Product Research Survey Template for a Product People Already Use
Product research is four different surveys. Which one you can honestly run is decided by what the people answering have already done, and the table below settles that in one read. Then here is the survey most people came for: 14 questions for people who have used the product. Answer it below, copy it, or open it in the editor.
- 14questions
- About 5 minto complete
- 4 routesone decision table
- Current userswho it is for
Which product survey you can honestly send
Four surveys go by the name product research, and they are not versions of one another. Each needs a different group of people in the sample. Find the row that describes the people you can reach.
| What the people you can reach have already done | What they can answer | Which survey to send |
|---|---|---|
| Read a description, seen a mock-up or watched a demo. Nothing is built yet. | Whether the idea appeals, what they think it is for, and what would stop them. | Send the concept testing survey questions. |
| Used the product. Recently, and for long enough to have an opinion of it. | What they use, what works, what is missing, and what to fix next. | This page. The 14 questions below. |
| Completed an order in the last few days. | How the buying went: finding it, checkout, delivery, and the first look at it. | Send a post-purchase survey while it is fresh. |
| Bought in the category, and has never been asked about your product by name. | How big the category is, who is in it, and what they buy instead. | Run a market survey, which names no product. |
The rule the table is built on. A question about a product can only be answered by somebody who has used the product. That is not a finding, it is what the words mean. "Which part would you drop" is unanswerable by a person shown a picture, and "would you buy this" is a guess put to somebody who already owns it.
Published question lists mix the stages, which is where this goes wrong. One list puts "what is holding you back from trying our product" three items from "what current features do you find the most valuable", addressing a stranger and a daily user on one form.[5] Another puts concept items about a product that does not exist yet beside "how satisfied are you with our product".[6] A third recommends five questions, then heads its own bank "70+ Product Survey Questions by Category".[7] If half your answers describe something people use and half react to a description, the two halves are not answering the same question, and their average means nothing.
Three more things are not on this survey, on purpose. For a mobile or web app, the app feedback questions go further and cover crashes, updates and the store rating. For a price, use the product pricing survey: asking what somebody thinks of a product and what they would pay in one run makes both answers worse. For one number to track release on release, that is the recommend question, and it works alone on the net promoter score survey. Net Promoter, NPS and Net Promoter Score are registered marks of Bain and Company, NICE Systems and Fred Reichheld.[4]
The 14 product survey questions
Every question, the scale it uses, and what an answer tells you. Copy one block, copy all 14, or open the survey in the editor and put your product's own words into it.
Who is answering
Two taps first, so every rating below can be split by how long somebody has used the product and how often. Without them, a daily user and a once-a-month user land in the same pile.
- How long have you been using it?Tenure. Split every other item by this before reading any of them: a first-week answer and a third-year answer are about two different products.
- How often do you use it?Weight. It tells you how much of the product a person has actually seen, and therefore how much their answer to question 8 is worth.
What it is for them
Three statements on one 5-point agree scale: the job it was bought to do, whether it does that job every time, and the verdict.
- It does the job I got it for.Basic fit. Disagreement from somebody with a year behind them means the thing you thought you sold is not the thing they use.
- I can rely on it to do that job every time.Reliability, asked as its own item rather than folded into the one above, because a product that works four times in five can still be rated satisfactory.
- Overall, I am satisfied with this product.The headline. Report it as the whole spread of answers, not as an average; the next section says why.
The parts they touch
One written line and one rating, which together separate the part carrying the whole product from the part people abandoned.
- Which part of it do you use most?The one thing you cannot break. Count the answers; the winner is what your roadmap is really protecting.
- There is a part of it I tried and gave up on.Abandonment, which a feature-use dashboard shows as a flat line with no reason attached. Agreement here plus a name in question 12 gives you the reason.
Fit and friction
Three statements about the gap between what the product does and what the person needs. The third is the one a roadmap can act on.
- It has everything I need it to have.Completeness. Split it by question 1 before treating a low answer as a roadmap item, since the two groups have seen different amounts of the product.
- It fits the way I already work.Whether the product bends to the person or the other way round. Disagreement alongside a high satisfaction answer means somebody is putting up with a shape that is not theirs.
- I have had to work around something it does not do.The workaround item, and the most actionable line on the form. Agreement is a specific missing capability that somebody is already paying for in time.
Direction, then the close
The two open boxes sit before the end rather than at it, because answers to questions positioned later in a questionnaire come back shorter and more alike than answers given near the beginning.[1] It closes on a rating, the cheapest question to answer.
- It has changed for the better since I started using it.Direction. It only means something from somebody with tenure, so read it split by question 1 and drop the newest group.
- What is the one thing you would change about it?The change list. Tag each answer to a part of the product and count the tags before anyone argues about them.
- What is it good at that you would not want changed?The do-not-touch list. It is the answer that stops a redesign from removing the thing people stayed for.
- I expect to still be using it a year from now.The closing rating. Everyone answering is a current user, so treat a fall here as the only forward-looking signal the survey can honestly give you.
Every item here needs a user to answer it. That is the constraint the page is built on, and it is why nothing asks whether somebody would buy it, what they would pay, or anything answerable from a screenshot. All 14 are in the survey builder already: open it, put your product's name into the wording, and cut anything you would not act on.
Keep the required split if you edit it. Questions 1 and 2 are required, because every other answer is read through them. The two open boxes are optional, because a required text box is where a survey loses the people who had nothing to add.
