1.

Levels of HR analytics
Descriptive, Diagnostic, Predictive , Prescriptive
The 4 types of HR analytics
1. Descriptive HR analytics
Descriptive analytics uses historical data to reveal past trends. It focuses on employee
metrics such as workforce demographics, retention patterns and productivity levels. This
information helps you gain a deeper understanding of what happened, and why.
This type of HR analytics is most useful when you need a clear picture of the current state of
your workforce.
For example, descriptive analytics tells you:
What’s causing voluntary turnover: Descriptive analytics shows you when employees are
overworked and makes it easy to spot signs of burnout. When these HR metrics precede a
sudden spike in unanticipated resignations
2. Predictive HR analytics
While descriptive analytics focuses solely on historical data, predictive analytics forecasts the
future of your workforce. It looks at patterns in employee activity to forecast what will
happen moving forward.
Using statistical analysis and machine learning to predict outcomes, predictive HR analytics
allows you to anticipate and plan for future events.
For example, many HR teams use predictive HR analytics to guide:
Future workforce planning: Predictive analytics empowers you to forecast future talent
needs, identify skills gaps and develop strategies to address them. You can compare your
current workforce to the one your company will need a year from now to develop targeted
strategies to bridge gaps.
3. Prescriptive HR analytics
Prescriptive HR analytics takes predictive analytics further by providing highly actionable
insights. Instead of showing what might happen, it recommends steps you can take to steer
your workforce in your desired direction.
This type of HR analytics is valuable when preparing for various HR scenarios. By combining
multiple data sources,
For example, prescriptive analytics helps you:
Reduce your attrition rate: While it’s possible to experience high turnover and still see
company growth, a consistently high attrition rate means business is shrinking. Prescriptive
analytics provides insights you can use to lower it. For instance, workforce analytics software
may alert you to early signs of employee disengagement that would otherwise go unnoticed.
4. Diagnostic HR analytics
Diagnostic HR analytics focuses on identifying the root causes of workforce challenges and
performance gaps. Instead of guessing what might work, you can develop targeted solutions
and policies.
 This type of analytics is instrumental in helping you understand why problems arise and how
to address them. For instance, you might find out why job candidates take so long to accept
offers or what’s causing so many people to leave your company.
For example, successful HR teams use diagnostic analytics to:
Reduce the risk of burnout: Employee burnout is a diagnosable condition characterized by
feelings of fatigue, distrust and reduced productivity.
HR metrics and its types
System of data that is used to measure HR and its relationship and effect on an organization.
Used to uncover the areas of excellence and also explore the areas of improvements
HR Metrics are often complex and comprises of ratios; they are different from industry to
industry
Two different benchmark to assess these ratios
- Talent Development Reporting principles (TDRP)
- Chartered Institute of Personal and Development (CIPD)
Efficient and Effective Metrics
- Efficient Metrics – Talks about “Doing the Things Right” – If anything performed at the
least possible time.
- Effectiveness metrics – Focus on the “Doing the Right Things” – If anything performed as
per the expected level of outcome.
Lead and Lag Metrics
- Lead Metrics – Is also called as process metrics
- Lag Metrics – Outcome metrics Highlight
Training and Development Metrics
Training Completion Rate: The percentage of employees who complete assigned training
programs.
Learning & Development ROI: Measures the return on investment for training programs
based on improved performance or productivity.
Time to Proficiency: The average time it takes for a new hire or trainee to reach full
productivity.
Recruitment Metrics
Time to Hire: The average time it takes to fill a position from the job posting to the offer
acceptance.
Cost per Hire: The total cost involved in hiring a new employee, including advertising,
recruitment agency fees, and on boarding costs.
Quality of Hire: Assesses the performance and retention of new hires, often measured
through performance reviews and turnover rates.
2. Analytics value chain
STRATEGY
Strategic Business Plan
Innovate
Produce, Sell
 Service
Customer: Analysed Needs
Price
Quality
Service
Operations:
Planning targets:
Process Times
Product Quality
Output volume
Talent Management:
Workforce Planning – Hiring
Deploying-
compensating-
developing-
engaging-sustaining
EXECUTION
Financial Outcomes:
Income Statement
- Revenue
- Expenses
Balance Sheet
- Assets and Liabilities
Customer responses:
Number contacted and number responding
Conversion Rate and Spend Satisfaction level
Return Rate
Operational Outputs:
Unit cost – Cycle time
Quantity: Output/input
Quality: Error rates, shrink, rework
HR Services:
Workforce and succession plan-hire cost-time to fill – quality pay ad benefits cost
L&D Spend-engagement program
3. Steps of predictive analytics,
Determine key performance indicator,
Analyse and report data,
Interpreting the results and
Predicting the future
1. Determine key performance indicator (KPIs)
• Through conversation with the VP, segment the data into three types of measures
 • Efficiency
• Effectiveness
• Outcomes
• Then gathering of data In which department does the data reside?
Who is the gate keeper?
Is it sensitive information?
What approvals?
Type of data
Format of data
Standard process for requesting data
Standard turnaround time for the request
• Communication to gather data Procedure / Request rising
Approvals
About Confidentiality
• Formatting the data for analysis Separated by rows and column
Vertical style
Cross tab structure is required
Where each column is a variable

