creates for a business.

Ross Frazier and Amit Rahul are associate partners in McKinsey’s Chicago office, where Naufal Khan is a senior partner;
Gautam Lunawat is a partner in the Silicon Valley office.

Copyright © 2020 McKinsey & Company. All rights reserved.

37 Products and platforms: Is your technology operating model ready?

 3. Reinventing
technology delivery
 How companies can win
in the seven tech-talent
battlegrounds
Companies have to hire the best, but that won’t be enough. They’ll
also need to rethink how they reskill and upskill their people.

by Matthias Daub, Ranja Reda Kouba, Kate Smaje, and Anna Wiesinger

© Getty Images

October 2020

39
 With the acceleration in digital, the demands upskilling will be at least half of the solution to their
on technology—for speed, flexibility, reliability, persistent skill gaps.²
security, and value—have radically increased.
For CIOs surveying how to transform their
organizations, one tricky question is emerging: Seven emerging tech-talent
Where do I find the people to do all the work? battlegrounds
To better understand what tech talent will matter
Few executives would debate the importance most in the next three to five years, we spoke
of talent or the difficulty that many have in with hundreds of global CIOs, analyzed talent
attracting and keeping top people. But companies developments over two years across three global
nevertheless aren’t treating tech talent with the markets, and reviewed more than 30 cross-cutting
urgency it demands. Respondents to a recent tech trends. We then mapped relevant skills
McKinsey survey report more significant impact and roles to the most significant emerging tech
from talent transformations than from any other trends and business needs. For example, given
technology-based play. Yet talent transformations the increasing importance of using data to make
are relatively rare. Only 27 percent say their better and faster decisions, the ability to rapidly
companies have pursued one in the past two years, build infrastructure and architecture for data
and just 15 percent believe they will do so in the (data-engineer skills) is likely to become more of
next two years.¹ a bottleneck than the ability to generate insights
(data-scientist skills).
Amidst this reality, the increasing complexity of
IT systems and the emergence of a broad range Through this analysis, we identified about 4,000
of new technologies, from cloud to artificial tech skills, which we broke down into seven
intelligence (AI) to machine learning, have battlegrounds, or clusters of need (see chart on
increased the challenges. One European CEO and the next page). (Note: while cultural and change-
football fan explained it this way: if you gave him management aspects, including social and
a big enough budget, he’d be confident he could emotional skills, are also important, our research
put together a winning team. But a cricket team? honed in on tech skills only).
He wouldn’t know where to start, since he doesn’t
know anything about the game. He used the Significant skills gaps in these seven areas already
analogy to point out how hard it can be for leaders exist, and we expect them to become more severe
to know what talent they actually need. over time. Executives expect skills mismatches
in functions that have already started adopting
A few companies, however, have started to automation and AI technologies, according to
crack the code. Companies winning in this arena McKinsey Global Institute analysis.³ The largest
have identified at a granular level the tech skills percentage of survey respondents (more than 30
they need to build value for the business, have percent) ranked data analytics, IT, mobile, and web
developed a clear view of their present and future design as the skills with the highest expectation of
talent needs, and are intentional about finding a mismatch over the next three years.
both top talent and adaptable learners. Crucially,
these leaders understand that it’s impossible to In Germany, 700,000 additional tech specialists
hire everyone you need; training and reskilling are needed by 2023 to meet the economy’s
the existing workforce has to be a core part of the demand for them.⁴ For agile skills, demand will be
strategy to win the talent battle. Some 82 percent four times greater than supply, and for big data
of global executives expect that reskilling and talent, 50 to 60 percent greater.⁵ Globally, 3.5

1
McKinsey Global Survey on IT and the Business, August 2020.
2
For more, see “Retraining and reskilling workers in the age of automation,” McKinsey Global Institute, January 2018, on McKinsey.com.
3
For the full McKinsey Global Institute report, see “Skill shift: Automation and the future of the workforce,” May 2018, on McKinsey.com.
4
Julian Kirchherr, Julia Klier, Cornels Lehmann-Brauns, and Mathias Winde, “Future skills: Which skills are lacking in Germany,” Stifterverband
and McKinsey & Company, September 2018, stifterverband.com.
5
Satty Bhens, Ling Lau, and Hugo Sarrazin, “The new tech talent you need to succeed in digital,” September 2016, McKinsey.com.

40 How companies can win in the seven tech-talent battlegrounds

Battleground Rationale Tech skills (sample set)

DevOps Faster and continuous delivery of features, more stable • Agile product-life-cycle
environments, and reduced operations time. (For more, management
read “Agile, reliable, secure, compliant IT: Fulfilling the • DevSecOps
promise of DevSecOps,” on McKinsey.com.) • Continuous integration and
delivery (CI/CD)
• Microservices architecture
Customer Significant shifts in customer behavior as a result of • Predictive/nudge analytics
experience COVID-19 and rising customer expectations; need to • Design thinking
deliver top experiences across a wide array of channels; • Test-and-learn at scale
prioritization of personalized over generic design • Automated testing
(while maintaining privacy); continuous test-and-learn • Prototyping
cycles. (For more, read “Elevating customer experience
excellence in the next normal,” on McKinsey.com.)
Cloud Infrastructure increasingly provided through next-gen • Kubernetes
cloud architecture, the time to market of services is • Docker
vastly improved, solutions are more easily scalable; • Multicloud and hybrid-cloud
acceleration of transformation and increased source of architecture
competitive value. (For more, read “Capturing value in the • Security
cloud,” on McKinsey.com.) • Smart distribution/metering
• Edge computing
Automation Significant number of tasks automatable: about 22 • Cognitive AI
percent of workforce activities across the European • RPA technologies
Union could be automated by 2030,⁶ for example, • Automation anywhere
through end-to-end automation across development, • Machine learning
testing, and deployment processes—accelerating • AI-enabled analytics
development and reducing errors. (For more, read “The • Quantum computing
imperatives for automation success,” on McKinsey.com.)
Platforms and Platform-as-a-service (PaaS) operating model provides • Life-cycle management across
products foundation for development with reusable code; platform layers
“building-block” product approach to development • Industrial Internet of Things
speeds up releases and makes process more flexible. (IIoT)
(For more, read “The platform play: How to operate like a • Vertical software as a service
tech company,” on McKinsey.com.) (SaaS)
Data management Need for real-time data-driven insights, data • Use-case life-cycle
democratization (nonexpert users making advanced management
data queries), acceleration of both data quantity • Synthetic data
and variability. (For more, read “How to build a data • Data governance
architecture to drive innovation—today and tomorrow,” • Automated machine learning
on McKinsey.com.)
Cybersecurity and Data breaches are increasing while data-privacy • Shift-left security
privacy concerns are resulting in varied regulatory changes, • Automated testing
forcing companies to rethink security and compliance • Zero-trust security
protocols. (For more, read “A dual cybersecurity mindset • Data-protection law and
for the next normal,” on McKinsey.com.) practices

6
For more, see “The future of work in Europe,” McKinsey Global Institute, June 2020, on McKinsey.com.

How companies can win in the seven tech-talent battlegrounds 41

 million cybersecurity positions are projected to be IT organizations are built around small cadres
unfilled in 2021.⁷ of high-performance people working in highly
self-motivated, self-managing, and agile
In addition to meeting the challenges of filling teams.
future roles, technology modernization requires
knowledge of how to transition from existing Finding these anchor hires and being
systems, which are often written in outdated prepared to pay more for them is more
programming languages, such as LISP, ALGOL 58, cost-effective in the long run—and greatly
or COBOL, and are understood mostly by an aging helps in recruiting additional people who
workforce.⁸ want to work with the best. Spotting
quality talent is notoriously difficult,
however, because companies are often
Closing the talent gap unclear about how to evaluate their talent
To succeed in the seven tech-talent battlegrounds, effectively. Top companies, however, identify
companies will need to use a set of well-considered top performance through a hierarchy of
strategies: hiring, reskilling (training employees for observable behaviors.
new roles), upskilling (training within an existing
role), reallocating, and sourcing. Which strategies to — Finding adaptable learners. Tech talent has
pursue depend on a company’s starting point and always been accustomed to lifelong learning
specific needs (see sidebar “Four archetypes for as their fields change and new ones emerge.
addressing talent gaps”). For this article, we focus Technology skills evolve so quickly that
on hiring, reskilling, and upskilling. focusing solely on credentials and specific
skills when hiring is not enough. In addition
The first step in closing the skills gap is rigorous to specialized talent, the best companies
discipline in identifying specific talent needs. In a look for “strong talent,” which has the ability
McKinsey survey, nearly twice as many respondents to learn and adapt. As one executive said,
who report successful transformations say their “We’re not looking for people with skills; we’re
companies set hiring goals based on specific looking for people who can learn skills.”
skills needs, compared with respondents whose
organizations don’t set those same kind of goals.⁹ The flip side of this coin is nurturing an
They do so by evaluating relevant trends, identifying environment for learning. In a survey of IT
the corresponding skills needed over the next three experts, the majority of respondents said
to five years. Importantly, they identify skills at a they regard employee training as a crucial
level of precision necessary so they can target the driver of career success, even more important
right hires and build out relevant training programs. than IT certifications.10 But beyond formal
training, the best companies explicitly provide
Hiring practices that work their people with time to learn, budgets to
We have found the following hiring practices to be fund experimentation, and access to new
most effective: technologies, as well as flexible career paths
that provide additional learning opportunities.
— Favoring quality over quantity. Given the scale of
the need, organizations tend to focus on quantity. — “Techies for techies” recruiting. To hire good
However, they should favor quality even more. tech talent, you need to involve your top
A single expert or highly skilled engineer is as talent in the recruiting process. The reality
productive as eight novices. The most effective is that techies want to talk to techies rather

7
Steve Morgan,“Cybersecurity talent crunch to create 3.5 million unfilled jobs globally by 2021,” Cybersecurity Ventures, October 24, 2019,
cybersecurityventures.com.
8
Patrick McGeehan, “He needs jobless benefits. He was told to find a fax machine,” New York Times, April 4, 2020, nytimes.com.
9
“Unlocking success in digital transformations,” October 2018, McKinsey.com.
10
“What IT pros think about IT training,” LinkedIn, January 2017, learning.linkedin.com.

