Retour

Survey Completion Rate: Formula, Diagnosis and Fixes

Matthieu Saussaye

Survey completion rate is the percentage of people who finish a survey out of those who started it: completes divided by starts, times 100. If 800 people open the survey and 600 reach the end, the completion rate is 75%. It measures how well the survey holds people once they begin, which makes it different from response rate, which measures how many of the people you invited answered at all.


Key points:

  • Formula: completion rate = completed surveys ÷ surveys started × 100.

  • Response rate uses invitations as the denominator, completion rate uses starts. A survey can have a low response rate and a high completion rate, or the reverse.

  • The number alone is not the insight. The useful view is where people drop out, question by question.

  • Length is the biggest driver. SurveyMonkey reports completion rates falling 5% to 20% for surveys taking longer than 7 to 8 minutes (SurveyMonkey).

  • Define "start" and "complete" before you report, or two teams will report different rates for the same survey.

Completion rate vs response rate and related survey metrics

Survey metrics are often mixed up because they share the same numerator and differ only in the denominator. Here is the full set, using one example: 5,000 invitations, 1,000 clicks, 900 who answered the first question, 150 screened out, 600 completes.

Metric

Formula

Example

What it tells you

Click or open rate

Clicks ÷ invitations delivered

1,000 ÷ 5,000 = 20%

How compelling the invitation and channel were

Start rate

Answered first question ÷ clicks

900 ÷ 1,000 = 90%

Whether the landing and introduction put people off

Completion rate

Completes ÷ eligible starts

600 ÷ (900 − 150) = 80%

How well the survey itself holds qualified respondents

Dropout (breakoff) rate

Eligible starts that did not finish ÷ eligible starts

150 ÷ 750 = 20%

The inverse of completion; the number to drive down

Incidence rate

Qualified ÷ screened

750 ÷ 900 = 83%

How well your sample source matches your target

Response rate (simple)

Completes ÷ invitations delivered

600 ÷ 5,000 = 12%

Overall yield of the whole fieldwork

Notice the choice in the completion row: we removed screen-outs from the denominator, because people who were terminated by design did not abandon the survey. Some tools include them, which makes completion look worse than it is. Write your definition down and keep it constant.

For formal reporting, AAPOR's Standard Definitions set out how to classify complete interviews, partial interviews and breakoffs and how to compute standardized response, cooperation and contact rates (AAPOR Standard Definitions). AAPOR also notes that a response rate alone is not sufficient to determine how much nonresponse error a survey has.

How to calculate survey completion rate correctly

  1. Define a start. Usually: answered the first real question. Page views and link clicks are not starts; they inflate the denominator.

  2. Define a complete. Usually: reached the final question or the end screen. If you accept partials, for example respondents who answered at least 80% of core questions, state that rule.

  3. Remove screen-outs and quota-fulls from both numerator and denominator. They were ended by you, not by fatigue.

  4. Remove duplicates and test responses before calculating.

  5. Calculate per segment. Device, source, language and country often differ sharply, and an average can hide a broken mobile layout or a bad translation.

Find where people drop out: the dropout curve

A single completion rate tells you there is a problem. A dropout curve tells you where. Plot, for each question in order, the share of eligible starters still present. A healthy survey shows a gentle slope. Problems show up as cliffs.

Pattern

Likely cause

Fix

Cliff on the first screen

Intro too long, length too high, or a demanding first question

Shorten the intro, start with an easy closed question

Cliff at a grid

Long matrix, especially on mobile

Split it, cut items, or convert to single questions

Cliff at an open question

Typing is effortful, especially on phones; the question feels vague

Make it optional or more specific, or allow voice answers

Cliff at personal questions

Income, health or contact details asked too early or without reason

Move them to the end, explain why, add "prefer not to say"

Steady leak after the midpoint

General fatigue; the survey is simply too long

Cut questions; move core questions earlier

Cliff after a page break or logic jump

Technical error, slow load, broken routing

Test every path on mobile and desktop

Cut the curve by device too. If mobile users drop out at a question where desktop users do not, the problem is layout, not content.

What is a good survey completion rate?

There is no universal benchmark worth quoting, because completion depends on length, audience, incentive and how "start" is defined. What matters more is the comparison with yourself: the same survey over time, different versions in an A/B test, or different segments in the same wave.

Useful rules of thumb instead of a single number:

  • Short surveys with a motivated audience (employees, active customers, paid panels) should keep almost everyone who starts. Large dropout there is a design signal.

  • Long or cold-audience surveys will lose more, and the loss accelerates after a certain length. SurveyMonkey found abandonment rising for surveys over 7 to 8 minutes, with tolerance higher for work or school surveys and lower for customer ones (SurveyMonkey).

  • Day of answering matters a little. SurveyMonkey's State of Surveys data shows higher-than-average completion rates for surveys taken on Wednesday, Thursday and Friday (SurveyMonkey).

