How long should a B2B survey be? What the research says

How long should a B2B survey be?

As short as the objective will allow (ideally 10 minutes or less, rarely more than 20). That's a judgment rather than a published standard, but it reflects what respondents tell researchers and what the studies show about response and drop off as surveys get longer. As a rule of thumb:

• 10 minutes or lessA good target for most B2B online surveys.
• 10 to 20 minutesCheck that every question earns its place, and that the incentive fits the time you're asking for.
• Over 20 minutesOnly with a very clear reason, a measured length and extra quality checks.

Two studies asked members of online panels directly about the ideal survey length and the most they'd accept:

Averages were a bit higher than the medians (ideal 12.6 to 15.4 minutes, maximum 22.7 to 28.7 minutes).

Neither study asked business respondents. The closest thing to B2B evidence is from a 2022 study by Seshadri and Broekemier. They asked 248 small business owners and managers in the US how likely they would be to complete an online survey of 10, 20, or 30 minutes, with six different incentive options (one of which was no incentive). On a scale from 1 (not at all likely) to 7 (very likely), the average rating across the six options was about 4.6 for the 10 minute survey, 3.4 for the 20 minute survey and 2.5 for the 30 minute survey.

Darker means more likelyOutlined: the two ratings compared below

A 20 minute survey with $20 cash came in at roughly the same level as a 10 minute survey with no incentive at all (3.77 vs. 3.80). The gift card and the donation both performed very much like cash at each length. The offer of the survey results did almost as well as cash at 10 minutes but fell further behind at 20 and 30 minutes. The authors concluded that length had a greater impact on willingness to participate than incentives did.

Two caveats. The study measured what people said they would do rather than what they actually did. Less than 1% of the owners and managers the researchers contacted took part, so the people who responded may not be representative. Treat the results as indicative of the likely direction rather than as a prediction of actual response rates.

How does survey length affect response rates?

Longer surveys tend to get fewer responses, but the size of the effect varies, and both the length you state in the invitation and the real length matter.

• Controlled trialsIn the 2023 Cochrane review of methods to increase response, a shorter questionnaire raised the odds of response to electronic questionnaires by about 51%, though results varied a lot between trials. Higher odds aren't the same as that many more percentage points: if 20% of people would respond to the longer version, 51% higher odds brings that to about 27%.
• Stated lengthGalesic and Bosnjak (2009) told people, at random, that a web survey would take 10, 20 or 30 minutes. The longer the stated length, the fewer people started and the fewer completed it.
• Length and contentA 2011 meta-analysis by Rolstad, Adler and Rydén of 20 studies found that longer questionnaires had lower response rates. The studies couldn't disentangle length from content, so length alone may not be the explanation for the gap. The authors recommended selecting questions on the basis of their content rather than their length.
• A recent testIn a 2025 experiment by Hanson and colleagues, a general population survey was carried out in Austria, with half the sample given the shorter version (announced as 35 minutes, 182 questions) and the other half the full version (announced as 50 minutes, 274 questions). Response was 38% for the shorter version and 34% for the longer one. The difference was statistically significant but very small, and the authors found no clear difference in who responded or in the quality of the responses.

While reducing survey length generally contributes to higher response rates, the increase from a single reduction may be modest. And a higher response rate doesn't automatically mean a more representative sample, as discussed in our guide to B2B survey response rates. Length also affects cost and data quality. Shorter surveys cost less per complete and can help maintain data quality by reducing respondent fatigue, which strengthens the case for keeping surveys as short as possible.

Does survey length affect data quality?

Often, and predominantly at the end of the survey. Galesic and Bosnjak rotated blocks of questions so that the same questions appeared early in the survey for some respondents and later for others. Answers to questions that came later in the survey were faster, shorter and more uniform than answers to the same questions that came earlier. In practical terms, that equates to less time per question, shorter answers to open ended questions and less variation in answers across a grid (at the extreme, this is straightlining). This reduction in effort later in the survey is often referred to as respondent fatigue, or survey fatigue.