42 How companies can win in the seven tech-talent battlegrounds

Four archetypes for addressing talent gaps

Four organizational archetypes determine how companies address their tech-talent challenges (exhibit).

The traditionalist: Upskilling

Extensive skills gaps paired with an inability to attract top talent predisposes this archetype to focus on upskilling and reskilling existing
employees. Digital learning platforms can help to make training scalable, applicable across locations, and also feasible during COVID-19
restrictions.

The digitizing incumbent: Skilling–hiring mix

This archetype in general still has a large tech-skills gap, especially in quantity of skills, with a slightly smaller gap in quality of skills. In
addition to reskilling employees, the focus is on hiring new tech talent, though that can prove to be a challenge.

The emerging digitalist: Redeploying and hiring

The emerging digitalist is prone to focusing on hiring to address a moderate—though widening (due to business growth)—skills gap. In this
case, redeploying talent to the most value-generating needs can be particularly effective. Our research reveals that, on average, leading
companies reallocate digital talent more than five times faster than their peers.

The digital native: Continuous reskilling and hiring

Leadership at digital-native companies is typically aware of technology’s critical role and the need to stay abreast of the competition. If
a skills gap arises at all, it is likely to happen because a tech firm undergoes yet another IT paradigm shift, from mobile-first to AI-first, for
example. They are then aggressive both in hiring the necessary talent and in reskilling relevant people.

Exhibit
IT organizations
IT organizations typically
typically fallfall
intointo
oneone of four
of four archetypes,
archetypes, eacheach
withwith varying
varying approaches
approaches
to closing to closing skills gaps.
skills gaps.

1 The 2 The digitizing 3 The emerging 4 The digital

traditionalist incumbent digitalist native

Characteristics IT assumes a support IT assumes an IT assumes a central IT assumes a strategic

of the IT function; outdated enabling function and function in the setup function, and an
organization working model and is understood to be of a new digital IT-first mindset
tech stack, tenured important for reaping business within an permeates all teams;
staff, and lack of efficiency gains or existing incumbent state-of-the-art tech
belief in the power of discovering new and is expected to stack, adaptable top
technology impede frontiers in an unlock new value talent, and tech-
reorientation incumbent‘s core pools in addition to forward IT operations
business those in the core model act as enablers

Sample Public-sector Large bank decides Automotive incumbent Leading tech firm
organization institution aims to to digitally transform builds digital attacker undergoes yet another
digitize processes its core business (new digital business IT paradigm shift (eg,
from the ground up in addition to core from mobile-first to
business) AI-first)

Strategies to Extensive qual and Large quant gap Medium quant gap Modest qual gap
close the gap quant gaps paired paired with openness paired with overall paired with overall
with limited ability to to revamp employee business growth business growth leads
revamp employee base leads to skilling leads to redeploying to continuous reskilling
base leads to and hiring focus and hiring focus and hiring top talent
upskilling focus

Upskilling

Reskilling

Hiring

How companies can win in the seven tech-talent battlegrounds 43

 than to HR people with limited tech knowledge. away from broad learning programs to targeted
Acquiring top talent also requires the use of learning journeys that focus on top-priority areas
a broad set of recruiting channels, such as for the business. In addition, the courses can be
developer conferences and hackathons; an short. Tech-learning providers offer introductory
open mind about educational qualifications and courses that take only a few hours or degree
an awareness that 85 percent of developers are programs that can be completed within three to
at least partially self-taught; aspirational goals six months, with less than 15 hours of learning
that inspire; and a demonstrated commitment to effort per week.15
building a diverse and inclusive workforce.11
Build learning journeys
— Moving quickly. Job seekers in the tech A learning journey is a set of connected learning
world are impatient. Applicants often have experiences that drive sustained performance
various offers and are used to rapid recruiting improvements (exhibit). Learning journeys have
processes: 57 percent of job seekers are been highly effective in closing skills gaps, as they
unhappy with the waiting time after an interview, blend a variety of different training formats, such
while 23 percent are willing to wait only one as digital, cohort-based, or on-the job learning.
week to hear back.12
COVID-19 has accelerated the full digitization of
all learning-journey components. These dynamics
Practical guide for reskilling and not only make it possible to scale learning efforts
upskilling more cost effectively but also offer greater
According to the World Economic Forum, around personalization for learners.16
54 percent of all employees will need reskilling and
upskilling by 2022. Of these, 35 percent will require For example, a leading US insurer identified 15
up to six months of training, 9 percent will need six to 20 critical talent pools among its more than
to 12 months, and 10 percent more than a year.13 The 17,000-strong workforce, to determine the
best programs will focus on the following practices. potential of displaced individuals to be reskilled
and redeployed. The insurer designed learning
Use budget strategically journeys to upskill and reskill current roles to the
Reskilling is cheaper than hiring. While reskilling an roles of the future, such as the business translator.
internal employee may cost $20,000 or less, the This learning-journey approach made it possible
cost of hiring often costs $30,000 for recruitment to reskill or redeploy 40 percent of the overall
alone, in addition to onboarding training. And new workforce.
hires are two to three times more likely to then
leave.14 Large tech players understand this and In another example, a European regional bank
often opt to invest more significantly in reskilling linked its learning journeys to concrete new
their workforce. career paths. In addition to learning-journey-
based reskilling, almost all of the more than
Effective reskilling and upskilling, however, don’t 30,000 employees used mobile-app digital
require large outlays. By using existing training learning courses to build skills identified as
budgets more strategically, companies can move important for the company’s future.

11
Tomas Chamorro-Premuzic and Jonathan Kirschner, “How the best managers identify and develop talent,” Harvard Business Review, January 9,
2020, hbr.org.
12
“Are you taking too long to hire?,” Robert Half, 2016, roberthalf.com.
13
The future of jobs report 2018, World Economic Forum, September 2018, weforum.org.
14
Josh Bersin, Rethinking the build vs. buy approach to talent, General Assembly and Whiteboard Advisors, October 2019, joshbersin.com.
15
Course overview of online learning providers Udacity and Coursera, 2020.
16
Sapana Agrawal, Aaron De Smet, Sébastien Lacroix, and Angelika Reich, “To emerge stronger from the COVID-19 crisis, companies should
start reskilling their workforces now,” May 2020, McKinsey.com.

44 How companies can win in the seven tech-talent battlegrounds

Exhibit
Individual learning
Individual learningjourneys
journeysare
aretailored
tailoredto
tospecific
specific skills
skillsneeds
needs over
over a range of
formats.
range of formats.
Tech company example

Digital Cohort-based learning On-the-job learning

Reskilling journey
(~6 months)

Digital
platform
Apply skills owner
Digital on the job
Peer • Practitioner
Boot learning
Apply skills coaching
camp with technical
Boot on the job
Pre- expertise
Technical camp
project learning • Integrator of
manager internal and
external
resources
• Trusted advisor
Has perspective Translates product Possesses solid Understands for executives
Targets on core topics road maps into theoretical dynamics of
• Agile problem
Knows personal platform or eco- knowledge platform
system strategies ownership solver and driver
development path Develops capabilities of innovation
First practical by peer coaching Drives
perspective innovation

Learning needs to be a top management priority relegated to being an “IT thing.” While people on the
The CHRO and CIO need to take joint ownership business side don’t need to know how to code, they
of a business’s tech-training program. The most do need to learn how to better use technology. The
effective partnerships make sure that their continued democratization of data can also allow
training investments align with the company’s for “laypeople” to use data to make better and faster
overall strategy, establish a governance model decisions without relying on complex IT processes.
with shared ownership among business leaders,
continuously assess skills gaps, design targeted Furthermore, CIOs often assume that only IT people
learning journeys, and integrate them into HR can be reskilled and typically underestimate the
processes.17 possibility of reskilling employees from nontech
departments. But increasing evidence shows
Don’t forget your nontech employees that reskilling nontech people for tech roles can
Nontech people need tech skills, too. With the be effective (see sidebar “Even people without
continued importance of technology in driving ‘adjacent’ skills can be successfully reskilled”).18
business value, technology can no longer be

17
Jacqueline Brassey, Lisa Christensen, and Nick van Dam, “The essential components of a successful L&D strategy,” February 2019, McKinsey.com.
18
Coursera blog, “Learned code and switched careers as a developer,” February 24, 2017, blog.coursera.org.

How companies can win in the seven tech-talent battlegrounds 45

 Even people without ‘adjacent’ skills can be successfully reskilled

Skills adjacency is defined as the proximity between the skills required for two different jobs. Among students at Udacity,¹ a for-profit edu-
cational organization offering online technology courses, 67 percent showed high skills adjacency between their previous job and the one
they found after completing their courses. Interestingly, however, a significant 33 percent found a new job with only medium or low skills
adjacency, indicating that reskilling someone from a nontech role to a tech role can succeed (exhibit).²

Reskilling can be successful even when skills adjacency is low.

Exhibit
Reskilling can be successful even when skills adjacency is low.
Examples

Type and length of reskilling

Front-end web developer

Driver Software
4 months; 5–10 hours learning/week engineer

Mortgage-loan Digital-marketing nanodegree Digital

processor marketer
3 months; 10 hours learning/week

Android-developer nanodegree
Hospitality Android
professional 6 months; 10 hours learning/week developer

1
McKinsey has a nonexclusive partnership with Udacity.
2
Udacity data analysis, nonenterprise, private customers, n = 463, August 2020.

are willing to dedicate the energy, focus, and

resources to continually closing—or, in some cases,
Given the rapidly changing nature of business
even leapfrogging—those gaps can win in the most
and technology, companies will always be facing
important talent battlegrounds.
technology-skills gaps. But organizations that

Matthias Daub is a senior partner in McKinsey’s Berlin office, Ranja Reda Kouba is an associate partner in the Vienna office,
Kate Smaje is a senior partner in the London office, and Anna Wiesinger is a partner in the Düsseldorf office.

