"Is 2% good?" is one of the first questions people ask after putting a survey on their site. The honest answer is that it depends on what the 2% is a share of.
This post collects published response rate benchmarks by channel, explains why they differ so much, and shows how to work out your own rate when your survey tool does not report one.
The short answer
There is no single good number, because rates depend on the channel and on what you divide by. In Survicate's 2025 data, website widget surveys had a median response rate of 7.64% of survey views. SurveyMonkey's 2025 data puts popup surveys at an average of 3.65% and email surveys at 49.17% of the people who received them. Compare your rate with the same channel and the same denominator, and above all with your own pages over time.
Survey response rate benchmarks by channel
Two survey companies publish benchmarks from their own customers' surveys. They measure differently, so read each row with its denominator.
| Channel | Response rate | Divided by | Source |
|---|---|---|---|
| Website widget | 7.64% (median) | Survey views | Survicate |
| In-app, mobile | 18.69% (median) | Survey views | Survicate |
| Popup | 3.65% (average) | People who received the survey | SurveyMonkey |
| Web link | 29.95% (average) | People who received the survey | SurveyMonkey |
| SMS | 18.54% (average) | People who received the survey | SurveyMonkey |
| 49.17% (average) | People who received the survey | SurveyMonkey |
Survicate's report is based on 4,332 surveys from 460 companies, and its overall median is 9.98%. Media companies had one of the lowest medians, 5.37%. Surveys with 2 to 3 questions did best, at 15.97%, against 6.87% for surveys with 7 or more.
A median is the middle survey; an average can be pulled up by a few very high rates. That is one more reason not to set one source's figure against another's.
Why the numbers differ so much
Most of the gap between a 4% popup and a 49% email comes from the denominator, the number you divide by. There are three common ones:
- Recipients. An email survey is divided by the people it was sent to. Those people are on your list, so they already know you, and many surveys go to small lists of recent customers.
- Survey views. A widget survey is divided by the times it was shown. That includes every visitor who arrived from search, glanced at the page and left.
- Page views or visitors. If your tool does not count views of the survey itself, you divide by the page's traffic. Readers who never scrolled down to an end-of-page question still count, so this rate comes out lowest of all.
What counts as a response varies too. Survicate counts a response when someone starts the survey or clicks its welcome screen. SurveyMonkey defines the rate as the share of people who started the survey. Neither is the share who finished. Your own tool may count only completed answers, which gives a lower figure again.
Response rates have been falling for years
Low rates are not just a website problem. In 2017, Research Live reported an industry analysis of eight large face-to-face UK government surveys, including the British Social Attitudes survey and the Crime Survey for England and Wales: their response rates had dropped by an average of nine percentage points in 15 years. In the US, Pew Research Center reported that its typical telephone survey response rate fell to 6% in 2018.
Pew adds a point worth keeping: studies of low response "have generally found that response rates are an unreliable metric of accuracy". A low rate does not make your answers wrong. Who answers matters more than how many, and that is a separate problem from the rate itself.
Tip: If you worry that only the most opinionated readers answer, the vocal minority post covers how to plan around it.
How to work out your own response rate
Many on-page tools, Asklet included, do not count how often the survey was loaded, so they cannot show a response rate. You can work one out in a few minutes:
-
Pick a page and a date range
Use dates when the survey and the page did not change.
-
Count the answers
Take the number of responses for that page and those dates from your survey tool. Count readers who answered, not individual answers to each question.
-
Take the page views from your analytics
Same page, same dates. If your analytics records scroll depth, also note how many views reached the part of the page where the survey sits.
-
Divide
Responses divided by page views, times 100. If you have the scroll figure, divide by that too: it is closer to the "survey views" in the benchmarks above. The response rate calculator does the sums for you.
Is your response rate good? Compare like with like
A benchmark from another company's surveys is a rough guide at best. Comparisons you can trust are closer to home:
- The same page over time, with the same question in the same place.
