An 8-Step Framework for People Analytics

Framework for People AnalyticsOnce you understand the what, why, and who of HR analytics, it’s important to apply a basic framework to your HR analytics projects.

A framework helps enhance clarity and consistency of a project’s progression and next steps. Each project, problem, and solution will always be unique, but developing an overall process will help guide you to the information you’re seeking.


[FREE DOWNLOAD]  The Beginner's Guide to HR Analytics: The What, Why, and How of HR Analytics


Here are 8 steps to follow for any people analytics project:


1. Understand


When beginning an HR data analytics project, the first step is to understand your purpose. Ask yourself the following questions:

  • What business problem are we trying to address?
  • Are we asking the right questions?
  • Who will champion this project?
  • What are the expected or hopeful outcomes of this project?
  • Will the results of this project lead to any meaningful changes?
  • Who will be impacted most, least, or not at all by these results?


2. Identify


After you have a strong understanding of the project, you should identify the data and sources of data needed to carry out the project. Here are some questions to consider:

What types of data do we need?

  • Behavioral data (e.g., absenteeism)
  • Perceptual data (e.g., employee engagement)
  • Demographic data (e.g., tenure)

Is the data structured or unstructured?

  • Structured (e.g., numeric data)
  • Unstructured (e.g., text, audio, or video)

Does the data already exist, or does new data need to be created?

What needs to be done to gather the necessary data?


3. Collect

After identifying the kinds and sources of data you need, you can begin collecting the data by:

  • Conducting interviews, focus groups, or surveys
  • Working with other departments to obtain necessary data
  • Tracking and documenting obstacles to keep in mind for future projects


4. Clean

Before you can analyze your dataset, it needs to be cleaned. Scan your dataset for:

  • Data points that don’t belong (e.g., numbers where text should be)
  • Incorrect data (e.g., a value of 12 in a 10-point scale)
  • Inconsistent data (e.g., a manager has 5 direct reports, but has data for 6 direct reports)
  • Extreme outliers (e.g., a value of 100 when the next highest value is 15)
  • Excessive amounts of missing data


5. Analyze

This is the step people typically think of when they see “HR Analytics.” Here are three tips for analyzing your data:

  1. Always start with descriptive analytics. This will help you to familiarize yourself with the data.
  2. Never jump straight into diagnostic or predictive analytics. If you do, you might overlook important insights that could render your advanced analytics unreliable or useless.
  3. Think of statistics and machine learning like toolsets. Each analysis is a separate tool, which means certain tools are better for answering certain questions.


6. Extract

Analysis is about creating results from data, and extraction is about determining which results are meaningful. Ask yourself the following questions:

  • Which results are noise?
  • Which results tell the most important or relevant stories?
  • Do the results support your hypotheses or expectations?
  • Are any results surprising?
  • Did any results spark thoughts for additional analyses or projects?


7. Communicate


After you’ve gone through the previous steps, now is the time to share what you’ve found. This step depends entirely on the project and your role. You might communicate insights:

  • Only to your immediate manager
  • To individual department and team leads
  • To the senior leadership team
  • To the entire organization

No matter who you present to, make sure your points are clear, concise, explainable, and rely on visuals whenever possible.


8. Evaluate

HR analytics projects never really end because new data is always being created. As such, an evaluation schedule should be created to assess:

The Project: What went well with this project? What didn’t go well?

Follow-through: Were any plans or strategies put into action?

Trends: Have employee behaviors or perceptions changed? What do the trends look like at 3, 6, 12 months after your project?

Impact: What impact did your project have on employees



Want to make your business better through data-driven insights? Download our ebook, The Beginner’s Guide to HR Analytics.

New call-to-action

Published December 18, 2019 | Written By Teresa Preister