Thursday, May 30, 2013

Is Everybody Happy?

Getting the most from Employee Satisfaction Surveys

Do you conduct employee satisfaction surveys? How did they go? Did you get useful information, and what did you do with that information? An even better question might be why did you conduct a satisfaction survey?

Some organizations conduct stand-alone employee satisfaction surveys. Others are getting the input and evaluation of a larger group of stakeholders (customers, donors, community partners, etc.) and include employees as part of a broader snapshot of how an organization is doing; how well it’s accomplishing its mission.

The quality of the responses you receive can depend on many things, some of which you really can’t control. The people most likely to respond are those that are very dissatisfied, or those who are very satisfied. Those in the middle aren’t often motivated enough to respond. As a result, the data can be a bit skewed.

There are a few things you can do to increase the likelihood of getting a representative sample of opinions. But the first step is knowing exactly why you want to do a survey in the first place. Have you gotten the idea that your employees are not happy? Do you have any idea why they might not be? Or, are you just taking a pulse and trying to be proactive? Are you committed to doing something meaningful and constructive with the data you collect? If you haven’t thought about these things, you probably shouldn’t be surveying your employees.

Communicating that the information you get is intended to be used to improve the workplace, and therefore not just doing it because you think you have to, or because an outside entity requires you to do so, may increase the response rate among your employees. Of course, once collected and analyzed, you have to provide some feedback, and make an effort to communicate that you heard their concerns and are committed to addressing them.

The way questions are worded is also important. You don’t want to ask leading questions, or ask respondents to rate statements that may lead to desired responses. An example of leading statement might be "My supervisor’s door is always open to me" and ask the employee to agree or disagree with that statement. At the very least, give a scale of agreement or disagreement that’s wide enough to encourage a more considered response.

Invite specific comments. A simple yes/no or agree/disagree isn’t always that helpful to you. For instance, receiving a "disagree" response to the statement of "our benefit package is sufficient" may not mean that employees think your benefit package stinks. If you invite an explanation to that response, you might find that your communications about your benefit package may need improvement because some simply don’t understand what you actually offer. Big difference.

Don’t lose control of, or lose sight of, your data. You don’t really want to know that Jane Doe made a particular observation or comment, but you do want to know that 5 out of 30 people made the same comment. If you’re doing the survey in-house, enlist your bean counter to help with the data analysis. Remember, averages are often of little real value. One high or low level response can throw off the number and be very misleading. A simple calculation of a median data point might be more useful. If you use an outside vendor to conduct or analyze the data, more sophisticated calculations will most like be done, assuming your response rate is high enough.
Allowing employees to self-select to participate in the survey is not the best method. Again, self-selection is likely to attract only those with very strong opinions one way or the other. For the highest reliability and validity, every employee should participate, or a random selection strategy should be employed. Offering some sort of incentive to participate can often be effective, as well.
The results should be communicated openly, honestly and completely. Even if they’re not what you really wanted to see. Doing so will go toward building an atmosphere of honesty, integrity and trust. Most importantly, you need to be prepared to actually do something with, and about, the results you receive. You also should be tracking progress and communicating any successes and failures along the way.
Knowing what your employees think about working conditions, pay, benefits and relationships with co-workers and supervisory staff can be beneficial and can give you direction on what may need to be changed, improved or discontinued. But, one of the biggest problems with employee surveys is the expectations the survey creates. For a survey to be taken seriously, employees must be convinced that it's being conducted for meaningful reasons. They need to believe that management wants to know what problems exist so problems can be resolved. Employees expect that management will do something about problems that come to light. Expectations created by a survey become a major problem if you do nothing with the information you receive. The damage that can be caused by what will be perceived as an empty effort will not easily be repaired.
In order to get the most out of any employee satisfaction survey:
  • Know why you’re conducting the survey, what you intend to do with the results
  • Choose your questions carefully. Make sure you really want to know the answers.
  • Make a carefully worded, but realistic commitment to address the survey's results.
  • Communicate the results, honestly and completely.
  • Follow through on your commitment to address the results
Employee satisfaction surveys are not for the faint of heart. You will see comments and observations that will be painful to you. But, if you’re committed to getting the pulse of your organization, you have to take the good with the bad and do so with your eyes open and a willingness to admit there may be weaknesses. Or to be positive about it, challenges that can be met.


  1. Last year one of our business units conducted an employee engagement survey. Among the "positive" statistics was that 60% of the employees felt management was trustworthy.

    While I think the employee/employer relationship will always make this a challenging statistic, I thought that just being above 50% was not a win. It's great to gather this information, but the management team has to do something with it.

    I know the first thing I did was meet with my management team, even though we were an outside division, and discuss employee relations. I wanted to make sure our trustworthiness was going to be perceived well about 60%.

  2. I agree, Rob. And finding out why the other 40% answered in the negative, is probably more important than knowing why 60% answered in the positive. Determining how to get that information is more challenging, but possibly more useful.