Find the assumption that could sink your plan
Put what you believe on the map, drag each note to where it belongs, and see what to test first.
Pick a test for what you need to learn
Nine tests that cover most situations. Set the pass signal before you start, so you can't talk yourself into a result afterwards.
Do they want it?
Customer interview
What people say · 1 week
Talk to 5 people who hit this problem in the last month. Ask them to walk you through the last time it happened. Don't pitch anything.
Passes if: At least 3 of 5 describe the problem without prompting and have already tried to fix it.
Run the conversation as an AI interview in Vouch, then add what you learned here.
Fake door
What people do · 1 week
Offer the thing before it exists, with one clear button. When clicked, say it isn't ready and ask to notify them.
Passes if: At least 5% of visitors click, and half of those leave an email.
Specky builds the page and collects the responses. You bring the visitors.
Concierge
Money or a signature · 3 weeks
Deliver the outcome by hand for 5 customers, as if the product existed.
Passes if: At least 3 of 5 come back for a second round or agree to pay.
Specky builds the page and collects the responses. You bring the visitors.
Can we build it?
Partner and supplier interviews
What people say · 1 week
Talk to 3 partners or suppliers this depends on. Ask what it would take, what it costs, and where it usually breaks.
Passes if: At least 2 of 3 confirm they can deliver at the cost and speed you need.
Run the conversation as an AI interview in Vouch, then add what you learned here.
Wizard of Oz
What people do · 2 weeks
Give customers what looks like the product, while a person does the work behind it.
Passes if: At least 3 of 5 finish the job and don't notice or don't mind.
Specky builds the page and collects the responses. You bring the visitors.
Extreme programming spike
Early signal · 3 days
Time-box a technical probe of the riskiest part. Write down what is now known and what is still open.
Passes if: The risky part works at the speed and cost you need, inside the time box.
You run this one. Specky keeps the plan and the result.
Will it pay?
Pricing survey
What people say · 1 week
Ask the segment at what price this feels too cheap, a bargain, expensive, and too expensive.
Passes if: Your planned price sits inside the range at least half consider acceptable.
Specky builds a one-question page and collects the answers. You bring the people.
Mock purchase intent
What people say · 1 week
Show the real price and ask people if they would buy, without taking payment. Try two prices with two groups.
Passes if: At least 5% say yes at the higher price.
Specky builds the page and collects the responses. You bring the visitors.
Letter of intent
Money or a signature · 4 weeks
Ask 10 buyers with authority for a short, non-binding letter saying they would pilot or buy.
Passes if: At least 3 of 10 sign.
Vouch runs the buyer conversation. Getting the letter signed is yours.
Keep your map connected to your tests
In Specky, picking a test writes the plan for you, and each result moves the note and updates the bet resting on it.
Try it in SpeckyWhat is an assumption map?
An assumption map is a chart of everything that has to be true for a product or plan to work. One axis is how badly the plan breaks if the belief turns out to be wrong. The other is how much evidence you have for it. Anything high on impact and low on evidence goes to the top of the list, because it could sink the plan and nobody actually knows whether it's true. Everything else can wait.
How to run an assumption mapping session
Get the people who build, sell, and support the product in one room for an hour. Write down what must be true for customers to want this, for your team to be able to build it, and for it to make money. One belief per note. Place each note together, and expect arguments about where they go: those arguments are the point, because they show where the team is guessing. Then circle the three to five notes in the top right and agree on one test for each before anyone leaves.
How to turn an assumption into something you can test
"People will like it" can't fail, so it can't teach you anything. A useful assumption names who, what, and how much: "Teams of five to fifty will pay €2,000 for this" can be proven wrong by next month. If a belief feels too big to test, split it into smaller ones that must all be true. If any one of them fails, the big one fails too, and it's far cheaper to find that out on a small piece.
Why most assumption maps go stale
A map on a wall is right on the day it's drawn. Tests finish, results land in a doc somewhere, and the notes stay exactly where they were. Specky keeps the map connected to the work: record a result and the note moves on its own, and any bet resting on it shows whether the evidence still holds up.
Assumption mapping questions
The short answers teams ask for when they run this for the first time.
What is the difference between an assumption and a hypothesis?+
An assumption is something you're treating as true for now and accepting the risk that it isn't. A hypothesis is an assumption you've decided to test, written precisely enough that a result could prove it wrong.
How many assumptions should we test at once?+
Three to five. The goal of mapping is to find the few beliefs that could sink the plan. Testing everything spreads the team thin and delays the answers you need most.
What counts as strong evidence?+
What people do beats what people say. An interview is a useful early signal, someone trying to use a feature is stronger, and money or a signature is strongest. Treat a belief as supported only when people have acted on it.
What should we do when an assumption turns out to be wrong?+
Stop and decide whether to change the plan, aim it at a different customer, or drop it. A wrong assumption found early is the cheapest bad news you'll ever get. If the result is unclear, run a sharper test or accept the risk and move on.
More free tools
Free AI PRD Generator
Generate a production-ready PRD in 60 seconds.
Feedback → Spec
Turn one piece of customer feedback into a defensible spec.
JTBD Interview Template
21 Jobs-to-be-Done interview questions, free download.
Product Roadmap Template
Now / Next / Later roadmap template, free CSV download.
RICE Prioritization Calculator
Score and rank your backlog with Reach × Impact × Confidence ÷ Effort.
Vibe Check Your Pitch
An honest PM gut-check on your landing page or pitch.