How many questions in a survey? What 49 example surveys ask

A good survey asks fewer questions than you have. Counted on 13 September 2026, the 34 survey forms built on this site ask a median of 10 questions and tell the respondent it takes 4 minutes, and none of them asks more than 13. The question banks they are built from hold 20 to 40 questions each, and across the 9 survey and feedback examples the median is 9 questions. The gap between the bank and the form is the finding: a survey is chosen from a longer list, not written to the length of it.

The 49 surveys and 40 banks behind the numbers

Two kinds of page are counted, and they measure different things.

The first is the 40 survey question banks under /form-questions/: employee, customer and education surveys, 909 questions between them, each tagged with its field kind and whether it is required. A bank is a menu. It holds 20 to 40 questions by design so a reader can pick the ones that fit. 34 of the banks also carry a small finished form built from a subset of 9 to 13 questions, and the other 6 point to an existing form example instead.

The second is the 9 survey and feedback form examples under /examples/forms/. The 40 banks and the 9 examples are the 49 counted in the title; the 34 built forms and the 9 examples are the 43 surveys a respondent would actually see.

Minutes come from the estimated time each form shows on its welcome screen, which is what the respondent is told before they start. It is a stated estimate, not a timed average. On a form, a rating grid counts as one question however many statements it holds; in a bank, each statement is listed separately, which is why banks look longer than the forms drawn from them.

These are the surveys this site contains, not a sample of surveys in general, and the set grows as banks and examples are added.

Question counts by survey type

The table under this article has the medians for each type.

Employee surveys have the largest banks, a median of 22 questions, because staff surveys measure many things at once and lean on agreement grids. The employee engagement bank is the longest in the set at 40 questions: 35 grid statements across the role, the team, the manager and the company, three written answers and two dropdowns. Its built form still asks 10 questions and tells people 12 minutes, the longest estimate of any built survey, because each grid is a screen of statements.

Customer surveys have banks of 21 and forms of 11.5, but the shortest estimates, a median of 2.5 minutes. Customer questions are mostly single choice with almost no grids, 2% of the bank questions, and customers owe the business nothing, so every question has to be quick.

Education surveys sit at 20 questions in the bank, 10 on the form and 4 minutes.

The examples split cleanly in two. The five short feedback forms, including the customer feedback survey, the event feedback form and the pulse survey, have a median of 6 questions and 2 minutes. The four reviews and interviews, a 360 review, a performance review, a self assessment and the exit interview form, have a median of 13.5 questions and 12.5 minutes, because each is a one off conversation somebody has set time aside for.

How many questions are required

Across the 909 bank questions, 36% are marked required, far below the 76% across all 140 form examples on this site. The typical survey asks a lot and insists on little.

The required share splits by kind of question. Grid statements are required 86% of the time, because a half answered grid cannot be averaged. Single choice questions are required only 16% of the time and long text answers 20%. Customer banks require just 17% of their questions against 44% for employee banks and 47% for education banks, and one wellbeing bank requires none.

The built forms sit in between, with 40% of their questions required. A required mark on a survey is a statement that the result is unusable without that answer, and most survey answers are not like that.

Open questions, and how few there are

Open questions, meaning long and short text answers, make up 13.6% of bank questions. Every one of the 40 banks has at least one, and the most common number is two, found in 18 banks. The built forms carry a median of 2 as well.

The exceptions show where open answers belong. The performance self assessment bank has 15 text questions out of 24, because a self assessment is written rather than rated. Reviews and interviews among the examples carry a median of 6 open questions against 3 for short feedback forms.

Placement follows a rule too: in 15 of the 40 banks the final question is a long text answer, asking for the reason behind the ratings already given. Pew Research Center’s methods work shows that an open question and a closed version of the same question produce different answers, which is why the open box comes after the scale rather than replacing it.

The rating scales in use

Nearly half of all bank questions, 47%, are single choice, and a further 24% are statements in a rating grid. Surveys are built from scales.

On the 34 built forms, the 163 single choice questions settle on two scale lengths: 72 offer five options and 52 offer four. Three options appear 18 times, and eleven options, the zero to ten Net Promoter Score scale, twice. Rating grids appear 41 times across the 34 forms, and they are how an employee survey covers six statements on one screen instead of six.

The employee pulse survey shows both in six questions: a grid of six statements on the same agreement scale, and a separate single choice question about workload with four options. A grid collects either one answer per row or several, and it exports as a single column with each statement and its answer joined.

A length rule you can use

Put the numbers together and a working rule falls out.

1. Start from a bank, not a blank page, and pick the 8 to 12 questions whose answers you will act on. The banks here hold 20 to 40 because the choosing is the work.
2. Count a grid as one question but budget time for every statement. Six statements on a scale take longer than one question, and the engagement form’s 12 minute estimate is what that looks like.
3. Tell people the time, and keep it near 4 minutes for staff and students or 2 to 3 for customers. SurveyMonkey’s analysis of about 100,000 surveys found completion falling once a survey passes seven or eight minutes.
4. Require the measure, not the reasons. Make the core rating required and leave the open questions optional, with one of them last.

For the wording of each question, how to write survey questions covers leading and double barrelled questions, and the page on survey fatigue covers what happens to answers as a survey drags on. GOV.UK’s guidance on question pages, one question per page, is the logic behind conversational surveys.

The rating matrix form page shows the grid used in most of these surveys, and the guide to creating a conversational form with AI builds a short one question at a time survey from a description.