Who to send it to, and when
The question the rest of this subject skips, and the one that decides whether the answers mean anything. Four decisions belong to this survey; the rest of a send sits with the general survey research methods.
Build the list from use, not from your customer list
The people who can answer this are the ones who have used the product recently. A month of activity is a fair line for something used weekly; for something used once a quarter, stretch the window. Anyone outside it is guessing, and a guess is worse than a blank: it looks like data.
Send it to five people first
Send it to a handful who match the audience and ask where they hesitated. Pretesting the parts of a survey before it goes into the field is a requirement in federal statistical practice, not a nicety,[2] and here it costs an afternoon.
Not in the week after a big change
A wave sent days after a release is a wave about the release. Either wait until the new thing is ordinary, or send it deliberately after one and label the wave so nobody compares it with the last by accident.
Say in one line who reads the answers
The template asks for no name and no email, so nobody can be contacted about what they wrote. Put that in the invitation. It is why people answer bluntly, and why you cannot follow any of it up.
Reading the answers, and who is missing from them
Two things go wrong at this stage, and neither is fixed by collecting more responses.
An average hides the shape. A rating that averages 3.4 can be everyone sitting on 3 and 4, or half the room at 5 and half at 2. Two different products, two different jobs to do next, one mean. Report the whole spread for every rated item, with the counts. Then tag each written answer to a part of the product and count the tags: one vivid comment is one person, five naming the same thing is work.
Everyone who left is missing by construction. The list is built from people using the product now, so anyone who tried it and gave up is not in the sample. Nothing inside the survey fixes that. Run it for two years and the ratings can climb while the business shrinks.
What that costs depends on who is missing, not just how many. The size of the error turns on the response rate and on how far the people who did not answer differ from the people who did. The US federal standard for statistical surveys treats a response rate below 80 per cent as the point at which that difference has to be investigated rather than assumed away.[2] Whether a rate relates to bias at all has been examined across fifty-nine methodological studies, because the answer is conditional.[3] Neither is a pass mark for a product email. Both are the reason to record who you sent it to, and how many answered, beside the numbers.
Product survey questions: FAQ
What is the difference between a product survey and a market research survey?
The respondent. A product survey asks people about a thing they have used, and names it. A market research survey asks about a category and names no product, so somebody who has never heard of you can answer it. The two cannot be merged: different people, and answers that do not pool.
Who should you send a product survey to?
People who have used it recently, which for most products means the last month. Two things follow. Everyone who tried it and left is off that list, so you are asking a room you have already filtered. And anyone who has not used it will guess rather than skip, which is worse than a blank.
How many questions should a product survey have?
Fewer than you want. The evidence is about length, not a magic number: the longer a questionnaire is stated to be, the fewer people start and finish it, and answers to questions placed later come back faster, shorter and more alike.[1] This one is 14 questions, two of them optional, at roughly five minutes. If you add one, take one out.
What is the product-market fit question, and should I use it here?
It is the item asking how you would feel if you could no longer use the product, and it belongs to a different survey with a different purpose: deciding whether to keep building at all, rather than what to fix next. It is left off this template on purpose. The reading attached to it is a rule of thumb rather than a measured finding, so no threshold for it appears here.
This asks somebody who owns the product and has lived with it for weeks. A single sitting with a coded sample is a different instrument, and the sensory evaluation form is written for that moment.
Product feedback covers what to fix and what to build next. Whether the price itself still stands up is a separate question, and the price questionnaire keeps it separate.
Methods and sources
What this template is based on
The length of the questionnaire and the position of the open boxes follow the published finding that fewer people start and finish a longer questionnaire, and that answers given later come back shorter and more alike[1]. The instruction to pretest before the full send, and the treatment of a low response rate as something to investigate rather than ignore, follow the US federal standards for statistical surveys[2][3].
Question licensing
There is no licensed instrument in this subject, and all 14 questions were written for this page, so they are free to copy, edit and send anywhere. Net Promoter, NPS and Net Promoter Score are registered marks of Bain and Company, NICE Systems and Fred Reichheld; the recommend question is named on this page and never reproduced in the survey[4].
Limits and disclosure
No benchmark is offered, because there is no defensible one to offer: a product survey has no standard scale and no comparable population. Published guidance on how many questions a product survey should carry disagrees with itself on the same page[7]. And the sample this template can reach is people who are still using the product, which is a limit of the design rather than of the wording.
SuperSurvey makes the editor this template opens in. The questions are free to copy and use anywhere, including somewhere else.
References
- Galesic, M. and Bosnjak, M. Effects of Questionnaire Length on Participation and Indicators of Response Quality in a Web Survey. Public Opinion Quarterly 73, no. 2 (2009): 349-360. academic.oup.com
- Office of Management and Budget. Standards and Guidelines for Statistical Surveys, September 2006. Standard 1.4 and Guideline 3.2.9. obamawhitehouse.archives.gov
- Groves, R. M. and Peytcheva, E. The Impact of Nonresponse Rates on Nonresponse Bias: A Meta-Analysis. Public Opinion Quarterly 72, no. 2 (2008): 167-189. academic.oup.com
- Bain and Company. About the Net Promoter System. netpromotersystem.com
- Sprig. Top Product Survey Questions. Sprig blog. sprig.com
- Qualaroo. Product Testing Surveys. Qualaroo blog. qualaroo.com
- Formbricks. Product Survey Questions. Formbricks blog. formbricks.com
Send it to the people who have used it
Fourteen questions, about five minutes, no name asked. Open it in the editor and make it yours.
Use this template