The authors wish to thank the team of Jutta Bodem-Schrötgens, Florent Erbar, Teresa Keller, Anna Lena Robra, Hannah Mayer,
Eileen Raßlenberg, Michael Scherbela, Surbhi Sikka, and Thaksan Sothinathan for their ongoing support and drive.

The authors also wish to thank Sapana Agrawal, Kerstin Balka, Sven Blumberg, Andrea Del Miglio, Anusha Dhasarathy, Vito Di
Leo, Amadeo Di Lodovico, Karel Dörner, Desiree El Chebeir, Peter Jacobs, Shweta Juneja, Naufal Khan, Harald Kube, Mahir
Nayfeh, Angelika Reich, Wolf Richter, Scott Rutherford, Henning Soller, Gisa Springer, Richard Steele, and Steve van Kuiken for
their contributions to this article.

Copyright © 2020 McKinsey & Company. All rights reserved.

46 How companies can win in the seven tech-talent battlegrounds

 CIOs are redefining what a
successful relationship with
their IT providers looks like
As CIOs lead their IT organizations through transformations,
long-established relationships with IT providers are set to change.

by Abhi Bhatnagar, Anuj Kadyan, Wasim Lala, and Abdallah Salami

© Getty Images

April 2020
47
 Note that these findings are based on a survey expectations, and players shaping this role will
done before the current coronavirus crisis, and differ significantly from those of the past.
depending on its impact, any sourcing strategy
going forward could look different.
Looking for help disentangling from
In the digital world, technology is no longer an legacy systems
enabler but a strategic asset and a competitive CIOs are eager to accelerate their businesses’
advantage. CIOs are at the helm of this digital digital transformations. After years of allocating
transformation journey and are under increasing external IT spend to legacy systems, the leaders
pressure to deliver the technology capabilities to we surveyed signaled an ambitious desire
enable businesses to generate value. to allocate more than half of their external IT
budget to next-generation services within three
In this context, significant questions are surfacing to five years (Exhibit 1). However, CIOs seeking
about the role of IT providers, which have been rapid progress are often thwarted by their
a mainstay of the technology landscape for the organization’s entanglement in the complexities
past two decades. Are IT providers hindering an of legacy IT. In fact, 87 percent of leaders cited
organization’s ability to innovate? Do they deliver the “complexity of existing infrastructure” as
on promised cost savings, and if they do, are the a key impediment to implementing next-
savings worth the effort? Are they driving strategic generation services. As a result, CIOs are
outcomes, or are they serving legacy goals? These looking for their IT providers’ help in simplifying
Article type and Year
questions are increasingly pertinent, as CIOs must and streamlining the legacy environment. This
both ensure that their organizations can keep up Article Title
will enable digital transformation by freeing
with innovation while maintaining established Exhibit X of X
up resources and funding that are currently
systems. engaged in keeping the lights on.

To address this challenge, many CIOs are choosing

to build up internal capabilities in an effort to Exhibit 1
increase speed, flexibility, and control over critical
technology, often in the context of tech-enabled How will you allocate your external IT
transformations. At the same time, this challenge
budget in three to five years?
is leading CIOs to redefine how they engage with,
and what they expect from, their IT providers. As
one CIO put it: “Given the shortage of capable
talent internally, our resources are focused on
working with IT providers to define the problem
and then partnering with them to execute.” 42%

This evolving relationship with IT providers comes 58%

through in our survey of 250 global CIOs and
similar technology decision makers.¹ More than
Legacy
half of IT leaders believe that “there is no other
(infrastructure,
way” to achieve their digital-transformation goals application
without a close relationship with their IT providers. development
Our survey and interviews point to an active role Next-generation
& maintenance)
for IT providers along the digital transformation (big data, automation,
journey of many companies. However, the focus, cyber)

¹ The survey was conducted in the summer of 2019 and covered 250 CIOs and similar decision makers in companies with IT budgets greater than
$250 million across more than ten industries globally.

48 CIOs are redefining what a successful relationship with their IT providers looks like
 Article type and Year
Article Title
Exhibit X of X

Talent and innovation are a must Exhibit 3

While cost is a major reason why most enterprises
were and are still working with IT providers, other
Do your external providers meet your
factors are now gaining importance (Exhibit 2). expectations?
Leaders
Article cite
typeaccess
and to talent and innovation as
Year IT services providers report card
crucial
Article drivers
Titleinfluencing their decision to engage
and
Exhibit X ofwith
continue X IT providers.
Executing digital 4.2/5.0
transformation
Exhibit 2 Designing digital 4.1/5.0
transformation
What is driving your need to leverage
external IT providers? Delivering cost 4.0/5.0
savings
% of respondents
Driving 3.9/5.0
innovation
28
27
Driving business 3.8/5.0
outcomes
20
18 Bringing right 3.8/5.0
talent

8
A B C D E seeking to engage with IT providers that strengthen
their internal talent’s ability to innovate and drive
the business’s digital strategy. According to one CIO,
A: Cost D: Vendor demanded “We are no longer impressed with a ‘hand the keys
B: Access to talent by business over and let them drive’ model.”
E: Service not core
C: Innovation
to business
Providers, like CIOs, have to serve the
entire business
This is also reflected in how CIOs define “good” IT When it comes to selecting IT providers, roughly 60
providers. Previously, IT providers were deemed percent of the CIOs surveyed believe that business-
to be good if they fulfilled service requirements unit leaders will be either the sole decision makers
within targeted cost targets. Today, while CIOs are or joint decision makers with the CIO (Exhibit 4).
generally satisfied with external providers’ delivery
on cost savings (score 4.0 on a 5.0 satisfaction For this reason, CIOs are shifting their role to
scale), they are less satisfied with and demanding become true working partners to the business and
more from providers in terms of innovation (score focusing increasingly on how technology can drive
3.9 out of 5.0), delivering business outcomes (3.8), business outcomes. While doing so, they expect
and bringing the right talent (3.8) (Exhibit 3). IT providers—who have traditionally primarily
interfaced with the CIO—to do the same. It is not
In addition, there is a pronounced desire among surprising, then, to find that some 60 percent
CIOs to sustain their ability to innovate, especially of CIOs consider the “ability to engage with the
in business-critical and customer-facing areas. As business” to be a key criterion influencing the
a result, CIOs—particularly those with IT budgets selection of IT providers.
greater than $500 million—are increasingly

CIOs are redefining what a successful relationship with their IT providers looks like 49
 Article type and Year
Article Title
Exhibit X of X

Exhibit 4 IT providers have a good starting point. When we

asked CIOs about where IT providers deliver value,
Who leads decision making on designing and executing digital transformations
selection of IT providers? came out on top (above 4.0 on a 5.0 satisfaction
scale). To succeed going forward, IT providers need
% of respondents to ensure these transformations align with and
deliver business outcomes.

With greater choice, organizations are

Business leads rethinking their provider portfolio
CIO leads 26 Traditional players, such as large systems
integrators (SIs), have enjoyed market dominance
41
over other IT providers for a while. However, as CIOs
look for more specialized talent and capabilities,
niche providers, in particular, have the potential to
Joint decision
plug gaps in talent and innovation where traditional
33 large-scale providers fall short (Exhibit 5). Said one
of the surveyed CIOs: “For analytics, traditional
SIs play the role of flex capacity; if we need data
scientists, we go to specialists.”

Web 2019
CIOs are redefining what a successful relationship with their IT providers looks like
Exhibit 5
Exhibit 5 of 5

Who
Who is your
is your preferred
preferred provider?
provider?

Consulting firms Specialist/niche providers Systems integrators

Strategy Design Implementation

Organization’s IT budget <$500M >$500M <$500M >$500M <$500M >$500M

Public-cloud
migration

Private
cloud

Next-generation
application
development &
maintenance

Big data &

analytics

Cybersecurity

Automation

50 CIOs are redefining what a successful relationship with their IT providers looks like
Specialist/niche providers are making significant The relationship between IT organizations and
inroads and establishing themselves as serious service providers is profoundly changing. Leaders
contenders. While working with niche providers are looking to IT providers to bring to the table
requires more trial and error than working with talent and innovation while continuing to deliver on
traditional SIs, CIOs said the outcomes often cost. More than ever, leaders have at their disposal
justify it. CIOs would do well, therefore, to seriously a diverse bench of partners, including niche and
consider the new players on the chessboard as specialist providers, that are hungry to distinguish
they redesign their partnerships portfolio for the themselves from their peers.
long game.

Abhi Bhatnagar is a partner in McKinsey’s Atlanta office, Anuj Kadyan is a partner in the Gurgaon office, and Wasim Lala
and Abdallah Salami are associate partners in the New Jersey office.

Copyright © 2020 McKinsey & Company. All rights reserved.

CIOs are redefining what a successful relationship with their IT providers looks like 51
 Unlocking value: Four
lessons in cloud sourcing
and consumption
Companies that are successful in sourcing and managing the consumption
of cloud adopt a more dynamic, analytical, and demand-driven mindset.

by Abhi Bhatnagar, Will Forrest, Naufal Khan, and Abdallah Salami

© Getty Images

November 2020

52
 Cloud adoption is no longer a question of “if” but In our experience, a major driver of value capture
of “how fast” and “to what extent.” Between 2015 is transforming the approach to sourcing and
and 2020, the revenue of the big-three public consuming cloud. Enterprises that approach this
cloud providers (AWS, Microsoft Azure, and task with a traditional sourcing and infrastructure-
Google Cloud Platform) has quintupled, and they consumption mindset are likely to be surprised
have more than tripled their capital-expenditures by the bill. The flexibility to consume cloud as
investment to meet increasing demand. And needed and cost effectively places responsibility
enterprises are ever more open to cloud platforms: on enterprises to maintain a real-time view of their
more than 90 percent of enterprises reported needs and continuously make deliberate decisions
using cloud technology in some way.¹ on how best to adjust consumption.

