People analytics is about gathering and analysing data about people in a workforce. It’s sometimes called HR analytics or workforce analytics. People data is found in HR systems, from other departments like IT and sales, and from external sources such as salary surveys. Using people data offers the opportunity to contribute to an organisation’s strategy by creating insights on what people can do to drive change.
This factsheet explores what people analytics is, why it’s important and how it’s used. It introduces key terms such as correlation, causation, predictive and prescriptive. It also discusses who is responsible for people analytics as well as the strategy and process.
What is people analytics and why is it important?
People analytics is about analysing data about people to solve business problems. It’s sometimes called HR analytics or workforce analytics. One academic paper defines it as ‘a number of processes, enabled by technology, that use descriptive, visual and statistical methods to interpret people data and HR processes. These analytical processes are related to key ideas such as human capital, HR systems and processes, organisational performance, and also consider external benchmarking data’.
Five reasons for using people analytics:
It can be used to measure a workforce, for internal and external stakeholders, in a range of areas such as performance, wellbeing, and inclusion and diversity. See more on workforce planning.
It enables more effective evidence-based decisions on improving workforce and organisational performance.
It can demonstrate the impact of HR policies and processes on workforce and organisational performance.
It can be used to estimate the financial and social return on investment of change initiatives.
‘Analytics and creating value’ is a core knowledge element in our Profession Map, with ‘people analytics’ as a specialist knowledge element.
People analytics can be applied to almost an aspect of HR activity. For example:
Enhancing employee morale: Organisations can measure the drivers of employee engagement and adapt their practices accordingly to enhance employee morale.
Improving retention: Organisations suffering from high turnover of key employee groups can use people analytics to anticipate areas with specific issues and tailor incentives to curb attrition. Find out more on turnover and retention.
There’re more case studies of people analytics in action in our Valuing your Talent web pages and in our research report Human capital analytics and reporting: theory and evidence.
Find out more about how HR and finance professionals are using people data in our report People analytics: driving business performance with people data in association with Workday, as well as the summary reports People analytics: international perspectives.
Does collecting data involve monitoring and surveillance?
Potentially. Technology makes it easy to seamlessly collect data about people. Websites visited, time spent on specific apps, comments made on the organisation’s social networking site. Organisations can monitor their workforce within the bounds of law where they operate. Even if it’s lawful, how an organisation collects and uses monitoring data can be contentious, particularly if employees feel that it’s irrelevant, unnecessary or too intrusive. Watch Don’t be creepy: how to use data for good by Dr Heather Whiteman of the University of California, Berkeley at ‘People Analytics & Future of Work’ 2020.
If introducing employee monitoring software, it’s important to:
- Be transparent. Explain clearly what you’re monitoring and why.
- Consult with employees to ensure the measures are relevant and necessary. Measures can be about ensuring compliance as well as helping employees become better at their jobs.
- Be mindful of cultural differences and monitor your system to make sure it does not discriminate against minority groups.
What is descriptive, predictive and prescriptive analytics?
People analytics uses workforce or HR data, either qualitative or quantitative, to investigate a certain concept with the help of computer programmes and modelling techniques. There are three main levels of people analytics capability. Most organisations are able to do level 1 only, very few are able to complete level 3 analytics:
Level 1a– descriptive analytics: Uses descriptive data to illustrate a particular aspect of HR, for example recording absence, annual leave, and attrition and recruitment rates. At level 1, no analysis is applied to the data beyond using it to describe a certain concept or illustrate its change over time (sometimes called trend analysis). See our factsheet which gives commonly-used measures of turnover and retention.
Level 1b – descriptive analytics using multidimensional data: Combines different data sets, or types of data, to investigate a specific idea can help to uncover interesting relationships between different HR activities and processes. Using two different types of data to create an analytics output is known as multidimensional analytics (for example, combining leadership capability data with engagement scores to measure leadership effectiveness).
Level 2 – predictive analytics: Uses data to predict future trends can help HR professionals to plan for future events and scenarios, and ensure they are able to deliver to the business. Predictive analytics for forecasting require high quality and robust data, and specialist technology and capability.
Level 3 - prescriptive analytics: Applies mathematical and computational sciences to suggest decision options to take advantage of the results of descriptive and predictive analytics. Prescriptive analytics specifies both the actions necessary to achieve predicted outcomes, and the interrelated effects of each decision.
What is quantitative and qualitative data, correlation and causation?
HR data is information about any aspect of employees or the HR management system. Data comes in many forms, and may be quantitative or qualitative.
Quantitative data can be measured and illustrated through numbers
- How many? How much?
- Facts are value-free / unbiased
- Report statistical analysis. Basic element of analysis is numbers
- ‘Counts the beans’
- Examples: number of employees, remuneration rates, productivity
Qualitative data can’t be measured and are often subjective assessments representing an individual’s view of something.
- What? Why?
- Facts are value-laden and biased
- Report rich narrative, individual; interpretation. Basic element of analysis is words/ideas.
- Provides information as to ‘which beans are worth counting’
- Examples: employee opinion survey feedback, appraisals and performance reviews, learning and development outcomes
For example, a fondness for chocolate is qualitative data as it relates to an individual’s preference towards chocolate, while the dimensions of a chocolate bar is quantitative data as it relates to its numerical size (length, width and height). In HR, an individual’s age or performance rating is quantitative data, whereas their engagement data (such as job satisfaction) is qualitative.
Data is held in many places in an organisation but ideally should be managed by a specific data owner who is responsible for maintaining it, keeping it secure and ensuring management in line with the data protection policy. Only those responsible for the HR data should be able to change any aspect of the HR data itself (for instance changing terminology or definitions for specific HR indicators). Find out more in our data protection factsheet.
