Many HR leaders feel pressure to do advanced analytics. They may have had some success with better reporting and basic analytics thanks to improvements in HR technology. They might have even done some advanced analytics on turnover, but that's an isolated success. They feel they should be able to somehow use "big data" to help with critical business issues like performance management, but how to do this is unclear.
Here's the reality: For most companies, extensive use of advanced analytics is years and millions of dollars away. What you can do now, with your existing team and technology, is use basic analytics to make your performance management efforts more effective.
Analytics for the average HR department is a lot of small, fairly simple projects to gather and analyse data to answer specific questions. The important thing is going from an opinion to data-driven evidence. More precisely, you are engaging in gathering the best available information to test a hypothesis that will lead to a decision.
Here's how you can get started.
An example of basic performance management analytics
Analytics always starts with a business issue and an intuition of some problem or opportunity. Perhaps you have the intuition that your performance management process isn't working well. The next step is to take that vague sense that it isn't working and narrow it down into a series of more specific hypotheses:
- Managers aren't submitting appraisals on time. What do you need to go from that sense that many are late to a more specific number? If you count how many are late, you'll end up with data that tells management how big the problem is, whether it's getting better over time and which departments struggle the most. This data will help inform a decision on whether to invest in new processes or technology. Rocket science? No. Useful? Yes.
- Quality of the appraisals is low. To prove this hypothesis, create a sheet of criteria or standards for what makes for a good appraisal. Collect a random sample of appraisals and then rate them against the criteria you've outlined. This provides data that you can share with management about how bad appraisals are, where the problem is most severe in the organization so they can made decisions on what, if anything, should be done. They might decide to provide training to the departments or managers who need support to improve their appraisals.
- Administration is a time-consuming nightmare. You need data that shows how much time is involved. You can do this by interviewing a random selection of managers. Help them make reasonable estimates of the time the process takes them. Add up the estimates to prove or disprove your hypothesis. If the numbers indicate a problem, it will help management make decisions about investing in a better approach.
An example of mid-level performance management analytics
Having dealt with some of the table stakes of whether the performance management process is working reasonably well, you might start to ask more sophisticated questions.
If the business issue you need to address is raising the quality of performance management, you might propose the hypothesis that there are certain groups in the organization where it's working well. If the hypothesis is correct, you can study those groups and extract lessons to apply across the enterprise.
The first step is to decide what factors might define a group. Is it age, gender, job type, personality? There are endless cuts of the data you could do. And if you're sitting on a big budget, you might slice the data hundreds or thousands of ways to find the clusters. However, if you're an average HR department, rely on your intuition to pick a manageable number of factors to analyse.
You may have to go on a bit of a data safari. If you think the group factor is age or department, it's easily found in your HRIS. But if you think personality is the driving factor, you may have to look in another system or generate that data by doing assessments on a randomized sample of managers. Do what makes sense within the limits of your resources.
Analysing performance management is important
It's also hard and that means anything you can do to make the process better is worth the effort. You don't need to win the Nobel Prize for your analysis, you just need some data that will help the organization make improvements.
I've done workshops around the world on analytics and a key finding is that the average HR department doesn't need help learning statistical software packagers or AI (artificial intelligence) tools. They need help clarifying the business issue so they know where to look for data to help inform decisions.
When you start looking for data, only some of it will be in your HRIS or talent management software. The data you need is somewhere out in that vast savannah of your organization and you need to go on a data safari to find it.
Armed with a clear understanding of the problem, and a willingness to go on a data safari, the average HR function can do analytics to improve performance management.
Bit by bit the world of bigger data and more advanced analytics will creep into the life of the average HR department. Let's not wait for that. You have the capability already to do useful analytics that will lead to better decisions about HR processes, even for really tough cases like the analytics of performance management. You can do it, you probably already are doing some of it, now just scale up that capability.