- Similar pages against each other: guides with guides, news with news.
- The same placement: an inline question at the end of an article reaches fewer readers than a side tab that is always on screen, so their rates are not comparable.
- The same denominator: never set a rate based on page views against one based on survey views.
An unusually high rate is worth checking as well. If one page gets five times the rate of its neighbours, something is sending readers there to answer, such as a link from a forum thread or a step that has stopped working.
Example
Example: a gardening site checks its rate
Say you run a gardening site, and every guide ends with "Did this guide help?" followed by "What was missing?". Last month your guide to pruning roses had 4,000 page views and 60 responses. That is 1.5%, and it looks poor next to a 7.64% widget median.
Then you check your analytics. Only 1,500 of those views scrolled far enough to reach the end of the guide. Divided by those, the rate is 4%. That is closer to the benchmark, and it uses a similar denominator: readers who could actually see the question.
Your other pruning guides sit between 1% and 2% of page views, so the rose guide is normal for your site. One guide, on hedges, is at 6%. Its referrers show a link from a gardening forum thread where people were arguing about its advice. You read those answers on their own, knowing where they came from, before you change anything.
The rose guide does not need a higher rate. It needs enough answers to decide something, and the sample size calculator shows how many that is.
How to raise your rate without skewing it
A higher rate is only better if the extra answers come from ordinary readers. These changes tend to bring in more of them:
- Make the first question one tap, such as Yes / No. Typing comes later and stays optional.
- Ask fewer questions. Two or three is enough for a page.
- Put the question where readers finish, and keep it in the same place on every page.
- Ask about the page they are reading. A specific question is easier to answer than a general one.
Prizes and repeated pop-ups can raise the rate too, but they bring in readers who answer for the prize or to make the pop-up go away. Their answers tell you less.
Where Asklet fits
Asklet shows a short survey on the page, with a one-tap first question, and keeps the path each answer came from. Results show the number of responses for each page, theme and website for the dates you choose.
It does not count widget loads or store anything that identifies a reader, so it cannot report a response rate itself. Take the page views for the same dates from your analytics and divide, as above.
Your next steps
-
Decide which denominator you will use
Page views, or views that reached the survey. Write it down and keep it the same.
-
Work out the rate for your ten busiest pages
Responses divided by page views for the same dates.
-
Compare them with each other first
Look for pages far above or below their neighbours, and find out why.
-
Check that you have enough answers
A rate is only useful if the page has enough responses to decide something.
-
Change one thing at a time
A shorter survey or a different placement, then compare the same number of weeks before and after.
Try these steps on your own site. Asklet adds the questions to every page with one snippet. Free plan, no card needed. Compare the plans.
Start freeCommon questions
What is a good response rate for a website survey?
In Survicate's 2025 benchmark, the median for website widget surveys was 7.64% of survey views. If you divide by page views instead, expect a lower figure, because many readers never reach the survey. Compare with your own pages over time rather than with one benchmark.
Why are email survey response rates so much higher?
Mostly because of what they are divided by. An email survey goes to people on your list, who already know you, and its rate is a share of those recipients. A website survey is shown to everyone who visits, including people who leave after a few seconds.
Does Asklet show a response rate?
No. Asklet does not count how often the widget loads, so it cannot tell what share of readers answered. Divide a page's responses by its page views from your analytics for the same dates.
Does a low response rate make my results useless?
Not on its own. Pew Research Center notes that response rates are an unreliable guide to accuracy. What matters more is whether the readers who answer are very different from those who don't, and whether you have enough answers for the decision you want to make.
Related
Sources
More from the blog
- Privacy and law EU cookie consent changes: where the Digital Omnibus stands in October 2026 9 min read
- Privacy and law Session replay and pixel lawsuits: what website owners should know 9 min read
- Privacy and law How to make a feedback widget accessible, and what the European Accessibility Act asks 9 min read