What many businesses forget is that the most important employee to listen to is the employee that isn’t. I don’t mean “the employee that isn’t” in any zombie Halloween sense, but much more mundanely: The employees that join a company, and the ones that stay, are not representative of the ones that don’t. And there’s good reason to think that the employees who aren’t with you can give a better insight into your company’s shortcomings than the employees who are.
In my book Thinking Statistically, I talk about the problem that statisticians call “Selection Bias”—namely, that if we take a non-representative sample of a broader population and pretend that it is representative, then any inferences we draw from it will be badly flawed. I give the example of feedback systems in companies and other organizations: Where employees sense a danger, rightly or wrongly, from giving honest negative feedback, if they are asked (say) “what do you think of the new project?” they are liable to share only the positive half of their views on the topic. The sample of opinions that reach senior management will then be unrepresentative of the employees’ true opinions, and six months later the project collapses and senior management thinks “it seemed like such a good bet!” while in another corner an employee mutters under her breath “I knew this thing would sink for exactly this reason.”
The opinion that you’ll never hear
The “feedback effect” continues to operate, however, on an even higher level than the feedback that existing employees do and don’t share with their supervisors. Not only do certain opinions within an organization never reach the people at the top, certain people are filtered out of an organization in such a way that their opinions aren’t even present in the first place to be (potentially) ignored or engaged. Depending on how that filter operates, this form of selection bias can let a company disintegrate without even noticing its own disintegration.
For example, a good friend of mine worked for a consulting firm that was entirely reliant on its human capital for its success, yet instituted policies that needlessly made employees unhappy. Every bureaucratic hassle imaginable, every poor management practice, every way to irritate an employee without meaningfully benefitting the company had found a place somewhere in this organization. Of course, the employees who were least likely to put up with the hassles and the most likely to find great alternative employment included many of the best and most competent workers. More than annoyed at the way he was being treated, my friend was surprised that upper management wasn’t noticing the impact its decisions were having on employee morale and retention: he wasn’t saying that the company’s policies were somehow wrong or unfair, just that they were bad for business. The company was hemorrhaging competence through petty decisions that would hardly benefit anyone. Why couldn’t they tell?
The answer, I think, was selection bias. The information that reaches senior management is much more likely to come from people who are still at the company. People who have stayed at the company are a non-random sample of people who were ever employed by the company, let alone people who could have come to work at the company but turned down job offers (or never wanted to apply). The people who stayed are those who (for whatever reason) found putting up with the unnecessary hassles worthwhile; therefore, the information that filters back to the top says that the organization is doing fine, even when it’s being driven into the ground.
Inside the tent
One implication here is that organizations where disgruntled employees can’t leave (or have a hard time leaving) might actually have an advantage in terms of correcting their course if they move off track. Startups come to mind as an example: an early-stage startup employee with lots of equity and little salary will be loathe to leave even if the company starts making terrible, company-killing decisions. Instead of jumping ship, she’s more likely to do everything in her power to make sure her critiques get taken seriously. The executives might not be happy to have passionate critics stay within the organization, but the information those critics deliver might be exactly what saves the company from unnecessary mistakes. The question for other types of organizations, where disgruntled employees have an easier time leaving, is how to replicate the same effect.
Now, I’m not trying to claim here that statistical thinking gives a singular insight that’s unavailable through other means—there are other ways to think about the problem that will also reach the same conclusions. For example, the idea that group members who perceive deteriorating quality face a choice between voicing their grievances and leaving the group is the core of Albert O. Hirschmann’s classic book Exit, Voice, and Loyalty. However, statistical thinking does give one useful framing to the problem that you can use to find the right answers. For senior managers, the key question to ask yourself is “How can I ensure that information also reaches me from the people who have been ‘selected out’ from my normal information streams?”
One answer could be phoning former employees immediately after they leave the company and asking for their evaluation of what the company could be doing better—perhaps even what the company could have done that would have made them stay. Or it might even mean calling potential hires who decided not to accept employment offers in order to really understand why they chose to work elsewhere. Of course, these kinds of suggestions require personal and corporate maturity to execute; they are not easy things to do, because they require staring head-on at a problem when it’s easier to look the other way. But a difficult thing to do is (of course) very often worth doing: if you ask her nicely, the employee that isn’t can tell you truths about your business that you would never have heard otherwise.