Most example lists pad their count and skip the two things that decide whether an open-end earns its place: the wording craft, and what happens to the answers afterward. Here are fifty questions that survive both tests, organized by use case, free to copy.
By the Wavefield Research team · Published Aug 18, 2026
The short answer
An open-ended survey question asks respondents to answer in their own words instead of picking from a list. Use them to discover reasons, language, and problems you did not think to ask about. The craft: ask about one specific thing, keep them optional, place them late, and plan how the text becomes numbers.
Every example below asks about one specific thing — the property that makes an open-end answerable in one sentence and codable afterward. Adapt the wording to your study; the shapes are the point.
Aim at moments and reasons, not general opinions — the specific memory produces the useful answer.
The goal is comprehension and objections in the respondent's own words, before you fall in love with your concept.
Honesty depends on trusted anonymity — pair these with the group-size reporting rules covered in our engagement guide.
Open-ends are where positioning language comes from — respondents hand you the words your next campaign should use.
The open-end is where voters explain themselves — the classic 'most important problem' question has run for decades because it works.
Ask at the moment of friction — an exit-intent open-end outperforms a general feedback box every time.
The highest-value open-ends in any survey: they attach an explanation to a number you already have. Piping makes them automatic.
For finding out how people actually work today — the questions that make the next survey's closed options possible.
“What do you think of our company?” produces mush. “What nearly stopped you from buying today?” produces a decision-ready answer. The narrower the moment you ask about, the more usable the text — specificity is the entire difference between an open-end that gets coded and one that gets skimmed and forgotten.
Every open-ended question adds typing effort, and forcing one measurably raises abandonment — especially on phones, where most surveys are answered. Required open-ends also fill with junk (“asdf”) that pollutes coding. Make them optional and let the motivated answer; low-effort text gets flagged automatically anyway.
One to three open-ends per questionnaire, positioned after the closed questions have warmed the respondent up and framed the topic. An open-end as question two is a wall; the same question second-to-last is a natural reflection.
An open-end you never analyze was a cost, not a question. Decide before fielding how the text becomes numbers: a codeframe you apply, themes you count, quotes you pull for the report. If you cannot say what you will do with the answers, cut the question.
Where open-ends sit among the other question formats — and when a closed question serves better — is covered with examples in survey question types. The follow-up probes above use answer piping, which inserts a respondent's earlier answer into a later question automatically.
Open-ended answers only become findings through coding: assigning each response to one or more themes in a codeframe, so free text turns into counts you can chart, crosstab, and trend like any closed question. This is the step the example listicles never mention, and it is where most open-end value dies — a spreadsheet column of 400 verbatims that nobody has time to read is a question that should not have been asked.
The traditional cost was days of analyst time. AI-assisted coding collapsed it: on Wavefield, the agent proposes a codeframe from your responses (or applies one you supply), assigns themes to every verbatim, and you review and adjust the assignments — minutes instead of days, with the human judgment kept where it matters. Coded themes then flow into crosstabs and exports like every other variable. The full analysis workflow is in how to analyze survey results.
Ten that work almost anywhere: What nearly stopped you from buying? What would make this a 9 or 10? If you could change one thing, what would it be? What is this product for, in your own words? What is the hardest part of this process today? What does this brand do better than anyone else? What one change would most improve working here? What were you hoping to find? What would you use if we disappeared? Anything else you want us to know? Each asks about one specific thing — the property that makes an open-end answerable.
Three properties. It targets one specific moment or reason, not a general opinion. It could not be answered better by a closed question — if you could list the likely answers yourself, list them instead and save the respondent typing. And it has a planned path from text to numbers: a codeframe, a theme count, or a report quote. Vague prompts, double-barreled asks, and mandatory essays are the three ways open-ends go wrong.
Coding: reading responses and assigning each to one or more themes in a codeframe, which turns free text into counts you can chart and crosstab like any closed question. Traditionally an analyst codes a few hundred verbatims by hand over days; AI-assisted coding now proposes the codeframe and assigns themes in minutes, with a human reviewing the assignments. Either way, the discipline is the same: consistent themes, every verbatim coded, counts reported with base sizes.
Whenever you already know the possible answers and need to measure how common each one is — closed questions are faster to answer, impossible to skip-read, and analyzable the moment fielding closes. The classic workflow runs open-ends first in a discovery round to find the categories, then closed questions at scale to measure them. A survey that is mostly open-ends is usually an interview wearing the wrong format.
Related: survey question types · customer feedback surveys · employee engagement surveys · political polling
The agent builds your survey with open-ends placed properly, and AI coding turns the verbatims into themes you can chart. Free to build, from $99 to field.
Start a project