What do medical researchers, email marketers, Google engineers, and Obama’s 2008 campaign team have in common?

They all know the power of running tests to compare results. Medical research is built on testing. Email marketers regularly run tests to see which subject lines result in the most opens. Google engineers famously tested 40 shades of blue to see which one people liked most. Using the winning shade of blue resulted in an additional $200 million a year in ad revenue.  And testing different web designs helped the Obama campaign capture an estimated additional 2.8 million email addresses, leading to $60 million in donations.

But for all the power of analytical testing, these methods are not widely used in HR.
They should be. Testing can help crack a problem that plagues corporate America: low employee engagement.  One 2012 study found companies invest about $720 million annually to improve employee engagement—yet for all this investment, 70% of U.S. workers are not engaged at work, costing an estimated $480-600 billion a year in lost productivity.

This shows that a lot of that $720 million is being invested in employee engagement initiatives that don’t work. How can you determine what does work? By finding evidence of what does. By testing.

Here’s how and why to test your employee engagement program.

Why We Need Evidence-Based Employee Engagement Programs

Many HR practices and beliefs are based on outdated ideas about psychology and organizational design. For instance, hiring decisions are often made on the basis of unstructured interviews, which do nothing to improve the quality of hires.

We are not as rational as we believe ourselves to be, don’t make decisions as well as we think we do, and this affects our employee engagement programs. In fact, we are prone to:

• Overgeneralizing from a few instances of personal experience. Staff lunches on Fridays improved employee engagement at my last workplace, so it will definitely work here too! Maybe it will, maybe it won’t; one example of something working in one situation is no guarantee it will work in another.
• Assuming the evidence we have is all the evidence that’s relevant. Employee engagement is up 23%! Our initiative is working. More evidence is needed to definitively show that’s the case. It may be that previously engaged employees are feeling even more engaged, but previously unengaged employees are not feeling any more engaged.
• Overweighing factors that come easily to mind. Numerous employees have complained about X. If we fix that we’ll fix our engagement problem. X may be a symptom of Y, and addressing X may do nothing to address its underlying cause.

Professional judgment is important, but the nagging persistence of employee disengagement shows that all the HR expertise in the world isn’t enough to solve an engagement problem. It takes more than expertise: it takes evidence and evaluation.

How to Evaluate Employee Engagement Programs With Testing

Like all important business programs, employee engagement programs need to be tested and validated. Testing counters the problems of overgeneralizing, overweighing certain factors, and other mistakes we make due to unconscious biases.

Testing also eliminates the problem of HiPPOs--Highest Paid Person’s Opinion. There are many pitfalls to following a HiPPO without data backing it up and you want your employee engagement program to steer clear of them. You want your employee engagement program to be validated with data.

A/B testing is a simple way of learning what works and what doesn’t. Try an employee engagement program or initiative on one randomly selected group of employees and compare the results to the control group, a group of employees that did not undergo the program.
This type of testing can be done for nearly any employee engagement initiative. For instance:

• An employee appreciation program. Perhaps an “employee of the month” makes employees feel appreciated, which in turn makes them more likely to feel engaged. Or maybe choosing an employee of the month feels like favoritism, and makes other employees resentful, with a negative effect on engagement. Run the program for one group of employees and compare its impact to those who did not participate in the program.
• Social events. Perhaps obligatory social events strengthen bonds between employees—or perhaps stressed employees resent compulsory events on top of all the work they have to do. Compare the effect on one group and the control group.

Prepare yourself for results you find counterintuitive or just plain puzzling. For instance, one A/B test in Israel found that imposing fines for picking up children late at daycares resulted in more parents picking their children up late, not fewer. Parents had previously seen picking their children up on time as a social obligation, and implementing fines made it an economic transaction.

Similarly, Google found that a program awarding exceptional performance with significant financial rewards wasn’t a success. It celebrated money above other values, and ended up pleasing nobody.

This is why testing is so powerful—it will show not only where your employee engagement initiatives are successful, but where they have the opposite effect as intended. Without testing, how will you know when you’re wasting time and energy on something that makes your engagement problem even worse?

Think Beyond Employee Engagement Initiatives

Once you begin testing to see what improves employee engagement, you’ll realize almost everything is fertile ground for experimentation. Consider testing how these variables affect employee engagement:

• Physical environment. Do employees with more light in their workspace feel more engaged?
• Diversity. Do more diverse teams have higher or lower engagement?
• Technology. Does technology make employees feel more connected to one another, or do these tools become a burden?

As you start to test different variables for their impact on employee engagement, you’ll see that everything at a company affects employee engagement in some way—and you’ll have data to prove it.

Done right, this will help move your employee engagement initiatives away from one-off programs to a way of thinking that is built into everything the organization does, from traditional HR functions to the physical environment to technology.

And at the very least, it will help you steer resources away from initiatives that harm employee engagement to those that will help it.