These trends reflect a world where enterprises Here are four ways enterprises can derive value
increasingly “consume” infrastructure rather than from cloud by transforming their sourcing and
own it. The benefits of this model are plentiful. consumption approaches.
Cloud adopters are attracted by the promise of
flexible infrastructure capacity, rapid capacity Lesson 1: Sourcing and managing consumption
deployment, and faster time to market for digital of cloud is a dynamic exercise
products. The COVID-19 crisis has accentuated the Over the years, enterprises developed a robust
need for speed and agility, making these benefits model for sourcing IT infrastructure assets. It is
even more important. From an infrastructure- episodic in nature based on asset refresh cycles
economics perspective, perhaps the most and follows a structured sequence: requirements
attractive innovation of cloud is the ability to tailor to request for proposal (RFP) to negotiations
the consumption of infrastructure to the needs of to award. Success in this model requires solid
the organization. This promises greater economic negotiation and contracting skills and the ability
flexibility by transforming underutilized capital to engage the business at the right touchpoints in
expenditures into optimally allocated operations the process. The RFP juncture came to constitute
expenditures. the major point at which value was captured. Once
the contract was signed, the organizational focus
While this concept is attractive in theory, many normally shifted to other areas until the next
enterprises are facing challenges in capturing negotiation cycle.
the value in reality. Enterprises estimate that
around 30 percent of their cloud spend is wasted. Cloud economics mandates a fundamentally
Furthermore, around 80 percent of enterprises different approach. While cloud service provider
consider managing cloud spend a challenge. Thus, (CSP) selection and negotiation are critical
even though more than 70 percent of enterprises components of the cloud journey—determining,
cite optimizing cloud spend as a major goal, for example, the price of services and discount
realizing value remains elusive.² levels—many of the decisions impacting value

1
2020 Flexera state of the cloud report, Flexera, April 2020, flexera.com.
2
Ibid.

Capturing value in the cloud

This article is part of a featured series of articles on “Capturing value in the cloud” at cloud.mckinsey.com, in which we explore how cloud is
revolutionizing the way businesses can create and capture value.

53 Unlocking value: Four lessons in cloud sourcing and consumption

capture come afterwards. The very flexibility Should the enterprise preselect instance types to
that cloud provides means that enterprises be used by teams or leave the decision to the teams
must continuously make dynamic consumption based on the use case? Which instance regions
decisions about which services and specifications should be selected? For example, does the cost-
are needed when and for how long. Each of these benefit ratio justify provisioning instances closer
decisions can have significant cost implications to the customer in order to minimize latency? How
if not deliberately managed. One manufacturing long is the capacity needed, and if the duration
company we know was able to leverage its is predictable, should the organization purchase
traditional procurement muscle to negotiate reserved capacity rather than on-demand, since
competitive discounts from its CSPs, only to reserved instances can be up to 60 percent
be surprised by the high cloud-consumption cheaper? And finally, how should the enterprise
projections—up to twice its spend commitment—a dynamically adjust these choices as it rolls out new
year into cloud adoption. This prompted the products and features or expands into new markets
company to consider renegotiations with its CSPs and geographies?
and to accelerate the shift in its internal approach
to cloud to a more demand-focused model. Given these variables, a deep understanding of
an enterprise’s demand is critical across the cloud
The need to continuously manage cloud journey. During the CSP selection and negotiation
consumption is accentuated by the rapidly phases, enterprises equipped with a proper
evolving vendor marketplace and its continuous understanding of the level and variability of their
introduction of new offerings, features, pricing future demand will be able to better negotiate
mechanisms, and regions. For instance, AWS discounts and make calculated decisions on
has changed prices—mostly dropping them— spend commitments, if any. Following that, on a
more than 60 times since its launch in 2006. It continuous basis, enterprises that capture value are
introduced more than 20 new top-level services ones that take a “consumption approach” to cloud,
last year alone. Sourcing and managing the continuously matching their demand to the best-
consumption of cloud in this world requires a fitting cloud services and pricing arrangements.
deep understanding of the cloud ecosystem and One technology company we know launched a
continuous engagement with the business as continuous consumption analysis focused on
partners. application-level assessment and analytical
projection of demand. It was able to harmonize
Lesson 2: Cloud economics is a demand rather the number of instance configurations for related
than a supply game workloads from more than 20 down to three and
With server and storage assets essentially being then, leveraging the analytical projection, utilized
commodities, enterprises purchasing traditional reserved-instance pricing arrangements for the
infrastructure optimized around two variables: relatively predictable portion of demand.
price and quantity. The latter is less flexible, as it is
mandated by the number of assets that need to be Lesson 3: Granular visibility and forecasting are
refreshed and by fluctuations in peak and average needed to optimize consumption of cloud
demand. This has encouraged enterprises to focus While visibility into and forecasting of spend
on supply-side solutions, such as consolidating are critical to any procurement category, they
volume, standardizing SKUs, and structuring are particularly important to cloud given it is
favorable contract terms. a continuously sourced (“consumed”) service.
Capturing value from cloud requires a clear
In a cloud world, enterprises have to solve for more understanding of actual usage costs in order
numerous, interconnected, and demand-focused to stem any value leakage from excessive or
variables. Take compute as an example: Which miscalibrated consumption. However, enterprises
instance types, of the dozens offered, deliver the often find themselves mired in an intractable sprawl
right balance between performance and cost?

Unlocking value: Four lessons in cloud sourcing and consumption 54

 of cloud services with inadequate visibility into the spend, often assisted by internal analytics or
corresponding spend. The large and growing range third-party tools
of cloud offerings and pricing arrangements in the
marketplace—as well as often obsolete managerial — deriving the unit-cost economics according
processes—do not make this problem easier. to the hierarchy of business and technical
drivers, based on detailed historical analysis
To gain greater control of their cloud spend, top- of consumption patterns, then developing
performing enterprises focus on developing three the analytical model and governance to
capabilities (see sidebar, “Visibility, forecasting, and accurately forecast consumption
optimization go hand in hand”):
— optimizing consumption (through economic
— understanding the business and technical drivers such as reserved instances, or
drivers of consumption, then establishing architecture drivers such as spot fleet) to
granular visibility to monitor and track cloud inform business decisions (for example,

Visibility, forecasting, and optimization go hand in hand

One technology company we know established control of its consumption through developing an integrated approach for visibility, fore-
casting, and optimization (exhibit). In this way, the company was able to achieve more than 90 percent forecasting accuracy and identify
approximately 20 percent in savings.

Companies
Exhibit can take a three-step journey to gain control of their cloud
Companies can take a three-step journey to gain control of their cloud consumption.
consumption.

Optimize spend
Forecast spend
accurately • Value levers including economic (eg,
Make cloud reserved instances) and architecture
consumption visible • Forecast model combining driver (eg, tiered storage) drivers
hierarchy, consumption patterns, • Value-capture execution with clear
• Hierarchy of technical and and business plans (eg, product owners, timelines, and milestones,
business drivers launches, etc) along with execution governance
• Rigorous tagging of consumption • Monthly governance to continuously • Set of unit economics to inform
according to driver hierarchy right-size forecast with spend business owners (eg, cost per
• Understanding of unit costs based owners product, cost per customer)
on historical consumption patterns

Outcomes

• Granular transparency into cloud consumption tied to business drivers

• Accurate forecast of cloud spend (90%+)
• Enhanced ability to manage cash through greater predictability of cloud spend
• Line of sight to 20%+ optimization of cloud spend

55 Unlocking value: Four lessons in cloud sourcing and consumption

 through deriving cloud cost per subscriber or Top-performing enterprises instead are deliberate
product) about bringing together technical, financial, and
sourcing talent into a cross-functional cloud
Lesson 4: Cross-functional FinOps is essential financial-operations (FinOps) team to manage cloud
to manage cloud sourcing and consumption sourcing and consumption (exhibit). In some cases,
Given the complexity and differentiated nature companies can be successful by supplementing
of cloud economics, existing capabilities and their existing sourcing or technology functions
organizational constructs cannot fully capture with relevant talent. This team is then empowered
the value at stake. For many companies, sourcing to orchestrate across stakeholders, translate the
organizations can bring financial and process business’s consumption needs into optimal cloud
discipline, but they often lack the technical depth offerings and pricing arrangements, oversee and
and ability to stratify business demand in sufficient make rapid decisions around resource allocations
detail. This often leads to rigid sourcing standards and cloud usage, and track enterprise-wide cloud
that delay and constrain flexible capacity spend to ensure financial discipline. Importantly,
deployment. On the other hand, entrusting product this cloud-management team is provided with
or technology teams with the task can maximize the right analytics, tooling, and automation, such
agility and grant developers the freedom to flexibly as automated dashboards to better track cloud
and rapidly stand up capacity; however, many consumption in real time and advanced analytics to
organizations have observed that this approach help project demand.
leads to fragmented decision making, poor spend
visibility, and insufficient financial discipline.

Exhibit
Cloud
Cloud FinOps
FinOps teams shouldfollow
teams should followfive
five guiding
guiding principles.
principles.

Context Deep understanding and appreciation of consumption technical and

business requirements, trade-offs, time horizons, and ability to speak
Traditional sourcing organization
language of product teams
focused on contracting and
tendering
Advanced analytical capabilities (to continually dissect demand)
Fragmented stakeholders across
powered by automation and monitoring tools to maximize value
sourcing, finance, engineering,
and product teams
Solid understanding of market dynamics, vendor offerings, and
Fragmented cloud consumption
pricing trends
across the enterprise, with no
coordinated decision making or
planning
Ability to stand up collaborative cross-functional joint decision
making with stakeholders (product teams, finance, etc) with clear
roles and responsibilities

Balanced set of KPIs to performance management and tracking of

actual vs plan with root-cause problem-solving discipline

Unlocking value: Four lessons in cloud sourcing and consumption 56

 require a mindset shift toward a dynamic model
that appreciates the nuances and complexities
As enterprises progress along their cloud journey, of cloud economics, the importance of deeply
transforming the way enterprises source and understanding demand, and the benefits of a
consume cloud will make the difference between revamped organizational approach to sourcing
value capture and value leakage. Success will and optimizing the consumption of cloud.

