The business community is not always careful in its use of language; often more concerned with fad and fashion than carefully crafted policies. ‘Talent’ is the worst recent example, as it can mean anything from the whole workforce to a select group of high potentials. The worry is that ‘HR analytics’ will go the same way. At conferences it seems data management, reporting, and basic statistics as well as what might be described as genuine HR analytics (HRA) are all included under the same heading.
In a review of HRA theory and practice we discovered that when researchers defined ‘analytics’ they focused on prediction and making connections between HR activities (and/or workforce data) and business outcomes. The expert users made the same points, but also emphasised the role of HR analytics in improving business decisions and driving action.
If to the aficionados HRA is a way of solving people-related business problems, often using data drawn from multiple sources, then it should be distinguished from collecting and presenting data through dashboards and the like. This is not to suggest that these tasks are unimportant. As shown in the figure below, the expert practitioners pointed out they only spend a fraction of their time on full-blown analytical tasks.
My paper for the Institute for Employment Studies suggests that to overcome the ill effects of linguistic confusion we might be better describing ‘predictive’ HR analytics to mean the process by which we use multiple data sources (people and business) as a business problem-solving or risk-mitigation process. We can then leave straight HR analytics to cover data management and reporting, including the simple correlative statistics useful for demonstrating the present workforce situation or describing the recent past (like retention, attendance or recruitment effectiveness).
Predictive HR analytics can then deal with the wicked business challenges; externally driven ones such as:
- What impact will Brexit have on our ability to hire the right staff?
- How would we respond to a ban on zero-hours contracts in meeting our resourcing needs?
- If there were a change to taxation of flexible benefits how could we best change our reward package so that it remains attractive and cost efficient?
Or internally driven problems such as:
- What management actions are most likely to create an improvement culture?
- Where are the areas in the workforce where productivity might grow with the least financial outlay?
- Can we realistically have a zero accidents target and be taken seriously?
To be able to tackle such questions you are likely to need investment in capability (skills in scenario planning and data modelling, and an understanding of causation); data quality; technology to manipulate the data; and good management processes. The latter may be the least obvious pre-condition of success, but the organisations that are good at this stuff ensure they have the time and space to address big questions and are not distracted by urgent but technically simple tasks. They have both effective prioritisation systems and ‘intelligent’ customers who have been trained to be largely self-sufficient in analytical terms so that they only raise complex issues – and ones that there is a chance of successfully answering. This requires managers to have readily accessible data that is easily manipulated so that they can do their own basic analysis, and HR business partners who are on hand to filter the demands on the HRA team’s time. Switched-on business partners can perform the essential task of defining questions in such a way that analytical processes can be brought to bear to answer them.
So it is all about focus and clarity in what you are trying to do with limited resources. Defining your terms correctly and communicating them helps achieve that goal.
Peter Reilly is the author of The path towards predictive analytics, published by the Institute for Employment Studies
