Best Practice Basics: Open-Ended Questions

In this article:

Open-ended survey responses are a great way to collect unedited consumer feedback and reaction to ideas. At their best, open-ended responses are critical to diagnose concept issues and frame opportunities for idea optimization. But how can we elicit thoughtful, detailed answers from survey-takers?

At Upsiide, we have found that great questions are more likely to garner great answers, and framing questions purposefully makes respondents more purposeful in their answers as well.

Here is our guide to writing better open-ended questions.

Instead of: “What do you dislike, if anything, about this package design?”

Consider: “This package design is still in development. Please help us make it better by telling us everything that you dislike or think might be improved.”

By learning the purpose of this question, the respondent can better understand that negative feedback will be used to make the package design better for everyone – this is empowering and motivating.

Instead of: “How do you feel about this list of ingredients?”

Consider: “In this list, are there any ingredients that you feel we should remove? If, so please tell us why you feel we should avoid these ingredients.”

Here, respondents are directed to the specific elements you want them to engage with, with precise instruction. Specific phrasing minimizes different interpretations of the question, and pointed requests discourage short, un-elaborate responses, like “Looks good” or “I don’t like it”.

Instead of: “Please list everything you like and dislike about this product. Explain why you like or dislike these ideas.”

Consider: “In the previous question, you said that you dislike the font choice on this package. How could we improve the font?”

Prefacing the open-end with a forced-choice question and using the open-end as an opportunity to elaborate is a great way to keep respondents engaged - the survey questions are customized to their earlier responses, which feels more like a two-way conversation.

This approach minimizes the amount of effort required to engage thoughtfully with the topic and avoids the sense of overwhelm associated with a line of questioning that is too broad.

When programming your numerical open-ended question, it is best to set restrictions on the maximum number that respondents can specify to avoid extremely unlikely outliers.

In report mode, your open end results will be presented in 4 numbers; minimum, maximum, mean and median.

Typically, the mean and median score's are the primary output used for analysis. If you are asking how many of a product respondents would buy on average, or if you are trying to determine what price on average respondents would expect to pay for a product, these will be the the most important metric for analysis.