Overview

Web analytics is the measurement, collection, analysis and reporting of

web data for purposes of understanding and optimizing web usage.

Used as a tool for business and market research, and to assess and
improve the effectiveness of a website.

Helps companies measure the results of traditional print or broadcast

advertising campaigns.

Helps gauge traffic and popularity trends which is useful for market
research.

Provides information about the number of visitors to a website and the

number of page views.

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 2

 Basic Steps of the Web Analytics Process

Collection of data:
• This stage is the collection of the basic, elementary data. Usually, this data is counts of things.
The objective of this stage is to gather the data.

Processing of data into information:

• This stage usually take counts and make them ratios, although there still may be some counts.
The objective of this stage is to take the data and transform it into information, specifically
metrics.

Developing KPI:
• This stage focuses on using the ratios (and counts) and infusing them with business strategies,
referred to as Key Performance Indicators (KPI). Many times, KPIs deal with conversion
aspects, but not always. It depends on the organization.

Formulating online strategy:

• This stage is concerned with the online goals, objectives, and standards for the organization or
business. These strategies are usually related to making money, saving money, or increasing
marketshare.

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 3

 Off-site and On-site web analytics

Off-site web analytics On-site web analytics

• Refers to web measurement and • Measure a visitor's behaviour once

analysis regardless of whether you on your website.
own or maintain a website. • Drivers and conversions; for
• It includes the measurement of a example, the degree to which
website's potential audience different landing pages are
(opportunity), share of voice associated with online purchases.
(visibility), and buzz (comments) that • On-site web analytics measures the
is happening on the Internet as a performance of your website in a
whole. commercial context.
• This data is typically compared
against key performance indicators
for performance, and used to
improve a website or marketing
campaign's audience response.

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 4

 Web Analytics Data Sources

Direct HTTP request data:

directly comes from HTTP
request messages (HTTP
request headers).

Network level and server generated data associated

with HTTP requests: not part of an HTTP request, but
it is required for successful request transmissions.
For example, IP address of a requester.

Application level data sent with HTTP requests: generated and

processed by application level programs (such as JavaScript, PHP,
and ASP.Net), including session and referrals. These are usually
captured by internal logs rather than public web analytics services.

External data: can be combined with on-site data to help augment the website behavior data
described above and interpret web usage. For example, IP addresses are usually associated
with Geographic regions and internet service providers, e-mail open and click-through rates,
direct mail campaign data, sales and lead history, or other data types as needed.

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 5

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 6
 Frequency /
Single Page
Hit Bounce Rate Session per Click path
Visit / Singleton
Unique visit

Page Views per

Exit Rate / %
Page view Impression Average Click
Exit
Session

Page Time Average Page

Event New Visitor Site Overlay
Viewed Depth

Page Visibility Active Time /

Visit / Session Return Visitor
Time Engagement KPAs for
On-site
First Visit / First
Session
Repeat Visitor
Page View
Duration
Average Page
View Duration. Analytics

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 7

 Web Analytics 3.0 is:

Web Analytics 3.0 is:

• The analysis of qualitative and quantitative data from
your website and the competition
• To drive a continual improvement of the online
experience that your customers and potential
customers have,
• Which translates into your desired outcomes (online
and offline)
You have so little actionable insight because clickstream data is great at the what, but not at the why. That is
one of the limits of clickstream data. We know every click that everyone ever makes and more. We have the
what: What pages did people view on our website? What products did people purchase? What was the
average time spent? What sources did they come from? What keywords or campaigns produced clicks? What
this, and what that, and what not? You need to find out ways to answer the why!

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 8

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 9
04-10-2024 Lecture Presentation | © Dr. A. K. Kar 10
 Web analytics – Cluster Analysis

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 11

 Web analytics – Classification Analysis

Mapping an input pattern, represented as an input vector, to

a particular class, set or category.
• Requires initial training of the classification model using the data of the
pattern against the specific criteria
• After training maps the input data/object to the classes with which it has
greatest similarity with
• Facilitates Cohort Analysis (grouping of users)
• Good for sending email campaigns to cohorts of subscribers.

E.g. determine E.g. Determining if a E.g. Determining

whether a given email visitor to a website is legitimacy of online
is spam or non-spam real or not purchase behaviours

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 12

 Recommendation – Association Rule Analysis

Discovering interesting relations between variables in large databases

without attempting to explain them

Basis of creating recommendation engines in websites.

Looks for popular sets of visits of pages by individuals to recommend

other pages of related content

Popular in e-commerce and content marketing websites

For example: If a person visits pages 1 & 2, often he visits page 3.
Rule form: “Antecedent -> Consequent [support, confidence]”.
Association Rule: Page 1, Page 2-> Page 3 [0.5%, 60%]

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 13

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 14
 Explore these tools

Check out Open Site Explorer

• https://analytics.moz.com/pro/link-explorer/home
• Domain Authority, Page Authority, Spam score

Check out Majestic Web Analysis

• https://majestic.com/reports/site-explorer
• Check trust flow (quality), Citation flow (quantity)
• http://websiteseochecker.com/bulk-check-trust-flow-citation-flow/

Alexa (shut down now for public from 2022)

• http://www.alexa.com/siteinfo
• Check Alexa rank
• What does it indicate?

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 15

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 16
 Create a GA profile

Configure Site Search Settings

Keep searcharg, author, title, SEARCH

Exclude startLimit, SORT, endLimit, (and maybe
searchscope)
 Copy and paste tracking code to
botlogo file. Also manually add to
pages that don’t use botlogo (view
source to check).

Restart server.