Abhi Bhatnagar is a partner in McKinsey’s Atlanta office, Will Forrest and Naufal Khan are senior partners in the Chicago
office, and Abdallah Salami is an associate partner in the New Jersey office.

Copyright © 2020 McKinsey & Company. All rights reserved.

57 Unlocking value: Four lessons in cloud sourcing and consumption

 Three actions CEOs can
take to get value from
cloud computing
Leaders need to accelerate their journey to the cloud in order to
digitize quickly and effectively in the wake of COVID-19.

by Chhavi Arora ,Tanguy Catlin, Will Forrest, James Kapla, and Lars Vinter

© Getty Images

July 2020

58
 If you are a CEO, you already know what the cloud 3. putting in place the HR, compensation, and
can do for your business in a post-COVID-19 world. location policies required to attract and retain
You’ve probably even told your organization to get the specialized engineering talent required to
you there already. So why is your move to the cloud¹ operate in the cloud
coming along so slowly, even though you may have
been talking about it for years? It might be because Together, these interventions will help the executive
you and your management team have yet to take a team unite around a coherent point of view about
sufficiently active role, or provide the air cover your the business-driven value that the cloud represents,
chief information officer (CIO) and chief technology how to capture that value, and how to evolve the
officer (CTO) need. company’s operating model accordingly. Without
this perspective, your company may continue to
CIOs and CTOs are on the front foot right now move too slowly toward cloud computing³ for a
thanks to their crucial role during the COVID-19 post-COVID-19 “next normal”—creating the risk of
pandemic. That makes this a good moment to disruption from nimbler attackers.
further elevate top-team support for the cloud
enablement needed to accelerate digital strategy,
the digitization of the company, its channels of Invest for business value
distribution, and its supply chains—all of which During the past 20 years, IT organizations have
already needed to be moving more quickly than they adopted a range of innovations—for example,
were. virtualization and Linux—that have made running
business applications much cheaper and that have
The CEO’s role is crucial because no one else can required only modest investments. Cloud adoption
broker across the multiple parties involved, which has a different economic profile. While exploiting
include the CIO, CTO, CFO, chief human-resources cloud requires investment in building capabilities
officer (CHRO), chief information security officer and migration applications, it’s more efficient in the
(CISO), and business-unit leads. As we explain in this long term, sometimes markedly so for companies
article, the transition to cloud computing represents that have not fully optimized their technology
a collective-action problem—one that requires a environment.
coordinated effort across the team at the top of an
organization. It’s a matter of orchestration, in other The biggest benefits accrue to the business from
words, and only CEOs can wield the baton. faster time-to-market, simplified innovation, easier
scalability, and reduced risk. Cloud platforms can
To get to cloud more quickly, CEOs should ask their help deploy new digital customer experiences in
CIO and CTO what support they need to lead the days rather than months and can support analytics
organization on the journey. Chances are good that that would be uneconomical or simply impossible
three interventions will emerge: with traditional technology platforms.

1. establishing a sustainable funding model Unfortunately, technology-funding mechanisms

to support the investments required to get can stymy cloud adoption—they prioritize features
business value from the cloud requested by the business now rather than critical
infrastructure investments that will allow companies
2. developing a new business-technology to add functionality more quickly and easily in the
operating model² that exploits cloud for speed, future.
agility, and efficient scalability

1
In this article, we use “cloud” to refer to the public cloud rather than companies’ private clouds, in which they attempt to create highly automated
and virtualized application-hosting environments on premises.
2
An integrated operating model organizes technology teams around user-facing products and the underlying platforms that
enable them. For more, see “Products and platforms: Is your technology operating model ready?,” on p. 30 of this compendium.
3
Nagendra Bommadevara, James Kaplan, and Irina Starikova, “Leaders and laggards in enterprise cloud infrastructure adoption,” October 2016,
McKinsey.com.
4
Technical debt is the implied cost of rework caused by implementing a quick but brittle or otherwise architecturally suboptimal solution.

59 Three actions CEOs can take to get value from cloud computing
Each new bit of tactical business functionality built Such questions are unlikely to be asked, much less
without best-practice cloud architectures adds to answered, without serious engagement from the
your technical debt⁴—and thus to the complexity of CEO and other members of the top team.
building and implementing anything in the future.
A big financial-information provider, for example,
CEOs can help the senior team recognize that determined that moving applications in its
infrastructure investments in cloud platforms customer-facing business domains to the public
represent a source of competitive advantage rather cloud could enable much faster and less expensive
than a cost to be managed. Once the top team gets entry into promising markets. Hosting these
that right, a lot else falls into place, including your applications in the cloud meant that technology
technology-funding process, which begins shifting operations in a new country could be set up in
toward products or platforms rather than projects. a couple of weeks at a negligible cost, versus a
Projects are one-time investments funded in a couple of million dollars of up-front investment
yearly boom-and-bust cycle. Products in general for each country. A health-insurance carrier,
(and cloud platforms in particular) require more meanwhile, examined its current project portfolio
stable, ongoing funding and consistent “ownership” and found that it could speed up the capture of
to optimize new functionality and mitigate technical several billion dollars in additional revenue by
debt. adopting the cloud. Moving the systems that help
the insurer interact with healthcare providers was
The top-team conversation will benefit, too, from especially attractive because of the opportunity to
a prioritized, sometimes multiyear road map accelerate the onboarding of new providers.
of domains in which the cloud will accelerate
performance and digital transformation. This will Then, once the investment is made, it’s up to the
help prioritize investments—and avoid defaulting to CEO to demand higher business performance
applications that are technically easiest to migrate. in return for the cloud investment—no more
By asking which business domains (such as order deflecting blame for subpar outcomes to a subpar
capture, billing, or supply-chain optimization) technology environment. If the strategic case for
would benefit most from the speed, innovation, and the cloud is real, it should translate into better
scalability that cloud platforms can provide, top performance. The CEO must demand that it does.
teams can arrive at the highest-priority areas for
movement to the cloud.
A new operating model
Inevitably, resource-allocation issues will arise. Once the funding model is straightened out,
Growth businesses, for example, may be most likely companies must ground the new partnership
to benefit from the cloud, but they are the least between IT and the businesses in an operating
likely to have high margins or excess cash to pony model that reflects and supports their growing
up for a cloud investment. More mature business investment in the cloud.
units may have higher margins, but where, exactly,
should they get the money needed for the cloud—by Here, it will help to think about an integrated
spending less on tactical functionality this year and system rather than a set of individual technologies.
next, or by reducing marketing expenditure? Does Doing so implies organizational change across all
a legacy business have the legs to support a long- of IT, and many of the business units and functions
lived cloud investment? Should the CEO transfer as well. This operating model combines cloud-
money from one business unit to another, or accept based digital technologies and agile operational
lower margins when a business invests in the cloud? capabilities in an integrated, well-sequenced

Three actions CEOs can take to get value from cloud computing 60
 approach that can rapidly accelerate digital strategy management. These quickly moving modular
and transformation. The model helps to coordinate platforms should be run by a platform owner who
end-to-end operations across silos—supporting takes end-to-end responsibility for providing a
customer and employee journeys, for instance— solution and operating the platform as a service.
while taking technology out of quarantine and
making the most of it across all lines of business. Accounting for the risks
Everything in enterprise technology implies risk.
A cloud-ready business-technology operating To mitigate security, resiliency, and compliance
model has many requirements. Here, we focus on concerns relating to the adoption of the cloud,
the few that need intervention from the CEO. companies must be clear-eyed about these
risks. Among other things, that means holding
Improving business interaction rigorous discussions about the best mechanisms
Achieving the speed and agility that cloud platforms for aligning the appetite for risk with decisions
promise requires frequent interaction—for instance, about the technology environment. Getting the
to define and optimize customer journeys— organization to take the right tone on risk will
between IT managers and their counterparts in the require particular attention from the CEO. It’s
business units and functions, particularly those easy to let worries about security, resiliency, and
who own products and capability areas. CEOs compliance stop a cloud program in its tracks.
need to encourage business leaders to appoint Instead of letting risks derail progress, CEOs
knowledgeable decision makers as product owners should insist on a pragmatic risk appetite that
for each business capability. reflects the business strategy, while placing the
risks of cloud computing in the context of the
Too often, business units appoint product owners existing risks of on-premises computing and
who are too new or too junior, and who lack either demanding options for mitigating risks in the
the knowledge or the organizational throw-weight cloud.
to make their decisions stick. Many of these product
or capability owners are “process jockeys,” whose Companies that get the operating model right can
expertise is coordinating stakeholders and tasks. see dramatic improvements. These include better
Look instead for more senior folks capable of target-state economics and lower transition
thinking broadly and strategically. costs. They will also see improved agility and
ability to innovate. One natural-resource company
Going agile in IT implemented agile ways of working for business-
If your company is to gain value from the cloud, application development, infrastructure, and
your IT department must become more agile, if security. In particular, it invested in creating
it isn’t already. That involves more than moving automated, API-based services that developers
development teams to agile product models. Agile IT could use to provision workloads on cloud
also means bringing agility to your IT infrastructure platforms securely and resiliently. As a result, the
and operations by transforming infrastructure company started releasing new capabilities in
and security teams from reactive, “ticket driven” days rather than months, while limiting risk and
operations into proactive models in which scrum technical debt.
teams develop the application program interfaces
(APIs) that service businesses and developers can
consume. Revisit talent
As your cloud investment picks up speed,
Counterintuitively, you should avoid inserting supported by a new, cloud-ready operating model,
translators between IT and the businesses. Instead, your CIO will no doubt be asking for the talent
look to organizational groupings that unite business, needed for cloud. Although cloud computing can
technology, governance, process, and people dramatically boost the productivity of technology,