2. Analyse and report data

Descriptive Statistics
Mean
Standard Deviation
Frequency distribution
This analysis is to understand the data
Inferential Statistics
Relationship amongst the variables
Correlation and Regression
T-Test
Analysis of Variance

4. Typical data sources, Typical Data Issues with HR Analytics,

Typical data source
HR Information Systems (HRIS): These systems store employee data such as demographics,
job roles, tenure, and salary information. Common HRIS platforms include Workday, SAP
Success Factors, and Oracle HCM.
Payroll Systems: Provide data on compensation, benefits, deductions, and bonuses. This data
is crucial for analyzing cost, equity, and incentive effectiveness.
Applicant Tracking Systems (ATS): Store data on job applicants, including resumes,
application status, and recruitment metrics like time-to-hire and cost-per-hire. Examples
include Greenhouse, Lever, and iCIMS.
Performance Management Systems: Capture employee performance reviews, feedback, goal
attainment, and development plans. Tools like 15Five, Lattice, and SuccessFactors are often
used.
Employee Surveys: Gather data on employee engagement, satisfaction, and sentiment.
Survey tools like Qualtrics, SurveyMonkey, and Glint are commonly used.
Learning Management Systems (LMS): Track data related to employee training, certifications,
course completions, and skill development. Platforms include Cornerstone OnDemand,
LinkedIn Learning, and Moodle.
Time and Attendance Systems: Provide data on employee work hours, absenteeism, and
overtime. Kronos and TSheets are popular examples.
Exit Interviews: Collect data on reasons for employee turnover, which can be analyzed to
improve retention strategies.
Employee Communication Platforms: Tools like Slack, Microsoft Teams, or internal social
networks provide data on collaboration, communication patterns, and organizational culture.
External Data Sources: Labor market data, industry benchmarks, and economic indicators can
provide context and comparison points for internal HR data. Sources include the Bureau of
Labor Statistics (BLS), LinkedIn Talent Insights, and Glassdoor.
Typical data issues
• Biggest Challenges of HR Analytics?
Finding people with the right skillset to gather, manage, and report on the data
Data cleansing
Data quality
Too much data to parse or not knowing what data is most important
Data privacy and compliance
Proving its worth to executive leadership
Tying actions and insight to ROI
Data Quality Issues:
Incomplete Data: Missing or incomplete employee records, such as missing demographic
information, performance ratings, or absence records.
Inaccurate Data: Errors in data entry, such as incorrect dates of employment or inaccurate
salary information.
Inconsistent Data: Discrepancies in how data is recorded across different systems, such as
inconsistent job titles or department names.
Data Silos:
Fragmented Data: Data spread across multiple systems (e.g., payroll, recruitment,
performance management), making it difficult to integrate and analyze holistically.
Lack of Integration: Challenges in combining data from different sources due to incompatible
formats or lack of integration tools.
Data Privacy and Compliance:
Confidentiality Issues: Ensuring employee data is kept confidential and compliant with data
protection regulations (e.g., GDPR, HIPAA).
Access Control: Restricting access to sensitive HR data to authorized personnel only.
Unstructured Data:
Qualitative Data: Difficulty in analyzing unstructured data, such as employee feedback from
surveys, interviews, or social media, which is often qualitative and non-standardized.
Outdated Data:
Timeliness: Relying on outdated or obsolete data that doesn’t reflect current workforce
dynamics, leading to poor decision-making.
Bias in Data:
 Historical Bias: Historical data that reflects biased practices (e.g., biased hiring or promotion
decisions) can perpetuate bias in predictive analytics.
Sampling Bias: Data that is not representative of the entire workforce can lead to skewed
analysis.
5. Cleaning and Supplementing of data.
- Remove Unwanted observations - This includes duplicate or irrelevant observations.
Duplicate observations
Duplicate observations most frequently arise during data collection, such as when you:
• Combine datasets from multiple places
• Scrape data
• Receive data from clients/other departments
Irrelevant observations
Irrelevant observations are those that don’t actually fit the specific problem that you’re
trying to solve.
• For example, if you were building a model for Single-Family homes only,
you wouldn't want observations for Apartments in there.
• This is also a great time to review your charts from Exploratory Analysis.
You can look at the distribution charts for categorical features to see if
there are any classes that shouldn’t be there.
• Checking for irrelevant observations before engineering features can save
you many headaches down the road.
- Filter Unwanted Outliers-
• Outliers can cause problems with certain types of models. For example, linear
regression models are less robust to outliers than decision tree models.
• In general, if you have a legitimate reason to remove an outlier, it will help your
model’s performance.
• However, outliers are innocent until proven guilty. You should never remove an
outlier just because it’s a "big number." That big number could be very informative for
your model.
• We can’t stress this enough: you must have a good reason for removing an outlier,
such as suspicious measurements that are unlikely to be real data.
Two ways to handle this
• Dropping observations that have missing values
• Imputing the missing values based on other observations

Supplementing data
• This basically sourcing data from various sources
• Something that completes or makes an addition
• Add something to something to make it larger or better