Verify code is visible in source and

GA has started tracking.
Top Content is the most commonly viewed report, so let’s use it to define the labels.
Switch from Explorer view to Navigation Summary to see common paths through the site.
If you compare date ranges, be sure to pick the same start day of the week (so waves match up) and note other events (like first day of the semester).
Other built-in reports: Mobile Devices Wordpress does include GA tracking code
Other built-in reports: Referring Sites
Other built-in reports: Keywords that direct people to your site
Other built-in reports: Site Search (search terms entered in the website)
 User behaviour

Affinity Categories

• General likings of users landing in web page

• Useful for general ads
• Less sticky

In market segments

• Intent while searching and landing

• Useful for affiliate marketing
• More sticky
The real utility is in creating Custom Reports; you set Metrics (which stats are reported) and Dimensions (what data is displayed)
Landing pages are the first page a user sees on your site. Get this custom report: http://goo.gl/sI1TV
How are people reaching this version of the screen? Click through the listed URL to see referring page.
Lots of people searching for link to catalog page (we also see many people searching for library homepage)
Keywords used to find catalog pages in general. Get this custom search: http://goo.gl/ePpBj
Advanced Segments lets us look at slices of data. Here I’ve created a segment that I’ll only use to view numbers of search terms:
Catalog searches w/ 2, 3, 4, 5, 6-10, 11+ Words
Get this custom segment: http://goo.gl/e8kLd
I used Test Segment to see
these figures (no need to view
a report). Otherwise easier to
view one by one.

No surprises up top: 2, 3, 4, 5,
6-10, 11+ search terms steadily
decrease in number

View more rows to see: Bump

in the 8-15 range (implies
people are searching using
sentences or article titles, but
less now with Summon)

Implications for instruction/

prompts/ help pages
We’ve created Filters (that permanently change the data collected) for My Millennium screens (My Library
Account) to remove patron number.

This protects patron privacy and helps group types of MLA pages.
We strip out scope number (not needed by us) and patron number.
 How do patrons find catalog page

• Keywords
• Referrers

How do patrons search in the catalog?

• Types of searches (keyword, author, title)

• Scopes, limits
• Number of search terms

How do patrons use the records they find?

• Track outbound links to e-resources

• Track outbound links to library webpages

Google Analytics can show user behavior

 What to include in help screens, prompt text on
search pages, error pages
 What is our default search (keyword vs title)
 What do we include in catalog (if very low click
throughs to an ebook package, do we rely on
links from vendor site (database) or Web Scale
Discovery Engine

Facilitate data driven decisions?

 How does our implementation of Summon affect catalog usage?

• Click throughs from Summon (referrers and record # searches)

• Decline in number of searches?

Does new emphasis on quality searches in instruction classes change

number of search terms?

• Decrease in single word searches and over 10 word searches?

What to look at in the future?

 Why Web Analytics

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 40

 Please explore these websites..

SeoClerks.com Like4like.org Moz.com Sendinblue.com

Digg.com Alexa.com Serplab.co.uk SEOMoz.org

TwitterFeed.com ManageFlitter.com aHrefs.com searchenginewatch.com

blackhatworld.com wickedfire.com submitexpress.com woorank.com

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 41

 Other Major SEO Metrics

MOZ Domain Authority (da), MOZ Page Authority (pa), MOZ Links In (links), MOZ External
Equity Links (equity),

Alexa Rank (a_rank), ALexa Links number (a_links), Alexa Country (a_cnt), Alexa Country
Rank (a_cnt),

SemRush Domain (sr_domain), SemRush Rank (sr_rank), SemRush Keywords number

(sr_kwords), SemRush Traffic (sr_traffic), SemRush Costs (sr_costs) (Less available now)

SemRush URL Links number (sr_ulinks), SemRush HOSTNAME Links number (sr_hlinks),
SemRush DOMAIN Links number (sr_dlinks).

Majestic Citation Flow (cf), Majestic Trust Flow (tf), Majestic ExtBackLinks (ebl), Majestic
Refered Domains (refd).

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 43

 In class assessment

• Identify the top 5 on-page and top 5 off page optimization

strategies: https://ahrefs.com/blog/seo-tips/
• Comment critically on the shortcomings of this article:
https://colibri.io/growth-hacking/does-link-building-still-
make-sense/

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 44

 Web server log file analysis

Web servers record some of their transactions in a log file. It was soon realized that
these log files could be read by a program to provide data on the popularity of the
website.

A page view was defined as a request made to the web server for a page, as opposed
to a graphic

A visit was defined as a sequence of requests from a uniquely identified client that
expired after a certain amount of inactivity, usually 30 minutes.

The emergence of search engine spiders and robots, along with web proxies and
dynamically assigned IP addresses, made it difficult to analyse human visits only.

Log analyzers responded by tracking visits by cookies, and by ignoring requests from
known spiders.

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 45

 Click analytics

Click analytics is a special type of web analytics that gives special attention to clicks. Commonly, click
analytics focuses on on-site analytics.

An editor of a website uses click analytics to determine the performance of his or her particular site,
with regards to where the users of the site are clicking.

Also, click analytics may happen real-time or "unreal"-time, depending on the type of information
sought.

Typically, front-page editors on high-traffic news media sites will want to monitor their pages in real-
time, to optimize the content.

Editors, designers or other types of stakeholders may analyze clicks on a wider time frame to help them
assess performance of writers, design elements or advertisements etc.

Ideally, a click is "logged" when it occurs, and this method requires some functionality that picks up
relevant information when the event occurs.

Alternatively, one may institute the assumption that a page view is a result of a click, and therefore log a
simulated click that led to that page view.

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 46

04-10-2024 Lecture Presentation | © Dr. A. K. Kar 47