  • A completion rate that is too high on a long, dull survey can mean people are speeding to the end for an incentive. Check time per question before celebrating.

How to increase your survey completion rate

Cut length first

Every other fix is smaller than this one. Remove questions that map to no decision, use skip logic so people only see what applies to them, and drop "nice to know" demographics you never cut by. Then state the real length in the invitation. Understating it raises starts and lowers completion, which is the worst of both.

Fix the first 30 seconds

Open with a one-sentence purpose, the honest length, and an easy first question. Respondents spend the most time on the first question, around 75 seconds on average including the introduction in SurveyMonkey's data, so that screen sets the tone (SurveyMonkey).

Design for the phone

Replace grids with single questions, use large tap targets, avoid sliders that are fiddly on touch screens, and keep one question per screen where possible. Test every path on an actual phone.

Show progress, honestly

A progress bar helps when it moves at a steady pace. If the first half of the survey is long and the bar barely moves, it discourages people. Group questions into short sections instead.

Make open questions easier to answer

Open questions are a frequent breakoff point because typing a thoughtful answer on a phone is work. Ask fewer of them, make each one specific ("What almost stopped you from buying?" not "Any comments?"), and consider letting people answer by voice.

Mix question types

Long runs of identical scales encourage both dropout and straightlining. Alternate single choice, ranking and a short open question. Our guide to closed-ended questions covers formats that are quick to answer.

Be careful with push-to-complete reminders

Reminding people who started but did not finish sounds like an obvious win. A 2026 experiment by the SOM Institute at the University of Gothenburg found push-to-complete reminders produced no statistically significant effect on completion rates, and instead increased nonresponse bias (Persson, SOM Institute, 2026). Fixing the survey usually beats chasing the people it lost.

Why completion rate matters for data quality

Dropout is rarely random. People who abandon a survey tend to be busier, less engaged with the topic, or on a device that made the survey harder. If they leave in large numbers, your completes over-represent the most motivated respondents, a form of nonresponse bias covered in our guide to survey bias.

Low completion also costs money: every breakoff is a respondent you paid to recruit and could not use. Plan for it in your sample size by inflating invitations, and if response rates are your bigger problem, see why survey response rates are crashing.

A completion rate checklist before every launch

  • Definitions of start, complete, partial and screen-out written down and shared

  • Real median length measured in a pretest and stated in the invitation

  • Every route tested on a phone and on a desktop, in every language

  • No grid longer than five items; no more than two or three open questions

  • Sensitive and demographic questions at the end, each with "prefer not to say"

  • Dropout tracked by question, device and source from the first day of fieldwork

  • A soft launch to the first 5% to 10% of the sample, reviewed before full release

Run through it once and you will catch most of the problems that would otherwise show up as a cliff in your dropout curve halfway through fieldwork.

Keep people answering to the end

Most breakoffs happen where the survey asks for effort: long grids, open questions typed on a phone, and questions that do not seem to apply.

SmartInterview runs the survey as a conversation, by voice or text. Respondents can answer out loud instead of typing, an AI follows up only when an answer is vague, and the conversation adapts to what they have already said, in their own language. You get the depth of open questions without the dropout they usually cause, and the answers come back coded into themes automatically.

Start free

Frequently Asked Questions

How do you calculate survey completion rate?

Divide the number of completed surveys by the number of people who started the survey, then multiply by 100. For example, 600 completes out of 750 eligible starts is an 80% completion rate. Exclude screen-outs and quota-fulls from both numbers and define what counts as a start before calculating.

What is the difference between response rate and completion rate?

Response rate measures completes against everyone you invited, while completion rate measures completes against those who started. Response rate reflects the invitation and the audience; completion rate reflects the survey experience itself.

What is a good survey completion rate?

It depends on length, audience and definitions, so compare against your own previous waves rather than a generic benchmark. Short surveys to motivated audiences should keep nearly all starters. Long surveys to cold audiences will lose more, especially beyond 7 to 8 minutes.

Why do people not finish surveys?

The most common reasons are excessive length, long grids, effortful open questions on mobile, sensitive questions asked too early, irrelevant questions and technical problems. A question-by-question dropout curve shows which one applies to your survey.

How can I increase survey completion rates?

Shorten the survey, state the real length, start with an easy question, design for mobile, replace grids, make open questions specific or allow voice answers, and test every route. Fixing the design usually works better than reminding people who dropped out.

Does completion rate affect data quality?

Yes. Dropout is rarely random, so a low completion rate means your completes over-represent the most motivated respondents. A very high completion rate on a long survey can also hide speeders, so check time per question too.

Table des matières

Articles associés

Inscrivez-vous gratuitement

Inscrivez-vous gratuitement

Inscrivez-vous gratuitement