Customer research
List assumptions before researching a product idea
Write down every belief the idea depends on and mark which ones you have seen for yourself and which you hope are true. Rank each by what it would cost if wrong, and attach a test that could prove it mistaken before you commit design or build time.
A text to start with
I am considering [product idea] for [audience]. Here is what I currently believe about who needs it, what they do today, why they would switch and what they would pay: [notes]. List each belief as a separate assumption. For each one, say what would change in my plan if it were wrong, what evidence I already have, and a small test that could show it is false.
Example request. Change the details to fit your situation.Assumption register
Product idea: [ ] Intended audience: [ ] Decision this list supports: [ ] Assumption (one claim only): [ ] Who holds it and why: [ ] Consequence if wrong: [ ] Evidence so far, with source: [ ] Evidence quality (seen, heard, guessed): [ ] Possible test: [ ] Result that would show it is false: [ ] Decision the result would change: [ ] Assumptions to test first: [ ] Date to review this list again: [ ]
Beliefs written as facts
Look for sentences that sound settled but rest on one conversation or on your own habits, such as "small teams hate spreadsheets" or "people will pay for speed." Rewrite each as a claim that could be checked. Then read the proposed tests and ask whether any result would make you change course. Friendly opinions about whether an idea sounds good may tell you little about the target audience's actual behavior. Ask which observation would challenge the specific belief. Split any line that bundles claims: separate the parts when each needs different evidence.
Ask another agent to check the result
Read this assumption register. Flag every entry that states a belief as established fact and suggest a checkable rewording, using only the evidence I supplied. Identify tests where no possible result would contradict the assumption. Point out entries that bundle several claims and show how to split them.
Choose what to test first
Sort the register by consequence if wrong and by how little evidence you have. The assumptions that score high on both justify a round of research. Write the decision each answer would inform, since GOV.UK ties research rounds to actionable questions and decisions. Plan a small test for each, and set a date to revisit the list.