61 Three actions CEOs can take to get value from cloud computing
it requires specialized and sometimes hard-to- Compared with traditional IT managers,
find technical talent—full-stack developers, data successful CIOs and CTOs in this environment
engineers, cloud-security engineers, identity- and will be both more plugged into a company’s
access-management specialists, and cloud digital transformation and more technologically
engineers. Such talent can be hired externally savvy. In a post-COVID-19 next normal, these
or upskilled from within. Just make sure current executives cannot rely on vendors to figure
HR policies and approaches don’t hobble your everything out for them. They must be open
approach. The basis of performance management to new ideas and willing to learn, to take risks,
and promotion, for example, should be expertise and to fail fast and then quickly correct course
rather than the number of direct reports someone when necessary. It helps if they’re compelling
oversees. communicators who can inspire both business
partners and their own teams to undertake
If your HR policies are not up to speed, you may dramatic change.
need to provide some air cover for your CIO with
the CHRO. Some policies, put in place a decade
ago to contain IT costs, can get in the way of
onboarding cloud talent. Over the years, companies The COVID-19 pandemic has heightened the
have adopted policies that limit costs per head need for companies to adopt digital business
and the number of senior hires, for example, and models—and only cloud platforms can
that require the use of outsourced resources in provide the agility, scalability, and innovation
low-cost locations. Collectively, these policies required for this transition. Although there
produce the reverse of what the cloud requires, have been frustrations and false starts in the
which are relatively small numbers of highly talented enterprise journey to the cloud, companies
and expensive people who may not want to live can dramatically accelerate their progress
in traditional low-cost IT locations. The location by focusing investments in it where they will
issue is why CEOs who are serious about the cloud provide the most business value and by building
have suggested that their CHROs reverse policies cloud-ready operating models.
encouraging the use of low-cost, commoditized tech
talent. In some cases, this new direction takes the But they have to get there first. And that’s
form of newly established tech centers, in places where CEOs have an important role to play—
such as the US West Coast, which are specifically first by becoming more technologically savvy
designed to attract cloud talent. than they have been in the past and next by
addressing the collective-action problem that
CEOs must also make sure their technology often prevents companies from embracing
leaders get sufficient voice in senior forums and new strategic roles for IT. If companies are to
management process given the increasingly fast be successful in a digital next normal, their
integration of digital and business strategy. At many CEOs must ensure that their management
companies, CIOs and CTOs have been among the teams understand the specific ways that cloud
heroes of the COVID-19 response by pivoting their computing can raise revenue growth and
organizations to enable pervasive remote working, margins and how, in close alignment, those
often in a matter of days. The cloud allows CIOs and teams will rally to capture value.
CTOs to play an even more critical role in making
business strategies successful.

Chhavi Arora is an associate partner in McKinsey’s Seattle office, Tanguy Catlin is a senior partner in the Boston office, Will
Forrest is a senior partner in the Chicago office, James Kaplan is a partner in the New York office, and Lars Vinter is a partner
in the Copenhagen office.

Copyright © 2020 McKinsey & Company. All rights reserved.

Three actions CEOs can take to get value from cloud computing 62
 Breaking through
data-architecture gridlock
to scale AI
Large-scale data modernization and rapidly evolving data technologies
can tie up AI transformations. Five steps give organizations a way to break
through the gridlock.

by Sven Blumberg, Jorge Machado, Henning Soller, and Asin Tavakoli

© Getty Images

January 2021

63
 For today’s data and technology leaders, the 1. Take advantage of a road-tested
pressure is mounting to create a modern data blueprint
architecture that fully fuels their company’s digital Data and technology leaders no longer need
and artificial intelligence (AI) transformations. In to start from scratch when designing a data
just two months, digital adoption vaulted five years architecture. The past few years have seen the
forward amid the COVID-19 crisis. Leading AI emergence of a reference data architecture
adopters (those that attribute 20 percent or more that provides the agility to meet today’s need
of their organizations’ earnings before interest and for speed, flexibility, and innovation (Exhibit 1). It
taxes to AI) are investing even more in AI in response has been road-tested in hundreds of IT and data
to the pandemic and the ensuing acceleration of transformations across industries, and we have
digital. observed its ability to reduce costs for traditional
AI use cases and enable faster time to market and
Despite the urgent call for modernization, we have better reusability of new AI initiatives.
seen few companies successfully making the
foundational shifts necessary to drive innovation. With the reference data architecture, data and
For example, in banking, while 70 percent of technology leaders are freed from spending
financial institutions we surveyed have had a cycles on architecture design. Instead, leveraging
modern data-architecture road map for 18 to 24 this blueprint, they can iteratively build their data
months, almost half still have disparate data models. architecture.
The majority have integrated less than 25 percent
of their critical data in the target architecture. All Take the case of a large German bank. By using
of this can create data-quality issues, which add this reference data architecture as its base, the
complexity and cost to AI development processes, organization reduced the time required to define
and suppress the delivery of new capabilities. its data-architecture blueprint and align it with
each stakeholder’s needs from more than three
Certainly, technology changes are not easy. But months to only four weeks. Before adoption of the
often, we find the culprit is not technical complexity; reference data architecture, business executives
it’s process complexity. Traditional architecture would become disillusioned as the CIO, CFO,
design and evaluation approaches may paralyze risk leaders, and business executives debated
progress as organizations overplan and overinvest in architectural choices and conducted lengthy
developing road-map designs and spend months on technology evaluations, even when product
technology assessments and vendor comparisons differences had no material impact on the bank’s
that often go off the rails as stakeholders debate goals. To shift tactics, the company’s CIO identified
the right path in this rapidly evolving landscape. the minimal deviations required from the reference
Once organizations have a plan and are ready to architecture and presented to all the stakeholders
implement, their efforts are often stymied as teams examples of companies across industries that had
struggle to bring these behemoth blueprints to succeeded with the same approach. Executives
life and put changes into production. Amid it all, agreed they had the setup, market positioning, and
business leaders wonder what value they’re getting talent pool to achieve similar results, and the CIO’s
from these efforts. team quickly began building the new architecture
and ingesting data.
The good news is that data and technology leaders
can break this gridlock by rethinking how they Importantly, this isn’t a one-and-done exercise.
approach modernization efforts. This article shares Each quarter, technology leaders should review
five practices that leading organizations use to progress, impact, funding, and alignment with
accelerate their modernization efforts and deliver strategic business plans to ensure long-term
value faster. Their work offers a proven formula for alignment and a sustainable technology build-
those still struggling to get their efforts on track and out. One global bank implemented a new supply-
give their company a competitive edge. based funding process that required business

64 Breaking through data-architecture gridlock to scale AI

Exhibit 1
A reference data architecture for AI innovation streamlines the design
A reference data architecture for AI innovation streamlines the design process.
process.
ML
AI toolbox Ops

Physical
Mobile Web CRM
channels
Systems of
engagement Application databases

APIs and management platform

Data lake
Unified data,
analytics core Curated data vault (by domain) Analytics
Raw data vault

Ingestion Batch Exotic

Real-time streaming

Systems
of record Core processing systems External data Unstructured data

Cloud-enabled data platforms and services (serverless, noOps) Data

Ops

units to reprioritize their budgets quarterly against completing the previous ones. In fact, in our latest
immediate business priorities and the company’s global survey on data transformation, we found that
target technology road map before applying for nearly three-quarters of global banks are knee-
additional funds. This new process helped the bank deep in such an approach.¹
overcome underfunding of $250 million in the first
year while gaining immediate business impact from However, organizations can realize results faster by
refocused efforts. taking a use-case approach. Here, leaders build and
deploy a minimum viable product that delivers the
specific data components required for each desired
2. Build a minimum viable product, use case (Exhibit 2). They then make adjustments as
and then scale needed based on user feedback.
Organizations commonly view data-architecture
transformations as “waterfall” projects. They map One leading European fashion retailer, for instance,
out every distinct phase—from building a data decreased time to market of new models and
lake and data pipelines up to implementing data- reduced development costs when it focused first
consumption tools—and then tackle each only after on the architectural components necessary for its

¹The McKinsey Global Data Survey garnered responses from more than 50 banks, representing various regions and company sizes. To ensure
comparability of results and identification of key trends, several questions on key industry trends and demographics were extracted.

Breaking through data-architecture gridlock to scale AI 65

 Exhibit 2
Each
Eachcommon
commonbusiness use
business case
use is associated
case withwith
is associated components of the
components ofdata
the data
architecture.
architecture.
Illustrative
Most common use cases, by component

AI tools
• Chatbots
• Marketing technology (eg, customer data platform or campaign management)
• Relationship-based pricing
• Intelligent dashboards showing spending patterns

Application programming interfaces (APIs)

• Data monetization
• Data ecosystems
• Virtual assistants
• ID proofing
• Master-data management

Data warehouse Data lake Data streaming

• Financial reporting • Campaign and performance • Personalization

(profit and loss, balance sheet) reporting • Anti-money-laundering (AML)
• Credit-risk reporting • Predictive marketing fraud and transaction monitoring
• Loan-application scoring • 360-degree customer view • Real-time data ingestion
• Compliance (drawing on historical
stores of multiple data types)
• New use-case and model testing

Shared ingestion layer

• Fast access and test-and-learn research and development via AI sandboxes

priority use cases. At the outset, leaders recognized as needed; and a process to shut them down
that for data-science teams to personalize offerings when they aren’t needed. Whereas physical
effectively across multiple online and mobile and virtual environments could once run up IT
channels, including social channels, they would bills for months and years, such environments
need fast access to data. Previously, data scientists can now be accessed on the cloud for less than
had to request data extracts from IT, and data were 30 minutes—the average amount of time that
often outdated when received. they’re actually needed—generating substantial
cost savings.
The retailer’s focus on the architecture its use
cases required enabled development of a highly Once organizations finish building the
automated, cloud-based sandbox environment components for each use case, they can then
that provides fast access to data extracted from a scale and expand capabilities horizontally
shared, company-wide ingestion layer; an efficient to support other use cases across the
manner to spin up analytics and AI sandboxes entire domain. In the case of the retailer, as

