Low Scores
Data from the online test, “How Good Is Your Employee Engagement Survey?,” reveals that only 22% of companies are actually getting good results (i.e., increased scores) from their employee engagement surveys.
Given that most companies conduct surveys annually, why are so many failing to actually increase employee engagement? One of the big reasons is that most companies work on the wrong issues. More specifically, most companies look at their lowest survey scores and assume those are the issues most in need of fixing. But that’s a big mistake.
Imagine a company conducts an employee survey and asks people to rate leadership, corporate values, company culture, employee recognition, and all the other issues that typically get assessed. Then imagine that, on a whim, executives decide to also ask employees to rate the corporate logo, the color scheme, the food in the cafeteria, and the layout of the employee parking lot.
Finally, suppose that when the survey results are compiled, those issues represent the lowest scores. Employees absolutely hate the company colors and corporate logo, and they’re not fans of the cafeteria food or the parking lot design.
Those are the lowest scores, but do those issues meaningfully impact employee engagement? If you could only choose between training managers to better support employees or improving the food in the cafeteria, which of those issues is going to help you retain your star employees? Or turn middle performers into high performers?
The problem with focusing on your lowest scores is that those issues might not actually impact an employee’s willingness to recommend your company or give their best effort at work. To know which issues actually do predict and drive engagement, you need a statistical tool like multiple regression analysis.
Multiple regression is a statistical analysis for predicting the value of one dependent variable based on multiple independent variables. Put more simply, it’s a way of figuring out which issues are actually the biggest predictors of your employees’ engagement.
Here’s an analogy to help explain multiple regression. Imagine that you’re choosing a restaurant for dinner tonight; what factors go into your decision? Maybe 50% of your decision will be the cuisine (e.g., do you feel like steak or pasta). Perhaps another 30% of your decision will be how far you have to travel (e.g., do you want someplace close to home or are you willing to drive 30 minutes to get there). The ambiance could be an additional 20% of your decision (e.g., are you looking for someplace quiet or boisterous). And so on.
With that restaurant analogy in mind, let’s look at an actual multiple regression analysis for a company’s employee survey.
Multiple Regression Analysis For Employee Engagement Survey
This analysis says that over half (53%) of an employee’s willingness to recommend this company is driven by whether their direct leader encourages and recognizes suggestions for improvement. On a survey that had dozens of questions, this regression analysis revealed that one single question literally predicted more than half of employees’ engagement. In other words, if the company fixes this one issue, they’re virtually guaranteed to see a huge spike in employee engagement.
An additional 21% of an employee’s decision to recommend this company as a great employer is driven by whether they understand the rationale behind the company’s business decisions. Now we’ve got two questions, both easily fixable, that drive the vast majority of employees’ engagement.
Managers don’t have infinite time to work on employee engagement, and they’re far more likely to prioritize traditional KPIs like quality, efficiency, profit, expenses, etc. But when you can statistically distill their entire survey into one or two simple issues, you’ve made it much more likely that they’ll take quick action and see immediate results.
There’s one other essential point; every company’s regression analysis will be different. The factors that drive employee engagement at your company will inevitably be different than your competitors’ issues. So you can’t just pick an issue at random; you truly need to analyze your unique workforce.
