The most exciting stories of people analytics involve big data yielding surprising findings that deliver an enterprise-wide impact. The least exciting stories involve essential, but basic metrics such as dashboards with accurate turnover data. However, these stories from the extremes (advanced and basic) completely overlook a crucial area: mid-level wins.
Dozens of analytic projects for the mid-level can be done every year. They directly address short-term business decisions by answering questions such as, “Is there any evidence this training is reducing accidents?”, “Should we continue to spend money on this job board to hire engineers?”, “If we need to cut payroll costs by 10% in two-years, to what extent can we rely on attrition to bring down headcount?”
The reason mid-level analytics projects are so important is that these reliably generate a string of successes that show stakeholders the worth of the analytics function. Think of what happens when a new department gets a big budget but fails to produce a clear ROI. After the initial hype has passed, other VPs start looking hungrily at that budget and suggest that some of it could be better spent elsewhere.
With mid-level wins, HR can sustain its analytics department so it can proceed with the essential basic analytics and the business-changing advanced analytics.
How to Succeed with Mid-Level Analytics Projects
The most crucial element for getting mid-level wins is to start with a clear business decision that needs to be made in the short term. Contrast that for a moment with basic and advanced analytics.
Basic analytics provides data everyone agrees is essential, but it isn’t usually tied to a specific, short-term business decision. Advanced analytics can often be like a R&D project where you hope to discover something new and important after weeks or months of work. It may or may not be tied to a well-defined business decision, and it rarely will be a short-term one since these projects can take some time, especially without the right analytic solutions.
The second crucial element for mid-level wins is being comfortable working with imperfect data. Perhaps you are doing an analysis on how to improve wellness. It probably won’t be certain which of three interventions will have the biggest impact, but if the data indicates that one choice looks the best, well that’s really helpful for management; it’s far better than making a decision with no data.
The final crucial element for mid-levels wins is having a high level of skill in presenting the findings so that they are taken seriously and so that it is seen as an analytics win. Often this skill is lacking in analytics professionals who focus on the numbers and are not as attuned to the people dynamics that are inevitably part of any project.
Good Insights Don’t Require Perfection
If HR wants to have a string of meaningful successes in analytics then they need to embrace the mid-level. The data doesn’t have to be perfectly clean or require a massive data set – it just needs to provide insight. You can pull together everything you need for mid-level wins in a few weeks with the right technology and start building success from there.
The key insight is that some data is almost always better than no data. If a manager complains that employees keep coming in late then it will be useful to know, at least roughly, how many employees and how late. Perhaps you can get precise and detailed data; perhaps you can only get an estimate. In either case, the action you take will be based on some data not just a sense or a rumour or an opinion.
The amount of effort a company puts into getting data for mid-level analytics depends on how critical the problem is and how much time they have. The trick is to put in an effort appropriate to the situation and refuse to let perfect be the enemy of good.
Embracing All Three Levels of People Analytics
HR should embrace all three levels of people analytics; each has a role to play in improving decision making. Where HR departments get into trouble is when they don’t make a distinction between the levels. If they think the only problems worth tackling are advanced ones then they’ll miss a huge number of mid-level successes. If they are obsessed with basic reporting then they’ll be forgoing the insights advanced and mid-level analytics can bring. Keep the three levels distinct and spread your investment in analytics across all three.