Research

Employee Engagement is a Tricky Predictive Metric

By Greta Roberts, Mar 15, 2016

In my day job, my work focuses on using predictive analytics to decrease employee turnover or increase employee performance. We work with our customers to define the problem to solve, and then together identify useful data inputs to include in predictive model development.

One topic that frequently comes up is employee engagement data, and whether it is meaningful to the analysis. There can often be years of employee engagement data, and the data is typically in HR’s control.

How Valuable is Employee Engagement Data in Predictive Work?

Businesses and their HR teams have long measured employee engagement, “believing” it is a leading indicator for business performance. Hundreds of millions are spent, annually; conducting engagement surveys, and creating follow on programs that target areas that measure low in employee engagement.

It has to be satisfying to see reports that highlight areas in the organization with and without engagement. Engagement reports often provide a false sense of control that engagement, training and other employee development programs will lead to an increase in actual, real business performance.

It seems that executives and their HR teams, at the highest levels, have bought into the myth that high engagement leads to high performance. I say myth, because of the limited number of companies (and the consultants that work with them), that have done their own research proving this connection.

Employee engagement data definitely has value. But it needs to be used for the right thing. Use this data consciously – perhaps in HR programs. We’re suggesting employee engagement data is best kept out of the predictive domain.

What follows are 6 reasons we believe employee engagement data is best kept out of the predictive domain.

  • Engagement is not Performance – Many make the assumption that employee engagement drives actual business KPIs, but usually don’t go so far as to prove it. We could build a model to predict engagement, perhaps based on management methods or compensation or demographics. But what does it really do? It predicts the outcome of a survey.

    It is just more rational to go the rest of the way, to predict actual KPI performance. The Line of Business manager can’t eat engagement.

  • Dangerous Correlations - What if you studied engagement as a function of demographics, and discovered differences in engagement by race, gender, or other protected categories? Likewise if you found correlations for zip codes, which are seen as a proxy for the above. What is the business outcome you’d suggest - hire more or less of a certain group? The real world is more complicated than that.

    There may be a well-intentioned reason to understand engagement differences in the existing workforce, but this is more of a reporting issue than a predictive issue. And, cultural differences and political expectations in different subgroups can easily drive different engagement results.

  • Engagement is Self-Reported - Some work cultures are open enough to accurately gather honest engagement feedback from employees. Many are not, and the fears in those cultures are likely to bias scores upward. As noted above, we can also expect that different cultures will tend to want to show “pleasing” results to management.

  • Feel-Good Metric - One explanation for the prevalence of employee engagement metrics is that it makes managers feel good. Who wouldn’t want a team of workers under them that are highly driven to the same objective?

    There is a disconnect between the well-intentioned goals of engagement survey creators, and the mindset of many managers who use the tool. As with performance, we have no proof of a link between upper managers feeling great about their employee engagement, and the performance of that business.

  • A Messy “Middle Value” - Implicit in the employee engagement exercise, we assume that (a) people and processes drive engagement, then (b) engagement correlates or even drives KPI performance. Why not just skip the middle value, skip the assumptions, and predict KPI values directly?

    Engagement work can also assume that the linkage between engagement and performance is the same in all roles and lines of business. This is not going to be true - engagement might drive performance in sales, but not in shipping. Or vice versa. The point is that we don’t know, and can’t assume.

  • Get out and Meet your Line-of-Business Customers - Engagement is an HR measure. Useful to HR purposes. It’s alluring because the data is right there inside of HR. For better or worse, actual employee performance happens and is measured in the Line-Of-Business (LOB) and the LOB systems.

    Employee performance reviews are also notoriously unreliable measures of performance - there is no variance - everyone “meets expectations.” Real performance is multi-faceted and much more variable.

In most companies, this means we need to leave HR, go across the building, and talk with an LOB manager to get the KPIs. Business KPIs are where the rubber meets the road and where your predictive project will finally gain supporters and more resources.

If we don’t go all the way to predicting the KPIs, we are proving nothing and not adding predictive value.

Employee Engagement Measures can be Useful – but Using Them as Inputs to Predict KPI Performance is Tricky

I’m not negative on employee engagement scores for any reason except that businesses and their HR teams have been led to believe high engagement equal business success. And they often don’t know or haven’t proven it. I actually applaud organizations in years past for trying to find and measure metrics that could lead to a prediction of future business performance.

Employee engagement was a good try but ultimately doesn’t reliably and repeatedly show a strong connection. Current predictive analytics methods and approaches and better systems and data make it possible to move beyond this “middle measure substitute” to predicting real business performance.

Employee engagement surveys can be useful when used the correct context – and “right sized” in terms of their importance to the organization. If they are used to gain a general measure of sentiment then they have value.

Predicting KPI performance is another issue, where there are more direct measures with a lot less bias and are more efficient in their predictive capabilities.

About the author

Author photo

Greta Roberts is an influential pioneer of the emerging field of predictive workforce analytics where she continues to help bridge the gap and generate dialogue between the predictive analytics and workforce communities.

Since co-founding Talent Analytics in 2001, CEO Greta has successfully established the firm as the recognized employee predictions leader, both pre- and post-hire, on the strength of its powerful predictive analytics approach and innovative Advisor™ software platform designed to solve complex employee attrition and performance challenges. Greta has a penchant for identifying strategic opportunities to innovate and stay ahead of the curve as evident in the firm’s early direction to use predictive analytics to solve “line of business” challenges instead of “HR” challenges and model business outcomes instead of HR outcomes.

In addition to being a contributing author to numerous predictive analytics books, she is regularly invited to comment in the media and speak at high end predictive analytics and business events around the world.

Follow Greta on twitter @GretaRoberts.


Tags

Six Models for Organizing Analytics

Best Practices

Why organize for analytics? Because companies that think strategically about analytics achieve the biggest impact from analytics. See which model is right for your organization.

Read the ebook »