Open-Ended Survey Questions: Write Them, Get Better Answers

What an open-ended survey question is

An open-ended survey question has no fixed answer options. The respondent replies in their own words instead of picking from a list or a scale. “What is the main reason for your score?” is open-ended; “How satisfied are you, 1 to 5?” is closed-ended.

The trade is simple. Closed questions are fast to answer and easy to count, but they can only report the options you thought of in advance. Open questions surface the options you did not think of, at the cost of answers that are slower to give and slower to analyze.

When to use one (and when not to)

Use an open-ended question when you need one of these:

• The reason behind a number. A rating tells you that something is wrong; the follow-up tells you what.
• The answer list is unknown. Early exploratory work, new categories, or a market you do not yet know well.
• Language you will quote. Verbatims are what make a research deck feel real to the people reading it.
• A catch-all. “Anything else you would like to tell us?” at the end collects what the structured questions missed.

Do not use one when the answer is a fact you can offer as a list (country, age band, product owned), or when you will not have the time to read and code the answers. An open question you never analyze still costs the respondent effort and still lowers completion.

Wording rules that change the answers

1. Ask for one thing. “What did you like and what would you change?” gets one half of the question. Split it into two.
2. Anchor it to something the respondent just did. “What is the main reason for the score you just gave?” works because the number is fresh. “Any comments?” asks the respondent to decide what you want.
3. Do not lead. “What did you love about our fast delivery?” plants the answer. “How would you describe the delivery?” does not.
4. Avoid yes/no phrasing. “Was the checkout easy?” invites “yes”. “What, if anything, made checkout difficult?” invites a reason.
5. Say how much you want. “In a sentence or two” or “Tell us in a few words” sets the expected length.
6. Limit how many you ask. One to three per survey is usual. Each one is more work than ten closed questions, and fatigue shows up in the later ones as shorter, blanker answers.

Examples by goal

If your survey is a customer-feedback one, the voice of customer survey guide shows where these sit in a full question set.

The real problem: short, thin answers

Most teams do not fail at writing the question. They fail at what comes back. On a phone, typing a considered reply is work, so a large share of respondents write a few words (“price”, “good service”) or skip the box. You get a column that is easy to count and impossible to learn from.

Three things help, in this order:

1. Better wording (above). Free, and it matters more than any tool.
2. Placement. Put the open question right after the rating it explains, and early enough that the respondent has not already tired.
3. A lower-effort way to answer. Speaking is easier than typing for many people, especially on mobile.

Letting respondents answer by speaking

Voice Capture is a widget you paste into the survey you already run — Alchemer, QuestionPro, FormAssembly, Sawtooth Discover or any platform that runs custom JavaScript. It does not replace your survey platform. The respondent taps a button on the open-ended question, speaks, and the answer is transcribed automatically (36 languages) and written back into that question’s field, so it appears in your normal export.

What to know before you try it:

• Keep the text box. Voice is an option, not a requirement. Respondents in a meeting or on public transport will type.
• There is a length limit. The maximum spoken answer is 90 seconds on the free trial, 180 on Essential and 300 on Professional. That is plenty for “why did you give that score?” and short of a long narrative.
• You get the transcript, not the audio. Voice Capture transcribes and discards the recording; only the text is stored.
• Check your platform first. Typeform and SurveyMonkey strip custom JavaScript from the respondent view, so the widget cannot run there. The integrations page lists what is supported.

For the practical side of audio responses, see the audio survey guide.

What changes in the data

On our production data (7 studies, 7,169 responses), spoken answers averaged 23.1 words against 10.1 for typed answers in the same studies: roughly twice as long. It is not a tenfold jump, and the figure comes from our own fieldwork rather than a controlled trial. A separate pre-product pilot suggested a larger gap; both are published side by side on the methodology page.

Longer is not automatically better, but in practice it means a reason where there would have been a keyword.

Analyzing open-ended responses

1. Read a sample first. Fifty verbatims show you the themes and the rough edges of the transcription.
2. Build a code frame from what you read (price, reliability, support speed…), then apply it to every answer.
3. Count and cross. The finding is “detractors mention onboarding three times as often”, not “people have opinions”.
4. Keep the best quotes attached to their theme so the report has voices as well as percentages.

The step by step for transcribed answers is in how to analyze voice survey responses.

Frequently asked questions

What is an open-ended survey question?

A question with no preset answer options, where the respondent replies in their own words.

How many open-ended questions should a survey have?

Usually one to three. Each takes much more effort than a closed question, and answer quality falls as respondents tire.

What is the difference between open-ended and closed-ended questions?

Closed-ended questions offer fixed options (a scale, a list, yes/no) and are easy to count. Open-ended questions collect free text, which can reveal answers you did not anticipate but takes longer to analyze.

Can people answer an open-ended question out loud?

Yes, if your survey platform allows custom JavaScript. A widget such as Voice Capture adds a record button to the question and writes the transcript into the normal answer field.

How much does it cost?

One credit is one transcribed response and credits do not expire. The free trial includes 250 credits and the Essential pack is $99 for 1,000. Current packs are on the pricing page.

Where to start

Pick the single open-ended question in your current survey whose answers you most wish were longer, rewrite it with the rules above, and test voice on that question alone with a small slice of the sample. The free trial is enough to compare typed and spoken answers side by side. For the NPS case specifically, see NPS with voice responses.