The effect isn't guaranteed. In the Austrian experiment, skipped questions and identical answers in the grids were about the same in the 35 and 50 minute versions. But the risk is concentrated toward the end, where people get tired, so you do need to plan for it:

• Put the most important questions first. These are your key measures and any open ended questions you intend to quote. Asking them early protects them from the fatigue that can lead to less careful answers toward the end.
• Check the later sections separately. Compare the time per question and the level of straightlining in the final section with that in the first. A respondent who rushes through the end may still have answered the first section carefully.
• Use the same quality checks as any B2B survey. Our guide to preventing survey fraud covers speeding, straightlining, and open end checks, and how to avoid removing honest respondents.

Does survey length affect drop off?

Drop off, also called breakoff, means starting a survey and not finishing it. Longer surveys tend to lose more people along the way.

• Across real surveysLiu and Wronski (2018) analyzed more than 25,000 real web surveys and found that completion rates fell as surveys got longer and the questions got harder. These were comparisons between surveys, not experiments, so other differences between them may play a part.
• On phonesA meta-analysis of 14 studies by Mavletova and Couper (2015), covering 39 samples of mobile web surveys, found an average breakoff rate of 6.6%. Compared with a 5 minute survey, a 30 minute survey had about 42% higher odds of breakoff, though the length effect fell just short of statistical significance. Surveys designed for mobile had lower odds of breakoff, and stating the expected length in advance made no measurable difference to breakoff.

Why do surveys take longer on a phone?

Couper and Peterson (2017) found that web surveys take longer on smartphones than on computers, whether people chose their device or were randomly assigned one. The time taken to move between pages accounted for about a fifth of the difference. Much of the remainder was attributable to the extra scrolling required on a phone, particularly with grid questions.

A survey that takes 12 minutes on a laptop will generally take longer on a phone (particularly if it has large grids). If some of your respondents are going to be answering on a phone:

How to measure survey length (LOI)

Length of interview (LOI) is how long the survey takes a typical respondent, usually reported as the median in minutes. It's the number a panel supplier quotes and plans on, and the number your invitation should state.

1. Time it with real respondents. Your team knows the questions, so their timings run short. Use the median time from a soft launch, the first slice of fieldwork.
2. Use the median, not the average. A few people leave a survey open in another tab, and their times pull the average up.
3. Time each path. If different groups see different sections, check the median for each one. The longest path is the real length for the people on it.
4. Include the screener. Qualified respondents answer it too.
5. State the real length in the invitation. Understating it may get more people to start, but it breaks the promise you made to them, and a supplier's price and field plan depend on the length you give.

How to shorten a B2B survey

Long surveys often include questions that don't change any decision. Before cutting individual words, cut whole questions.

• Tie every question to a decision or an analysis. If no one can say what they'd do differently based on the answer, drop it.
• Don't ask what you already know. Panel suppliers often hold profile data such as company size, industry, and job function. Ask what they can append to the data instead of asking again, and keep the questions that confirm someone qualifies in the screener.
• Use skip logic so people see only the questions that apply to them.
• Trim grids. Remove items that measure the same thing, and split long grids into shorter ones.
• Limit the number of open ended questions to reduce the burden on respondents, and put the essential ones at the beginning of the survey. In Galesic and Bosnjak's experiment, answers to open ended questions were shorter when the questions came later, so asking early gets you fuller answers.
• Move nice to know topics to a later study, or show each respondent a random subset of modules. A subset design needs more completes to get a usable base for each module, so check the numbers with a sample size calculator first.

Don't split one long survey into two requests to the same people. In an experiment by Peytchev and colleagues (2020), asking for a survey in two parts, with the request for the second part made at the end of the first, led to more nonresponse, not less.

When a long survey can't be avoided

Some designs need time. A conjoint study, a MaxDiff exercise, or a detailed usage audit can run past 20 minutes even after cutting. In that case:

• Say so up front, and pay for the time. Our guide to B2B survey incentives covers setting the amount.
• Put the core exercise early, before the profiling and background questions.
• Design for phones, or tell respondents in the invitation if the task is easier on a larger screen.
• Check the late sections for speeding and straightlining before you accept the data.
• Budget for it. Panel suppliers price on incidence and length, so each extra minute raises the cost per complete. Our cost per complete benchmarks use a rule of thumb of about 3% per minute beyond 10 minutes, and the cost per complete estimator shows the effect for your audience.