66 Breaking through data-architecture gridlock to scale AI

new personalized offerings become ready for programming interfaces (APIs). This approach
deployment, the organization moves the selected reduced time to market and made it easier to use
data features into curated, high-quality data fast-paced analytical modeling, enabling new
environments for production access. customer-360 and master-data-management use
cases, while reducing the complexity of the overall
environment.
3. Prepare your business for change
Legitimate business concerns over the impact any
changes might have on traditional workloads can 4. Build an agile data-engineering
slow modernization efforts to a crawl. Companies organization
often spend significant time comparing the risks, In our experience, successful modernization efforts
trade-offs, and business outputs of new and legacy have an integrated team and an engineering culture
technologies to prove out the new technology. centered around data to accelerate implementation
of new architectural components. Achieving this
However, we find that legacy solutions cannot requires the right structural and cultural elements.
match the business performance, cost savings, or
reduced risks of modern technology, such as data From an organizational perspective, we see a push
lakes. Additionally, legacy solutions won’t enable toward reorienting the data organization toward a
businesses to achieve their full potential, such as product and platform model, with two types of teams:
the 70 percent cost reduction and greater flexibility
in data use that numerous banks have achieved — Data platform teams, consisting of data
from adopting a data-lake infrastructure for their engineers, data architects, data stewards, and
ingestion layer. data modelers, build and operate the architecture.
They focus on ingesting and modeling data,
As a result, rather than engaging in detailed automating pipelines, and building standard APIs
evaluations against legacy solutions, data and for consumption, while ensuring high availability
technology leaders better serve their organization of data, such as customer data.
by educating business leaders on the need to let
go of legacy technologies. One telecom provider, — Data product teams, consisting mostly of data
for example, set up mandatory technology courses scientists, translators, and business analysts,
for its top 300 business managers to increase focus on the use of data in business-driven AI
their data and technology literacy and facilitate use cases such as campaign management. (To
decision making. As part of the training, the data see how this structure enables efficiency across
leadership team (including engineers, scientists, even the larger, more complex organizations, see
and practitioners) shared the organization’s new sidebar, “Sharing data across subsidiaries.”)
data operating model, recent technology advances,
and target data architecture to help provide context The cultural elements are aimed at improving talent
for the work. recruiting and management to ensure engineers are
learning and growing. A Western European bank is
In addition to educating business leaders, cultivating a learning culture through a wide range
organizations should refocus efforts from of efforts:
their legacy stack to building new capabilities,
particularly in the infrastructure-as-a-service — Providing engineers with clearly documented
space. A chemical company in Eastern Europe, for career paths. This includes establishing
instance, created a data-as-a-service environment, formal job levels for engineers based on their
offloading large parts of its existing enterprise productivity, with promotion rounds based on
resource planning and data-warehouse setup to qualitative feedback, their contributions to
a new cloud-based data lake and provisioning the open-source communities, their management
underlying data through standardized application skills, and their knowledge, all assessed against

Breaking through data-architecture gridlock to scale AI 67

 Sharing data across subsidiaries
Across industries, regulators and companies’ risk, compliance, supply chain, and finance departments are increasingly asking
for granular data access covering the headquarters and subsidiaries. On the regulatory side, for example, companies exporting
products that can be used for both civilian and military applications must provide regulators full transparency across the value chain.
On the operational side, such transparency can help provide more advanced insight into global supply chains and operations and
improve productivity, reducing the resources needed to build and manage an end-to-end data architecture in every country.

In response, organizations are moving toward defining data-architecture strategies that can transfer learnings from headquarters
to subsidiaries or vice versa. Companies that do this well, such as Amazon, Google, and Microsoft, harmonize their business and
technology delivery models. This entails setting up a global team with a clear product owner, who owns the global data model, and
dedicated data architects and engineers, who create a shared data vault containing the granular transaction data of the subsidiaries.
Local engineers within the subsidiaries then make any customizations they need while remaining aligned with global teams.

By taking this approach, a French bank drastically improved the quality of its anti-money-laundering and know-your-customer
reporting while lowering the cost of the data architecture for subsidiaries by 30 percent. These positive results have laid the
foundation for groupwide scaling of another data lake to support other use cases, such as calculating risk.

a structured maturity grid. The bank also and virtual conferences. To support this, bank
revised its compensation structure to ensure leaders have instituted an agile performance-
that engineers at the highest job levels receive management model that emphasizes
compensation comparable to that of senior both knowledge and expertise. At other
managers in IT, data, and the business. organizations, the performance measurement
of top executives and team members includes
— Adopting a pragmatic approach to assessing their industry contributions; their success
expertise levels. Research indicates that expert metrics might include, for example, the
engineers are eight times more productive number of keynote presentations they deliver
than novices, so the success of modernization throughout the year.
efforts depends on effective recruitment,
management, and organization of talent. — Emphasizing engineering skills and
To provide a consistent measurement for achievements. To emphasize technical skills,
recruiting, upskilling, and advancement, the the bank encourages everyone in IT, including
bank used the well-known Dreyfus model for managers, to write code. This creates a spirit
skill acquisition to identify five aptitude levels of craftmanship around data and engineering
from novice to master, rate observable behavior and generates excitement about innovation.
through key indicators, and develop individual
training plans based on the feedback.
5. Automate deployment using
— Establishing a culture of continuous technology DataOps
learning. Continuous learning requires the Changing the data architecture and associated
sharing of expertise through formal and data models and pipelines is a cumbersome
informal forums, peer reviews, and freedom to activity. A big chunk of engineering time is spent
pursue online training courses, certifications, on reconstructing extract, transform, and load

68 Breaking through data-architecture gridlock to scale AI

(ETL) processes after architectural changes have standards and developed a code library to optimize
been made or reconfiguring AI models to meet new code reuse. It is currently defining an easier way
data structures. A method that aims to change this to deploy models in production to reduce time
is DataOps, which applies a DevOps approach to lags between model development and use. Once
data, just as MLOps applies a DevOps approach completed, this will reduce the typical time required
to AI. Like DevOps, DataOps is structured into to deploy models and apply results, such as
continuous integration and deployment phases with identifying the right mixtures, from weeks to hours.
a focus on eliminating “low-value” and automatable
activities from engineers’ to-do lists and spanning
the delivery life cycle across development, testing,
deployment, and monitoring stages. Instead of Today, most data technologies are readily available
assessing code quality or managing test data or in the cloud, making adoption a commodity. As
data quality, engineers should focus their time on a result, the difference between leaders and
code building. A structured and automated laggards in the data space will depend on their
pipeline, leveraging synthetic data and machine ability to evolve their data architecture at a brisk
learning for data quality, can bring code and pace to harness the wealth of data collected over
accompanying ETL and data-model changes into decades and new data streaming in. Organizations
production much faster. that can’t move as quickly risk derailing their
digital and AI transformations. The five practices
One large pharmaceutical company is working we have outlined, along with a positive vision
to bring biometric insights to its front line more and a compelling story for change, can enable
quickly using DataOps. It has defined automated organizations to move at the necessary speed,
ways to test new biometric analytics models against building momentum and value along the way.

Sven Blumberg is a senior partner in McKinsey’s Istanbul office, Jorge Machado is a partner in the New York office, Henning
Soller is a partner in the Frankfurt office, and Asin Tavakoli is a partner in the Düsseldorf office.

Copyright © 2021 McKinsey & Company. All rights reserved.

Breaking through data-architecture gridlock to scale AI 69

4. Future-proofing
the foundation
 Managing large technology
programs in the digital era
To successfully implement large technology systems, first accept the
complexity, and then take these six actions.

by Katya Defossez, Mark McMillan, and Hrishika Vuppala

© Getty Images

November 2020

71
 Every IT executive wrestles with implementing integrator, the vast majority of their large-program
large technology programs.¹ In fact, two out of problems can be solved.
three large programs regularly exceed initial
budgets, miss schedule estimates, and underdeliver Unfortunately, success can occur only when tech
against business objectives and benefits, often by leaders sufficiently acknowledge the complexity.
significant margins.² In practice, that means driving superior execution
across ten domains (Exhibit 1). Each of these
Our past research has found that 25 to 40 percent domains is a significant topic unto itself, requiring
of programs exceed their budget or schedules cross-functional skills and capabilities for effective
by more than 50 percent. This failure rate is execution.
especially debilitating for the business because
large programs are typically of critical importance— But the main consideration is how to balance the
for example, for consolidating multiple financial tremendous complexity of the program against the
systems to enable better operational insights or for practical need to make progress. In our experience,
implementing health-insurance enrollment systems. hitting that balance successfully requires
organizations to prioritize five to ten success factors
That failure rate does not have to be the norm. A for each of the ten domains and to develop large-
number of new digital practices and technologies program management capabilities accordingly.
that have emerged in the past few years can Traditional project management is simply not up
drastically improve large program implementations. to the complexities of managing a large number of
When combined with disciplined managerial and interdependent workstreams, the need for technical
talent practices to effectively deal with the vast mastery across many domains, and the importance
complexity of large technology programs, success of adjusting many dependent variables during the
rates can be as high as 90 percent or more. inevitable setbacks and challenges of a program at
this scale.

