Average completion time

Average completion time is the mean time respondents take to finish a survey, from opening it (or answering the first question) to submitting the last answer, calculated over completed responses only. It shows how long the survey really takes, helps you promise an honest length in the invitation, and flags responses that were too fast to be read properly. Because a few people leave a survey open in a tab, the median completion time is often a better summary than the mean.

Completion time is read together with the completion rate and the dropout rate: a survey that takes much longer than promised usually loses more people along the way.

How to calculate it

Average completion time = sum of completion times of all completed responses ÷ number of completed responses

Partial responses are left out, since people who quit after 10 seconds or after 8 minutes would distort the figure in either direction. Define the start point (first page opened or first answer) and keep it the same across surveys.

Worked example

Eight people complete a 12-question survey. Their times in seconds:

150, 170, 180, 200, 210, 240, 260, 1,500

• Mean: 2,910 ÷ 8 = 364 seconds, about 6 min 4 s
• Median: the middle of the sorted list, (200 + 210) ÷ 2 = 205 seconds, about 3 min 25 s

One respondent left the tab open for 25 minutes. That single value almost doubles the mean, while the median stays close to what most people experienced. In the invitation, "about 3–4 minutes" is the honest promise here, not "about 6".

How to interpret it

• Compare it with what you promised. If the invitation says 5 minutes and the median is 11, fix the estimate or shorten the survey. A misleading estimate damages trust and increases survey fatigue.
• Look at the distribution, not one number. Report the median and, for long surveys, the 25th and 75th percentiles. A wide spread often comes from branching logic: people on a short path and a long path are averaged together.
• Watch for speeders. Responses far below the median are often given without reading. Set a rule before fieldwork, for example flagging anyone who finishes in less than a third or half of the median, then check those responses for straightlining and empty open answers before excluding them.
• Segment it. Times on phones and computers, or for different respondent groups, can differ. A breakdown explains more than one overall figure.

What affects completion time

• The number of questions and the share of open-ended and matrix questions
• Logic that skips or adds questions for some respondents
• Complex wording, long instructions and reading-heavy screens
• The device and the setting in which people answer

Common mistakes

• Including partial responses in the calculation.
• Reporting only the mean when a few very long sessions are present.
• Deciding the speeder cut-off after seeing the results. Fix the rule in advance and write it into the methodology.
• Skipping the pilot. A pilot survey gives a realistic time estimate before the main fieldwork, when shortening is still cheap.
• Treating a long time as engagement. It can mean careful reading, or it can mean confusion and interruptions. Look at where the time goes.

How to estimate and measure it

1. Before launch, estimate the length from the number and types of questions with the survey length calculator.
2. In the pilot, time 5–10 test respondents from your audience and compare with the estimate.
3. During fieldwork, export the start and submit timestamps of completed responses and compute the median, mean and percentiles.
4. In the report, write something like: "Median completion time was 3 min 25 s (n = 412 completed responses); 9 responses under 70 seconds were excluded as speeders."