You know you’re a good writer. You’ve written blog posts on all kinds of topics, racking up tons of views. You’ve written articles that break down complex information so well that a first-grader could understand it. You’ve written speeches and video scripts. Now you need to write questions for a customer survey. That should be easy, right?

Wrong. It’s really, really hard to write a question that will generate valuable customer feedback. And, if you get it wrong, the information gathered from that question will be worthless. There’s more to the art of writing surveys than I can tackle in one blog post, so I’m going to focus on the biggest problem: When you think you’ve written a yes/no survey question, but respondents see it as open to interpretation.

The fatal flaw of a customer survey question that depends on the respondents’ interpretation

I call these “double-barreled” questions. They’re usually of the true/false or agree/disagree type, and the problem arises when you try to interpret the negative answers. Here’s an example:

“I don’t have a dog because I’m allergic to animals.”australian shepherds

If a respondent chooses the “false” answer, then you can conclude…what, exactly? Stop and think about that for a minute. What does a negative answer tell you?

I can think of at least three possibilities:

  • The respondent isn’t allergic to dogs.
  • The respondent doesn’t have a dog, but it’s for some reason other than allergies.
  • The respondent is allergic but has a dog anyway.

It might seem like a subtle difference…but not if you’re trying to determine the market potential for a dog shampoo that would allow allergic dog-lovers to have a canine best friend.

Let’s look at another one. Say you’re analyzing the results of this survey question, “I don’t shop at ABC because they have poor customer service.” Again, a “disagree” response could mean one of at least three things:

  • I do shop there, and I think the customer service is fine.
  • I do shop there even though I think the customer service stinks.
  • I don’t shop there, but for some other reason.

Once again, useless results. And dangerous results, if your business takes a particular course of action based on a flawed interpretation of flawed data.

How to write a customer survey question that delivers meaningful answers

It’s a lot harder than you may think to avoid writing double-barreled questions. In fact, it takes a conscious effort NOT to write them. So don’t think you’re immune, because no one is.

The answer isn’t about how to write a question as much as how to review it. After you’ve written your question, ask yourself, “What would I know with absolute certainty after analyzing responses to this question?” If you conclude that you’d have no way of knowing why a customer chose a certain response, you need to rewrite the question. Most of the time, the best approach will be breaking it down into multiple questions.

grocery store parking lotLet’s take another look at the one about ABC store. A better way to write that particular question would be to break it into:

  • “Do you shop at ABC?”
  • “If no, please indicate all factors that influence your decision not to shop at ABC:”
    • Prices
    • Location
    • Customer service
    • Quality of products
    • Availability of the products you need
  • “If yes, please indicate all factors that influence your decision not to shop at ABC:”
    • Prices
    • Location
    • Customer service
    • Quality of products
    • Availability of the products you need
  • “Please describe any other factors that influence your decisions about shopping at ABC.”

 

Written that way, the question could provide plenty of actionable information:

  • Which things matter most to people who shop at ABC
  • Which things matter most to people who don’t shop at ABC
  • Which problem areas should be prioritized to draw people back to ABC
  • Which stores are limited by problems that can’t easily be fixed, like location

Sending out surveys is easy. Writing good survey questions is hard.

It used to cost a lot of money to conduct a survey: focus groups, phone banks, mailings, etc. Now, it’s almost too easy — and too cheap. You can send out a survey without a lot of thought or expense. What’s the harm, right? Well, the harm could be that the company makes bad decisions based on bad data.

The best defense against writing a double-barreled customer survey question is to ask yourself, “What will I know without a doubt to be true after respondents answer this question?” If the results are open to interpretation, it’s not a good question. To get unambiguous data, you have to start with unambiguous questions.