A product research survey puts a number on what customers need. How often the problem happens, what they do about it now, which features would actually change their mind, and what they would pay. We write the questions, field them on our own proprietary platform to people screened to your criteria, and report the numbers against the decision you are making.
Book a free 30-min callAsk people to rate ten statements from one to five and you get rows of fours, a quarter of the sample picking the same answer down the whole grid, and no way to tell the two groups apart. The grid is comfortable to build and almost impossible to act on.
We write for the decision instead. Questions about what people did, not what they think they would do. Select-all rather than agreement scales. A real option for people the question does not apply to, so they are not forced into an answer that then pollutes the base. The result is shorter to answer and far easier to read.
Five interviews tell you a problem exists. A survey tells you whether it affects most of your market or a vocal few, which is the difference between a roadmap item and a distraction.
The same question read by segment, so you can see that the feature everybody asked for is wanted by one type of buyer and nobody else.
Pricing sensitivity, and the point at which people stop. We also see where they abandon the survey itself, question by question, which tells you which question they were not willing to answer.
About fifteen questions. Length drives drop-out and cost, so the budget is set before drafting rather than trimmed afterwards. If a question does not change a decision, it does not go in.
Frequency and severity in the respondent’s own week, not in the abstract.
The current workaround, which is your real competitor far more often than another product is.
Select-all and forced trade-offs, so the list comes back ranked rather than all important.
Which framing lands, carried over from concept testing where the decision needs a head to head.
Where interest falls away, and which tier people choose when made to pick one.
What would have to be true before they moved, and what would stop them.
Their own words, themed and counted, so a quote arrives with a number attached.
Only the demographics you will actually split the data by. Everything else is wasted length.
We agree the decision first, then write the questionnaire against it and send it to you before anything is fielded. Participants are screened to your criteria, and you see who is taking part before fieldwork starts.
Fieldwork runs in the browser or over WhatsApp, depending on the market, and open text comes back themed and counted rather than as a spreadsheet of raw answers. Where a number raises a question a number cannot answer, we follow it with interviews. User research covers that half.
We built the platform ourselves, so a survey here is not limited to text questions. The things that usually need a separate study can sit inside the questionnaire.
A Figma prototype or a Lovable build embedded in the survey, so people react to the real thing rather than a description of it.
Packaging, screens, ad creative, listing images and video shown inside the question, with the choice captured alongside the reason.
People reply by voice or by typing, whichever they prefer, and voice notes come back transcribed. Open text stops being one-word answers.
Fielded on the channel the audience answers on, which in several markets is the difference between a response rate and none.
You also see where people drop off question by question, so an abandoned survey still tells you something. More about the platform.
The decision you are trying to make, and the audience. That is enough to start. If you already have a draft questionnaire we will review it and tell you which questions are doing work and which are padding, before anybody fields it.
A product research survey asks a defined audience about a problem, what they do about it now, which features would matter and what they would pay. It is used to size a decision rather than explore it, which is why the questions have to be written before fieldwork rather than cleaned up afterwards.
About fifteen. Length is the main driver of drop-out and of fieldwork cost, so the budget is set before drafting rather than trimmed afterwards. If a question does not change a decision you are about to make, it does not go in.
Around 100 for a single market read, more if you need to compare segments, because every split divides the base. One hundred responses read by three age bands is roughly thirty people per band, which shows direction rather than proof.
Yes, and it is often better. If you have a list, we write and field the survey to it and no recruitment is needed. Where you need people you do not have, we recruit and screen them to your criteria.
Yes. A Figma prototype or a Lovable build can be embedded in a question, along with images, designs, packaging and video, so people respond to the real thing. Surveys can be fielded in the browser or over WhatsApp, and people can answer by voice, which comes back transcribed.
No. Long agreement batteries produce rows of fours and a pattern of straight-lining that tells you nothing. We convert them into a single select-all question, which takes seconds to answer and gives a cleaner read.
Tell us the decision and the audience, and we will come back with a questionnaire, a sample size and a figure.