Constructing a Survey That Works

Over the years I’ve conducted numerous market research projects for my clients. There are so many pieces that have to be done right to make the study a success. And you won’t be surprised to know that one important piece is the thing at the center of the research…the questionnaire. Constructing it well makes a lot of difference to how much you learn. So, here are a few tips to help get the survey’s questionnaire right.

Keep It Short and Focused

It never fails. The desire is there to ask pretty much EVERYTHING you’ve ever wanted to ask a customer – all at once. Mistake. The truth is that you only have so much of someone’s attention. And you have to use it wisely. One wrong move and they’ll hop right off. Or if they get fatigued, they’re gone. Respect your survey respondents. Take the time to whittle down your survey to only the most important questions. 

How do you know which questions are the most important? Well I could write for hours on that one. It depends on the purpose of the research. In a nutshell, you should think about what might be going on in the customer’s mind and how he or she makes purchase decisions for your product or service. Consider what you believe makes your company unique. Zero in on those elements that will help you better understand (1) what’s important to that given customer and (2) how you are doing along those lines. DO NOT waste their time on things that simply don’t matter much. It’s tempting to want to ask a lot more. Resist it.

Build It Right

Make the survey instrument easy to use and clear. Way too often surveys are poorly designed (or not designed at all). Don’t let your survey become a dump of questions. Consider the questionnaire from the survey taker’s perspective. Does this question make sense? Is it logical that this question follows that one?

And make the survey smart. Consider its flow from question to question and topic area to topic area. For instance, if I’ve asked you several questions about your decision-making process it might feel jarring to go from there to detailed questions about the physical layout of my store. Consider how you can make a smoother transition from area to area. 

Make sure the questions “learn” from each other. For example, if I just said that I’ve never visited your website, don’t ask me a bunch of questions about the website. The survey should “know” how I just responded to certain questions and serve up only relevant questions. This is called “piping” by the way.

Wording, Grammar, and Sentence Structure

I know. I know. Duh. Boring stuff. But it goes deeper than that. What impression does it leave when you read something from a company that has typos, misspellings, or confusing or unnatural wording? It’s not positive is it? Is the survey full of jargon that you assume your reader understands? Using acronyms is a huge place where this shows up a lot. Read and re-read your survey instrument.

One thing in particular to look for is multiple questions posing as a single question. For instance, asking someone if they like your front door and reception desk seems innocuous enough. But when looked at critically, it’s easy to see that you’re really asking two separate questions – one about the door and one about the reception desk. Someone might have differing opinions about the two. Either split this into two questions or (if appropriate) raise the question to a higher level; ask about their impressions of the entry area. OK, this may not be the strongest example, but surely you get the point. 

Avoid the Slant

It’s the political season, so we’re all inundated with polls and surveys about voter opinions. And we often take the “summaries” of these research pieces as being solid, unbiased research. But deep down we know that more often than not these research pieces are biased simply by the questions that are asked or how they are asked. Political research tends to be among the worst offenders, but everyday business surveys often reflect our biases as well. If we ask someone whether they prefer to eat a flavorful, juicy steak or a bland slab of tofu we’re kind of leading them to the answer we want, aren’t we? Then if we turn around and report that there is an extraordinarily low number of vegetarians in our population, it’s simply dishonest.

Reread your survey questions objectively. Think about how you might be biasing the responses. Or better yet, have someone else read it from that perspective.

Another form of bias is what I call “unquestioning”. I have had a number of clients over the years try this. When faced with asking a question that they suspect might paint a more negative picture of the company, they simply decline to ask the question. It’s harder to tell me you don’t like me if I don’t ask whether you like me, right? Again, it slants the research in a particular direction.

Open Up The Box

Don’t forget to allow open-ended questions in your survey. Give respondents a place to further explain their responses when appropriate. That “flavoring” of responses can be very valuable. 

It’s also a good practice to provide an open-ended question at the end that provides the survey taker the opportunity to tell you something you may not have covered in the questions. They will feel like you gave them the chance to say exactly what’s on their minds; not just be completely limited to the questions you chose to ask. One thing I’ve seen is that a number of questions on subsequent surveys get surfaced here. If I didn’t ask about my product’s color but a bunch of customers brought that up in the open-ended questions, it’s a signal that maybe I should  be asking about that issue. It’s a potential question to be added for the next survey.

Avoid Yes/No

I’m not a fan of the yes/no question. I get it. It’s a clear response. But really there’s very little black and white in our world. Most beliefs and feelings are gradients. I somewhat like your product. I greatly disagree with that statement about how much your service meets my needs. I figure that if you have someone’s attention enough to answer a question, you might as well get the subtleties around the person’s feelings. Nuanced questions offer a more informative experience and one that is much more actionable.

Keep in mind that “To what extent…”-style questions are more difficult or challenging for a respondent to answer. Asking a lot of those types of questions can eventually wear down a survey taker. It’s making them think a little more deeply than simply answering yes or no. So keep that in mind as you construct the questionnaire. 

Open and Close (and Follow-up)

It’s pretty common to overlook the “frame” around the survey. Be sure to introduce the survey, why its being done, set their expectations about how long it will take, etc. Give the prospective respondent (after all, they haven’t actually started the survey yet – and can bail before they do) a sense for what you are trying to accomplish and how that could be to their benefit. If I am a customer and I feel that my responses to a survey could lead to an improved experience with your company I’m more likely to invest the time required to complete it. But what if I get the sense that you are simply interested in checking off a box labeled “Conduct Customer Satisfaction Survey” on a to-do list that your company leaders gave you? Ummm, much less likely.

And don’t forget to close graciously. Thank respondents for their time. Remind them again why you are conducting this research and what you will be doing with the info you collect. (And as a follow-up, actually do it!)

OK. I don’t assume that I’ve covered the waterfront as it pertains to surveys here. My intent was to help you think more critically about the research you do and the different ways the actual survey questionnaire can impact the results you get. 


Market research is one of those things that seems easier than it really is. Is it easy to throw a survey out there? Sure. Is it easy to throw a good, actionable survey out there? Not so much. And that's where I can help. I've done tons of market reserach over the years for lots of different types of businesses. And I can help you conduct yours.

Please feel free to reach out and get in touch and let's explore how I can help you and your business succeed. No pressure. Just an informal discussion to explore some ideas. You can reach me at (713) 907-8429 or BCohen@IDiscoverConsulting.com

I hope you are enjoying these blog posts If so, please help spread the word. Tell others about IDiscover Consulting Group and IDiscover Journal. Share these posts. Comment on them. I’d really love to hear your ideas!