Six Rules for Writing Your Market Research Survey

Writing a market research survey is hard.

You probably haven’t done it before, and neither has anyone in your office. And your business or research questions are unique to your product, so there’s no blueprint to copy.

But the reality is, there’s nothing to worry about. You should be careful, of course, to make sure you’re asking the right questions in an effective manner. But while market research surveys are scientific, there isn’t just one way to do it. You’re not going to screw the whole thing up by doing something wrong. And at PeopleFish, we review every single question of your survey, so you can rest assured that you’re getting your money’s worth with each and every respondent.

That said, here are nine rules to follow when writing your market research survey.

1. Keep your survey less than 8 minutes long.

Respondents shouldn’t have to spend too much time taking your survey. If it’s too long, they’ll start to rush.

So try to keep the whole thing under eight minutes in length.

How many questions can you fit into eight minutes? It depends on what you’re asking. It might mean 10 questions, it might mean 30 questions. Respondents can get through a survey pretty quick when the questions are basic and easy.

But to be sure you don’t forget to ask anything important, we recommend you write down every question you want to ask, then whittle it down. It’s better to start too long and cut back than to forget to ask important questions.

2. Use short, clear sentences.

Your survey respondents shouldn’t have to think about how to give you their opinion. The meaning of survey questions should be obvious and intuitive. Answering should be instinctual.

So rather than:

Thinking about the two pictures below, please tell us which one you most prefer.

…ask:

Which of these do you like best?

Your respondents will thank you, and reward you with easy-to-interpret findings.

3. Stick to single-select, multi-select, and rank-order question types.

If you login to Qualtrics or SurveyMonkey, you’re going to see a million shiny new question formats. Complex ranking grids, heatmap sensors, and max-diff matrices, to name a few.

Rule of thumb? Ignore these. They may look interesting, but these new formats often confuse respondents and rarely yield data that’s any more useful than what can be done with these three familiar question types.

Again, the goal is to make your survey easy and intuitive. Don’t make respondents have to figure out how to answer your questions. Stick to single- or multi-select multiple choice, or simple drag-and-drop rankings.

4. Use scales wherever possible.

Everyone is different.

Scales give your respondents the opportunity to express themselves with a bit more nuance than if they were asked simple yes or no questions.

So less:

Were you satisfied? (yes/no)

…and more:

How satisfied were you? (Not at all satisfied to Extremely satisfied).

When making your scales, be sure to label every point—don’t just use numbers. This may seem odd, but numerical scale aren’t intuitive in all cultures and languages. So rather than numerical scales, use short (5- or 7-point) scales and label all points. Strongly disagree, Somewhat disagree, Unsure, Somewhat agree, Strongly agree. This makes your question text shorter, too—no need to explain the meaning of a label-less scale.

5. Choose question types and scales strategically, not arbitrarily.

From agreement to affinity to understanding, scales should align perfectly to the question being asked, and the answers being sought. Make sure your survey questions yield the exact findings you need.

To do this, think about what you want to be able to say when your survey is done—the claims you want to be able to make. If you want to say “__% of people love our brand,” make sure you use the word “love” in one of your scales. Love, Like, Indifferent, Dislike, Hate. If you want to be able to say “__% of people have used our product,” be sure to ask about use in your survey.

So in other words, use the kinds of words and phrases in your survey that you want to be able to report on when the survey is complete.

6. Minimize open-ended questions.

Open-ended questions are fine, but not in large quantities.

The fact is, most respondents take surveys on mobile. This is unavoidable.

This means they aren’t keen on typing out long answers to open-ended questions. It’s much easier, and more effective, to ask them to choose answers from lists of options.

So it’s fine to include two or three open-ended questions, but ideally not all in a row. And ideally not too open-ended. So rather than “What do you think about…”, pick something a bit more specific, like “What’s a word you’d used to describe…”.


These are some of the rules we use when designing and editing our clients’ surveys. Take them to heart, and you’ll be well on your way to a successful PeopleFish survey project.

Oh, and last thing…don’t worry about your survey! Before we field any survey, we make sure it’s ready to go—no glaring issues, no mistakes or typos. We never field a survey unless it’s going to yield data we’d trust ourselves.

To get started on your own survey research project, leave your email below.