An online survey builder, and a workflow for data you can use

On the free plan you get unlimited surveys, 500 responses a month, and 24 question types, including rating, star, matrix, ranking, dropdown, file upload and open text. Conditional logic, answer piping, hidden fields for tracking where respondents came from, multi-language surveys, anonymous mode and partial-response saving are all included. NPS, e-signature and payment questions can be built and previewed on Free but need Pro ($14.99/mo) to publish.

The same questions can run in three layouts: conversational (one question at a time, strongest on phones), single page (fastest for short surveys at a desk), and document (questions inline with text, useful for research consent and briefings). If you already have a survey in Google Forms, Typeform or Jotform, the importer rebuilds it rather than making you retype it.

The most common survey failure isn't a bad question; it's a survey with no stated purpose. Before drafting anything, finish this sentence: if the answer comes back X we will do A, and if it comes back Y we will do B. If you can't, the survey has no job yet, and every question in it is optional by definition.

That sentence also tells you who to ask. A survey meant to justify a pricing change needs responses from people who pay you, not from everyone on your list. And it tells you what to cut: any question whose answer would leave both A and B unchanged.

Measure length in minutes, not questions. SurveyMonkey's analysis of about 100,000 surveys found respondents spend roughly 5 minutes on a 10-question survey, and that abandonment rises for surveys taking more than 7–8 minutes. It also found people spend less time per question as surveys get longer, which means later answers get less thought.

Before you send, estimate the time with our survey completion time calculator at /tools/survey-completion-time-calculator and put the honest number in the invitation. Our guide on how many questions a survey should have traces other commonly quoted length figures to their sources.

Write each question to measure one thing in words every respondent will read the same way. Wording moves results: Pew Research Center reports that when a question about military action in Iraq added that U.S. forces might suffer thousands of casualties, support fell from 68% to 43%. Read every draft question aloud looking for the answer you hope for, and split anything that asks about two things at once.

Order matters too. Put general questions before specific ones on the same topic, screeners first, and demographics and sensitive questions at the end. Then pretest: AAPOR recommends pretesting with people like your real respondents. Five people talking through each question will catch most ambiguity in an afternoon. Our guide to writing survey questions has before-and-after rewrites.

Use logic wherever a question assumes a premise. "How often do you use the mobile app?" should only appear for people who said they use it; everyone else will pick the least-wrong option and quietly contaminate the data. In Zunoform, logic shows or hides individual questions based on earlier answers, on every plan. Google Forms can only jump between sections, so a single follow-up often means restructuring the survey into sections.

Logic also shortens the survey for most people. A satisfaction survey can ask "what went wrong?" only of people who scored 6 or below, and "what should we keep?" only of those who scored 9 or 10. Answer piping can then quote an earlier answer back ("You said checkout was difficult; which step?"), which makes follow-ups read as a conversation.

Send the survey where respondents already are, and track each channel separately. Zunoform's share page gives you a link, an inline embed, a popup embed, a full-page embed and a QR code. Add a hidden field such as source to each link (?source=newsletter, ?source=app) and every response records the channel it came from, which shows up in the response view, the CSV export and Google Sheets columns.

Plan your sample before launch. For a large population, a 95% confidence level with a ±5% margin of error needs about 385 completed responses (1.96² × 0.25 ÷ 0.05²). If you'll compare segments, each segment needs enough responses on its own. Our sample size calculator at /tools/sample-size-calculator and response rate calculator at /tools/survey-response-rate-calculator turn those into an invitation count. On Zunoform's free plan, remember the 500-responses-a-month ceiling is account-wide.

Start with where people quit. Zunoform saves partial responses as people answer, and the Summary tab shows a drop-off funnel, so you can see which question lost respondents. It also charts each question and opens with quick insights such as leading answers, splits and trends, on every plan; an AI-written summary of themes in free-text answers is Business-only.

For deeper work, export a CSV or connect Google Sheets on the free plan so every response lands as a row. Report percentages with their margin of error (our margin of error calculator at /tools/margin-of-error-calculator does the arithmetic), read every open-text answer on small samples, and check grids for straightlined rows before trusting their averages.

Zunoform fits surveys of up to a few hundred responses a month that need per-question logic, a one-question-at-a-time layout on phones, anonymous mode, multi-language versions or scoring, all on the free plan, with branding removal at $14.99 a month on Pro.

It doesn't fit if you need thousands of free responses (Google Forms and Tally have no stated cap), panel recruitment, or advanced statistics built into the tool; research platforms such as SurveyMonkey or Qualtrics sell those. AI survey generation and AI summaries are Business-only ($24.99/mo), so don't pick Zunoform's free plan for AI features.