Correlation and causation
People analytics can help identify cause and effect relationships, by investigating the relationship between two sets of data to be investigated, and determining whether the relationship is correlational or causal.
Correlation is when two or more things or events happen around the same time which might be associated with each other, but they aren’t necessarily related in a cause-effect relationship. It implies a mathematical relationship between two things which are measured, and is often described numerically with a value between 0 and 1, where 0 is no relationship and 1 is a fully predictive relationship. For example, there is a 0.01 correlation between eye colour and height, so knowing someone’s eye colour does not mean you know how tall they are. They are virtually independent. But there is a 0.8 correlation between smoking and incidence of lung cancer, so it's possible to say that smokers are more likely to develop lung cancer. However, this doesn’t necessarily imply smoking causes lung cancer in every case.
Causation is when one event or thing happens and as a result of it happening, another event or thing happens. If the first event did not happen, then the second does not happen. There is not a mathematical/probabilistic relationship between the two, but instead a time-based cause-effect relationship.
Just because two things correlate doesn’t necessarily mean there’s a causal (or cause-effect) relationship between them. Other factors are likely to influence relationships between two things as no organisation is a closed system. Therefore, it’s important to analyse as much data as possible before drawing conclusions.
Who is responsible for people analytics and managing people data?
It varies. Large organisations may have a centralised people analytics team that provide insights to stakeholders in the organisation. Some organisations prefer a decentralised approach where individual HR analysts within small centres of expertise provide insights within their specialist domain. Others prefer to outsource their analytics. In practice, organisations usually take a hybrid approach.
Although data is held in many places in an organisation, it should ideally be managed by a specific data owner. The data owner is responsible for ensuring that data is maintained and kept secure according to the organisation’s data protection policy. Only those responsible for the people data should be able to change the structure of the people data itself, such as the definitions for specific HR indicators. Employees and managers can view and update some of their personal data through self-service. Find out more about data protection in the UK.
What are the aims of a people analytics strategy?
People analytics projects should align to both the business and the HR strategy. Solving a critical business issue is likely to create the most value for the business and create further demand to create insights from people data.
A people analytics strategy should have three aims:
Connect people data with business data to inform business leaders and help them make decisions.
Enable HR leaders to use insights from the analytics to design and implement appropriate HR activities.
Measure HR’s effectiveness in delivering against its objectives. A sizeable minority of the people profession find this part challenging. Almost a quarter of respondents to our People Profession Survey 2020 said that they don’t have clear measures of success for measuring their impact.
Our practitioner’s guide explores the first steps to building a people analytics strategy, developing simple analytics capabilities. In our research report Human capital analytics and reporting: theory and evidence, we summarise key academic concepts to apply in a people analytics strategy.
What is a people analytics process?
The people analytics process should follow nine steps from planning through to evaluation. In practice, the process can be shorter. For example, a recent data audit can be reused, or when analysis and reporting have been automated.
Plan: Develop the goals and purpose for the analytics activity. Map the requirements of the customer and plan questions/queries which will be answered by the analytics process.
Define critical success factors: Define the measures that will show if the project has been a success. Examples of what these can be based on include: delivery on time, impact of project, feedback from users..
Data audit: Map the data which is currently available and grade its quality. This will illustrate where any gaps in data may be, which should be filled before progressing.
Design the process: Define roles and set objectives for team members. Define resource requirements and map stakeholders for the project.
Design the data collection strategy: Design the collection and processing stages of the analytics activity.
Data collection: Collect data from existing data sets (for example, absence records) or collect new data (for example, by running an engagement survey).
Analyse data: Analyse data and create insights, in line with the stakeholders’ requirements.
Report data: Report a solution to the problem clearly and recommend further areas of investigation if needed.
Evaluate: Review the process and evaluate impact. Update process as required.
Books and reports
EDWARDS, M. and EDWARDS, K. (2016) Predictive HR analytics: mastering the HR metric. London: Kogan Page.
KHAN, N. and MILLNER, D. (2020) Introduction to people analytics. London: Kogan Page.
MARR, B. (2018) Data-driven HR: how to use analytics and metrics to drive performance. London: Kogan Page.
Visit the CIPD and Kogan Page Bookshop to see all our priced publications currently in print.
BASKA, M. (2018) Six ways analytics will future-proof HR. People Management (online) 6 June.
GARCIA-ARROYO, J. and OSCA, A. (2019) Big data contributions to human resource management: a systematic review. International Journal of Human Resource Management (online). 9 October. Reviewed in In a Nutshell,.
GREASLEY, K. and THOMAS, P. (2020) HR analytics: the onto-epistemology and politics of metricised HRM. Human Resource Management Journal, Vol 30, Issue 4, November. pp494-507. Reviewed in In a Nutshell.
JEFFERY, R. (2019) Amazing insights you can learn from people analytics. People Management (online). 21 February.
RASMUSSEN, T. and ULRICH, D. (2015) Learning from practice: how HR analytics avoids being a management fad. Organizational Dynamics. Vol 44, No 3, July-September. pp236-242.
CIPD members can use our online journals to find articles from over 300 journal titles relevant to HR.
Members and People Management subscribers can see articles on the People Management website.
This factsheet was last updated by Hayfa Mohdzaini.
Hayfa Mohdzaini: Senior Research Adviser
Hayfa joined in 2020 as the CIPD's Senior Research Adviser in Data, Technology and AI. She started her career in the private sector working in IT and then HR, and has been writing for the HR community since 2012. Previously she worked for another membership organisation (UCEA) where she expanded the range of pay and workforce benchmarking data available to the higher education HR community. Hayfa has degrees in computer science and human resources from University of York and University of Warwick respectively.
She is interested in how the people profession can contribute to good work through technology.