The average survey response rate, and where the numbers come from

The basic formula is completed responses divided by the number of eligible people you invited, times 100. The words that matter are completed and eligible. Bounced emails and people outside your target group come off the bottom of the fraction; partial responses usually stay off the top unless you have decided in advance what counts as complete.

Survey researchers use the American Association for Public Opinion Research (AAPOR) Standard Definitions, which set out six response rate formulas (RR1 to RR6) that differ in how they treat partial interviews and cases of unknown eligibility. You don't need all six, but you should say which rule you used when you compare against a benchmark.

Vendor benchmarks often use a different denominator entirely. Survicate's in-product report divides responses by views, not by invitations, so a 10% figure there is not comparable with a 10% email survey rate. Our survey response rate calculator keeps completed and partial responses separate for exactly this reason.

Illustration: you email a customer survey to 1,200 people. 48 addresses bounce, leaving 1,152 eligible. 230 people finish the survey and another 40 start but stop. Your response rate is 230 ÷ 1,152 = 20.0%. Counting partials would give 270 ÷ 1,152 = 23.4%, which flatters the number and hides the fact that 15% of the people who started gave up.

If you had divided by the original 1,200, you would report 19.2%. Small choices like these move a rate by several points, which is why a benchmark without a stated formula is hard to use.

A good rate is one that beats your own last send to the same kind of audience, and one high enough that the people who answered are not obviously different from those who didn't. Published averages help you spot a problem, not set a target.

Relationship is the biggest dividing line. The high academic figures come mostly from researchers surveying defined groups, often with pre-contact and follow-up. Wu and colleagues found that surveys sent to a clearly defined, refined population did better, and that inviting more people did not raise the rate. General-population surveys with no prior relationship sit at the other end, as Pew's single-digit phone rates show.

Channel is the second line. In controlled comparisons, web surveys trail other modes by about 12 points, and the 2023 Cochrane review found the odds of response to a postal questionnaire were more than twice those for an electronic one. Holtom's team, looking at published management research, found no difference between online and mail by 2020, a reminder that averages across studies and head-to-head trials can disagree.

The honest answer is that nobody publishes a trustworthy cross-industry figure for email surveys today. The best traceable series is old: Sheehan's 2001 review found published email surveys fell from 61.5% in 1986 to 24.0% in 2000, largely because novelty wore off as inboxes filled.

A current, well-documented example is NSSE, the US student engagement survey, which invites students by email with a series of reminders. Its 2026 average institutional response rate was 23%. That is a warm audience with institutional backing, so a cold marketing list should expect far less, and an internal team survey with a manager's nudge can expect far more.

Usually, and the form of the incentive matters more than its size. In Church's 1993 meta-analysis of mail surveys, prepaid cash sent with the questionnaire raised response by 19.1 percentage points on average, gifts by 7.9 points, and incentives promised only on completion did not significantly raise response. Singer and Ye's 2013 review of later research reached the same broad ordering: prepaid beats promised, and money beats gifts.

Online, prepaying is harder, and the effects are smaller. Göritz's 2006 meta-analysis of web studies found incentives raised the odds of starting a survey by about 19% (odds ratio 1.19) and of finishing it once started by 27% (1.27). Holtom's team found published surveys reporting incentives averaged 74% against 65% without in 2020, while Wu and colleagues found no significant incentive effect in education surveys. Expect a modest lift, not a rescue.

Reminders help, with diminishing returns. The Cochrane review found electronic and telephone reminders raised response to postal questionnaires, and Wu and colleagues found phone reminders helped online surveys. Earlier work summarised by Holtom's team found that one or two reminders to web surveys had only a nominal effect, while three to five were associated with lower response. One or two reminders, a few days apart, is the defensible default.

Length works twice. A shorter questionnaire raised the odds of response by about half in the Cochrane trials, and Galesic and Bosnjak found that the longer the stated length of a web survey, the fewer people started it and the fewer finished. Later questions also got shorter, less careful answers. State the real time honestly: our survey completion time calculator estimates it from your question mix.

Most of what moves a response rate happens before anyone opens the form: who you invite, the invitation, the incentive, the reminders. A form builder can't fix those. What it can do is avoid losing people who do open it, and show you where they drop.

Zunoform saves partial responses as people progress, so you can see starts separately from completions instead of guessing. You can run the same questions as a one-question-at-a-time conversational form or a single page and compare, pass the channel in a hidden field (for example a UTM parameter) to see which send performed, and close the form on a date so your deadline is real. Free covers 500 responses a month; if you expect more than that, Google Forms, Tally and Youform have no response cap on their free plans.