Home > Writing Advice & Resources > Building Effective Customer Surveys

Building Effective Customer Surveys

Updated Feb 2006
Well-designed customer surveys can yield valuable information for your business. However, a poorly worded survey can set you marching off in exactly the wrong direction. Below are some tips on designing surveys to get reliable, useful data.

Keep it Short
To increase the completion rate, keep your survey short. This means that you must do some serious thinking about what it is you really want to know.

Ask Questions that Have Meaningful Answers
A recent survey I received from a local fitness facility included this question: “on a scale of 1 (poor) to 5 (excellent), how do you rate the friendliness of our staff?”

This question is essentially meaningless. This fitness center has twenty staff members, some friendly, some not. To award an “excellent” ranking would be to ignore the behaviour of one or two employees. To award a “poor” ranking would be to ignore the excellent service given by most of the staff. There is literally no way to answer this question in a meaningful way. Faced with such a choice, many people will split the difference and award a “3” (average) ranking, which totally masks the true situation. No matter what ranking is given, the reality of the situation is never uncovered.

In the same way, questions about conditions that vary over time (“Which road is busier, Calgary Trail or 109th Avenue?”) can’t be answered meaningfully. You must ask questions specific enough that unambiguous answers are possible. (e.g. “From 7-9 AM on a typical Saturday, which road etc.….”)

Avoid Double Negatives
Compare these two questions:

“I do not want to lose my long distance dialing privileges.” Yes/No
and
“I want to keep my long distance dialing privileges” Yes/No

The first question is confusing. Some people will answer with a “no”, meaning, “no, I don’t want to lose long distance dialing”, other will interpret “no” to mean the question is false—in other words, they don’t mind give up long distance. The resulting data will be muddled and invalid. By contrast, the second question is unmistakably clear.

Cover All Bases
Consider this question:

I’d like to see more green widgets in stock. Yes/No

What happens if the person is indifferent to green widgets? They’ll probably never want a green widget…. but they might, someday—so just to be on the safe side, they may put “yes”. So you get a result like this:

70% yes
30% no

…and based on this, you wind up stocking a lot of green widgets that don’t sell. Always provide for the possibility that the person may not care either way about an issue. (Exception: employee surveys. Think twice before giving employees the option of not caring.) In the example above, adding a simple “don’t care” option, would give a much clearer picture of the true level of support for the new product.

And remember: “don’t care” is a valuable opinion. Why waste time changing the color of your restaurant walls when 85% of diners don’t even consider it in deciding whether to come back? Spend you energy and money on things that DO matter.

Guarantee Anonymity
Asking someone to complete a survey in front of you puts them on the spot. It’s like asking your best friends what they think of your website; you’re unlikely to get the full truth. So let your survey respondents be truly anonymous. Provide a locked box or similar container that they can put completed surveys into.

Test Your Surveys
You probably know at least one person who’s a great Devil’s Advocate. Hand this person your survey and ask them to spot any ambiguities or shortcomings. Revise the survey to address any concerns, then have a small number of customers complete the survey. Tally up the results. Do the answers raise questions in your mind? Revise the survey as needed. Only when the questions are clear and the answers meaningful is your survey ready to launch.

Interpret Results with Caution
Here are some points to take into account when interpreting survey data:

• the fewer completed surveys you have, the more unreliable the data. If you’re contemplating big changes, get lots of people to take the survey first
• people who elect to take a survey may have a very different outlook on your company than the ones who skip it
• if you do an on-line survey, the results will not represent the portion of your customers who are not Internet-comfortable