Analytics is the business function that has developed to extract meaning and value from the large and growing amounts of data that are routinely generated and stored by the IT infrastructure that organisations use in their day to day operations.
There is a widespread consensus that the HR function has made less progress in using analytics than other areas of management.
HR maturity models offer a useful starting point in understanding why this is. Perhaps the best known of these has been developed by Josh Bersin, who conceptualises HR analytics developing in four stages:
- basic reporting
- advanced reporting
- strategic analytics
- predictive analytics
Basic reporting
Basic reporting involves interrogating databases that capture information on aspects of HR and people operations, for example a payroll database, to provide insight for managers.
This could involve building a picture of exactly how many people on different job grades different business units employ, perhaps to benchmark costs or to understand future training and workforce requirements.
There is evidence that many organisations embarking on an HR analytics journey have not progressed much further than basic reporting. We can identify two key reasons for this.
First developing the capabilities to do basic reporting well is hard. IT for HR operations is set up to do things (e.g. to make sure staff are paid the right amount on the right date). As long as the basic operation is carried out successfully, people tend not to worry about the underlying quality and structure of the data.
This means that one business unit may use a payroll database in a different way to another business unit, putting information into different database fields or recording different types of information in these fields.
These differences multiplied over many different business units is what makes basic reporting so tricky. Organisations have to expend a lot of effort and resources identifying and fixing these differences before they can do any meaningful reporting.
They also need to put in place protocols for data governance to stop these problems recurring.
Second, managers can be naïve in the data they ask to be reported. Instead of using data to inform decision making, data may be requested to confirm or legitimise decisions already made. Managers may also request too much data, resulting in the production of large spreadsheets of reporting results.
The problem here is that it is difficult to work out which data is important and why. As a consequence, organisations may get little value out of the effort they put into basic reporting, so don’t invest in doing advanced reporting.
Advanced reporting
Advanced reporting has two purposes.
- First, to automate reporting to make it quicker and easier to produce analysis and reports.
- Second, to make data easier to interpret
Business Intelligence (BI) software allows the results of automated analytics to be displayed in easy to understand visual formats which make it easier for managers to interpret key data and to understand how they are performing against comparable business units.
Adoption of advanced reporting using BI software has been limited, but is growing. Organisations seem to be focusing advanced reporting around key HR metrics of employee attrition, sickness and absenteeism and employee engagement.
Advanced reporting allows managers to easily identify if these metrics are out of line with other areas of an organisation, leading to quicker diagnosis of problem areas to save costs and improve morale and productivity.
Organisations that have adopted advanced reporting tools claim that the initial costs are quickly off-set through lower attrition related costs, improved morale and productivity.
Integrated talent management (ITM)
Another approach to advanced reporting comes through the adoption of integrated talent management (ITM) software suites.
This involves replacing a range of different software and IT tools that are used for different aspects of HR operations (recruitment and selection, on-boarding, performance management, learning and development and payroll) with a single software suite that does everything, and which will typically also integrates with wider enterprise resource planning software.
The process of migrating from multiple pieces of software to a single ITM is often a costly and time-consuming. It may make sense for large organisations that are likely to secure productivity and cost savings from the standardisation of HR operations.
The ability to do advanced reporting is then a happy bonus. If organisations simply want to move to advanced reporting, the use of BI software is likely to be an easier and more cost effective method of getting there.
Strategic and predictive analytics
At the heart of strategic and predictive analytics is the search to identify cause and effect relationships.
- How do people contribute to business performance?
- What factors cause employee attrition to increase?
Once cause and effect relationships have been identified, the results of the analysis can be used to make predictions about the future. Which employees are most likely to quit in the next 12 months? What would happen to business unit performance if staffing arrangements were changed?
The results can then be used to help managers to manage better, for example to better understand how an increased bonus might change the chances that a key employee will quit, or how increasing employee engagement in a business unit is likely to improve the performance of that business unit.
At a (relatively) simple level, strategic and predictive analytics can be undertaken by researchers with advanced (MSc/PhD) level quantitative skills in psychology, economics or sociology.
Data scientists with a background in computing are increasingly moving beyond this approach to use machine learning to automate and repeat the sort of analysis that social scientists do through bespoke analysis.
Technically, none of this is particularly hard for an organisation that has already developed the capabilities to do advanced reporting.
However, evidence suggests that very few organisations have progressed to using their data to understand the cause and effect relationships that underpin business success.
What aren’t organisations making more progress in this area?
The reasons for this lack of progress seem to be threefold.
First, HR leaders are not convinced that the human factors in their business can be modelled quantitatively. Second, there is a shortage of people with the right technical skills to undertake such analysis.
People working at the coalface of HR analytics are typically very good at programming database queries and extracting data for simple analysis in spreadsheet software, but have less understanding of the principles and practice of modelling complex cause and effect relationships.
Third, people with the right technical skills may often lack the business insight and influencing skills to develop analytical models that are trusted by managers. Without this trust in the models, managers are unwilling to use the results in decision making.
Three predictions about the future
In the light of this short overview of the HR analytics landscape, it is possible to make three predictions about the immediate future of HR analytics.
- Advanced reporting through BI tools is likely to grow in popularity as organisations come to appreciate the value they can get from it.
- Adoption of ITM software as a means to achieving HR analytics is likely to be more limited because of the high costs of migrating to an ITM suite. Some organisations will do it but for operational reasons; improved advanced reporting capabilities are a useful side effect, not the main purpose of ITM suites.
- The spread of strategic and predictive analytics is likely to be slower. Although the small number of organisations that have already developed the capabilities to do advanced and predictive HR analytics have found the results to be transformative, development is likely to be limited by a shortage of people with the skills (technical and managerial) to make it work.
