About » Thinking » Top 10 Tips

Top 10 Tips for Designing an Employee Survey

Here are our tips for getting the most from your employee survey.

1. It’s got to be anonymous.

If employees think they will be identified they will skew their answers positively. Even if you can’t identify someone’s responses directly, the demographics can give them away (just how many male, 20-30 yr old, 3-5 year tenure, managers do you have in your marketing department?).

2. Keep the survey short.

You won’t get any more insight from 100 questions than from 30 – but employees will get fatigued and rush through the questions. Better to get higher quality responses on fewer questions. As an indicator, fewer than 5% of responders should fail to finish the survey once they have started.

3. Keep the questions short.

Responders struggle to absorb long questions and get frustrated. At WorkplaceDynamics we have a limit of 14 words per question, and most of our questions are less than 10 words. Over 90% of responders who finish the survey should have answered every “agreement” question.

4. Test your questions.

We’ve tested hundreds of questions that seemed good, but that fail to get a meaningful response. Either use a standard survey that has already been tested, or test your own questions on a subset of responders to ensure they are measuring what you think and showing good variance. Spending time on this up front will save your employees a lot of time later.

5. Get a benchmark.

Without a benchmark your results are actually dangerous. Decide if you are looking for year-on-year comparisons (so it will take a couple of years to get meaningful results) or for external benchmarks (where you’ll need to use a third party survey). Looking at one year percentage agree figures in isolation can get you focusing on the wrong thing (people are nearly always most negative about their pay)

6. Be wary of demographic comparisons.

Survey scores are heavily affected by the demographics of the responder. People with higher salaries, for example, will score the survey higher on average. This means that department comparisons are skewed – your legal department will nearly always come out better than your mailroom. To compare departments fairly you really need to allow for demographic bias.

7. Craft your open-ended questions carefully.

Simply asking what people like and dislike about your company will yield the answers “the people” and “the pay.” Craft questions that get to your objective – such as “How could this company help you do your job better?” Our online survey is designed to ask different questions based on how the responder has responded so far.

8. Don’t get bogged down in statistics.

Correlations, covariance, Cronbach’s alpha, confidence levels… there is never-ending scope to get statistical about your survey. With tried and tested questions and a good benchmark you can avoid all that and spend time on what is most important – listening to what employees are actually saying.

9. Take on the scary topics.

Your employee survey is your big chance to get your employees’ help on making the company better. By taking on the scary topics you are making a statement that the company is willing to fix things. If you are unwilling to ask about topics such as strategy, leadership and ethics then it’s better not to ask anything at all.

10. Over-communicate.

Before during and after the survey. Tell people it’s coming and that it is being taken seriously; remind people to take the survey and show them how the response rate is going; present the results openly to the senior leadership team and the company as a whole.