What drives failure rates? For example, traditional project-manager training

So what drives success (or failure)? If you ask teaches managers to develop a sequential and
100 technology leaders, you are likely to get precise timeline of actions from project start to
almost as many different answers, from unclear finish and then to manage rigorously against those
objectives to ineffective change management, actions. For large, complex programs, this quickly
poor team capabilities, or vendor deficiencies. becomes an impossible task, and the amount of
Others would say excessive customization, the work and rework needed to create this level of detail
wrong platform decision, or ineffective decision is not worth the effort. While big programs still
making. In many ways, they are all right, and this need integrated schedules, they should not pretend
reflects an overriding reality of large technology- to have it all worked out up front. Because of the
implementation programs: they are extremely interplay and dependence across domains (the
complex. While that shouldn’t come as a big surprise, architecture, for example, depends on the sourcing
technology leaders continually underestimate the strategy, and the sourcing strategy depends on
extent and depth of that complexity. For this reason, the architecture), the complexity can be effectively
there is a natural tendency among IT leaders to think managed only by working through the domains
(wishfully, perhaps) that by employing a handful iteratively and in parallel, not by laying out every
of simple fixes or by finding the right systems possible step in advance.

1
While there is no standard definition of a large technology program, more than $25 million in one-time investment can serve as a useful
threshold.
²
Michael Bloch, Sven Blumberg, and Jürgen Laartz, “Delivering large-scale IT projects on time, on budget, and on value,” October 2012,
McKinsey.com.

72 Managing large technology programs in the digital era

Exhibit 1
Effective implementation of large programs requires superior capabilities
Effective implementation of large programs requires superior capabilities
across ten domains.
across ten domains.
Strategy Technology Governance Adoption

1. Vision and strategy 4. Customer experience 7. Program management 9. Operations and

Align stakeholders on and design inputs and governance management
prioritized objectives and Define user journeys and Establish agile control Transition to future-state
a clear future-state requirements based tower to oversee and operating model, supported
visualization on human-centered design enable efficient delivery by changes driven by the
2. Value drivers and cost 5. Platform and data 8. Talent planning and transformation (eg, new
Quantify value and cost of architecture capability building business processes)
use cases and capabilities Design and implement a Assemble or build needed 10. Culture and change
directly linked to the modern architecture, skills to execute the management
business case starting with a minimum program and maximize Incorporate change
3. Product road map viable product (MVP) value realization management across the
Sequence and execute 6. Vendor strategy and organization, and build a
the product and management culture of continuous
architecture road maps Design and execute sourc- learning
ing strategy and vendor
relationship management

Planning Execution

Change management

Six actions that make a big difference that have embraced select agile methods: clear
In our experience working on more than 500 product ownership, prioritized product backlog
large technology-implementation programs, the and road map, small cross-functional teams,
chances for successfully executing against these iterative releases with time-boxed sprints, modular
ten domains significantly increase when tech architecture, objectives and key results (OKRs)³
leaders take six specific actions. Four of them take to manage value capture, and a commitment
advantage of new digital capabilities, while two to a minimal viable product (MVP) and iterative
others are proven, long-standing approaches but releases. Agile mindsets can also be powerful in
are often neglected. supporting a willingness to respond to change, test
and learn, and collaborate.
Digital approaches
One large organization, for example, started a
1. Use select agile methods. Even enterprises claims-system modernization program by using
committed to agile development often are resistant the standard waterfall method. After spending the
to using agile for large programs. There have entire $200 million budget, the program was only
been, however, significant successes in programs a quarter of the way through development—and

3
An approach to defining and tracking desired outcomes (versus activities).

Managing large technology programs in the digital era 73

 what had been developed was a frustrating user principles in cross-functional workshops and
experience and full of defects. The organization interviews to collect current pain points and
made the difficult decision to start over using an requirements. Using the “digital twin” approach,
agile approach focused on small cross-functional it essentially created a digital simulation of a
teams working in sprints through active test- PLM system for the modernization team to work
and-learn cycles with a stable of smaller vendor with. This enabled the team to identify clear
partners. The results represented a stunning issues, such as complex collaboration processes.
turnaround for the program, improving delivery Based on this effort, the company created
velocity and productivity more than threefold, with a “data exchange” for suppliers, providers,
a massively improved user experience and a first and developers (among others) to drive better
release in months rather than years. collaboration around product design specs and
order management. This approach significantly
2. Ground the work in design thinking. Many large improved collaboration among teams and
programs may meet requirements but not user accelerated the release of features.
needs. Successful large programs use design
thinking—a method of problem solving anchored 3. Use cloud-based services. Most enterprise
in end users’ needs—to address this issue. The leaders still tend to reduce the benefits of
practice helps deliver products and services that cloud to efficiencies around infrastructure
users want and need and are therefore more likely management. The capabilities, components,
to use. Another benefit is savings, since teams and services that many cloud service providers
develop only those features that are needed. (CSPs) offer, however, allow companies
managing large-program implementations to
For large programs, design thinking starts with innovate much more quickly and get to market
uncovering user needs at the outset, typically faster through rapid environment provisioning
through a blend of survey-based quantitative and and simpler ways of prototyping or exploring
field-based qualitative research. These efforts novel solutions.
derive a clear picture of how people use the
service or product, signature moments, and unmet By migrating to the cloud while also rapidly
needs. Regular and immersive user engagement scaling cloud-native features for analytics,
throughout the program delivery—for example, in database management, and content
prototyping and user testing—then ensures the management, for instance, a state-government
program maintains alignment with user needs over agency was able to consolidate and modernize
time. three disparate legacy systems across millions of
residents and become 30 percent more efficient
One leading automotive company decided to in terms of operating costs. Running reports,
modernize its product life-cycle management which had been resource intensive and slow,
(PLM) systems. Instead of the traditional process happened much more quickly on the cloud. The
of collecting requirements from R&D, production, agency also took advantage of the CSP’s call-
sales, and after-sales, it applied design-thinking center-management application, which greatly

So what drives success (or failure)? If you ask

100 technology leaders, you are likely to get
almost as many different answers.

74 Managing large technology programs in the digital era

simplified a system that had previously relied on comparable. There is just no substitute for that
multiple providers. kind of experience and “pattern recognition.”
Without it, failure is far more likely. As might be
For large programs, leaders need to systematically expected, these people are hard to find, especially
evaluate how best to take advantage of the cloud. since these sorts of large programs happen
Selecting a software-as-a-service (SaaS) solution, infrequently for most organizations.
for example, can avoid the effort of a custom build
and can result in a best-in-class solution that IT leaders naturally try to address this issue by
is easier to maintain. Or leveraging a platform- bringing together a team of the best people they
as-a-service (PaaS) solution can enable greater can find. But ensuring this team addresses their
developer productivity and access to an ecosystem talent gaps requires an honest assessment of
of thousands of innovative services. the team’s existing talent and a willingness to
bring in the right people, either by hiring them or
4. Use modular architecture to increase flexibility contracting with vendor partners. This can be time
and vendor competition. Many organizations are consuming, but it is necessary. Hiring a systems
moving to more modular, flexible architectures, integrator to fill the holes often won’t work, since
such as microservices. This move not only creates it has different incentives—scope creep or delays
longer-lasting, more “future-proof” applications increase its revenue—and is focused on delivering
but also allows companies to use a multi-vendor against the contract rather than ensuring you are
sourcing strategy and thus solve one of the doing everything needed to manage the program
longest-standing challenges with large-program effectively.
delivery: keeping vendor incentives aligned with
your own. With single-vendor solutions, it’s nearly 6. Be aggressive about necessary course
impossible to apply steady cost pressure, as often corrections. Any program of this scale is going
a significant risk premium is worked into a fixed- to run up against issues. When that occurs, CIOs
price contract and change orders are common. and the leadership team analyze the problem and
Alternatively, time-and-materials contracts create recommend a correction. But these interventions
incentives for vendors to extend and expand are often not aggressive enough to get to the root
programs and thus grow their revenue stream. cause of the problem. That’s understandable, since
these programs are so complex. Their multiple
Instead, modular architectures allow companies to interdependent systems can make it difficult both
work with multiple vendors who can be replaced as to pinpoint the source of the issues and to muster
needed, leading to significantly better outcomes. the often significant effort needed to course
For example, one public-sector organization correct.
awarded a development master-services contract
to four development vendors. For each phase of However, an unwillingness to admit—or the inability
the program, the vendors either competed or were to realize—that the issues are more complex and
directly awarded small packets of work, such as require more work than anticipated means that
front-end design services or development and problems continue and often get worse. The
testing services for each component. Over time, research is quite clear on this point. Early cost and
the strongest-performing vendors—those bringing schedule overruns end up, on average, much worse
their A team at reasonable cost—won more of the in most programs, often costing twice as much as
work, leading to superior outcomes. anticipated—and that’s despite the interventions
of program leaders (see Exhibit 2 for an example
Proven approaches analysis).

5. Get people with large-program (ideally Fortunately, there are many examples of
comparable) experience. Given the complexity of successful interventions. One public-sector
large program implementations, it is crucial to have organization, having invested $60 million of
people who have already done them, or something its $200 million budget for a tax-processing

Managing large technology programs in the digital era 75

 Exhibit 2
Despite
Despiteinterventions,
interventions,issues typically
issues get
typically worse
get over
worse a project’s
over life life
a project’s cycle.
cycle.

Intervention
200 Performance to date
Typical perceptions of
how the program will
150 get back on track
Schedule
slippage Actual track record of
% comparable programs
100

50

0
25 50 75 100

Project progress
%

modernization program that was way behind

schedule, decided to forfeit the initial investment
and start over by making some aggressive changes. These and other examples show that organizations
It first hired a new systems integrator and software can be successful with their most important
vendor. It then developed a new business case as technology investments—if they master a broad
a “north star” to guide the relaunched program. array of success factors, take advantage of new
The results: a successful project for less than digital techniques to de-risk delivery, and ensure
$125 million—less than its original budget, even they have the right capabilities from the start.
accounting for the initial sunk investment.

Katya Defossez is a partner in McKinsey’s Houston office, Mark McMillan is a partner in the Washington, DC, office, and
Hrishika Vuppala is a partner in the San Francisco office.

Copyright © 2020 McKinsey & Company. All rights reserved.

76 Managing large technology programs in the digital era

McKinsey Digital
May 2021
Copyright © McKinsey & Company

McKinsey.com
@McK_Digital
McKinsey Digital