Goal
Test my startup idea before building it
You test a startup idea by putting its riskiest assumption in front of real customers before you build anything: ask them about their past, not your future, deliver the service by hand to a few of them, and see whether they commit time or money. Most ideas fail not because they were bad but because nobody checked; a week of honest conversations and one small experiment tells you more than six months of building.
What is really going on
The usual way to test an idea is to describe it to people and count the smiles. Friends, former colleagues and strangers at events all say it sounds great, because saying no to a hopeful founder is unkind and because a hypothetical future costs nothing to agree with. The founder collects the encouragement, treats it as validation, and builds. Months later the product launches to silence, and the founder concludes that the idea was wrong. Usually the idea was never tested; the questions were.
The second failure is subtler. Some founders do talk to customers, hear real problems, and then build the full solution they imagined rather than the smallest thing that would check whether people will actually change their behaviour. Wanting a problem solved and paying to solve it are different facts, and only an experiment with a price on it separates them. The full build tests everything at once, which means it tests nothing in particular, and when it fails there is no way to know which belief was wrong.
There is also the market question that almost nobody asks early: even if the idea works, who exactly is the first customer, and how will you reach them? An idea that needs a big, vaguely described market and a channel to be found later is two untested bets stacked on top of each other. The cheapest time to find out is before the code exists, and the books on this road exist because the founders who wrote them found out after.
Five ideas from five books
- 1
Ask about their past, not your idea
Rob Fitzpatrick's three rules turn a compliment machine into a fact machine: talk about the customer's life instead of your idea, ask about specific events in the past instead of opinions about the future, and talk less. When did this problem last happen, what did you do, what did it cost you? A person describing last Tuesday cannot be polite about it; a person imagining next year can. End every conversation by asking for something that costs them, fifteen minutes next week, an introduction, a deposit, because a commitment is the only proof that the problem matters. Twenty conversations run this way, before a line of code, are the cheapest test an idea will ever get.
- 2
Build the smallest thing that could prove you wrong
Eric Ries's minimum viable product is not a small version of the product; it is the smallest experiment that tests the riskiest assumption with real customers. Write your beliefs as a list, circle the one that kills the company if it is wrong, and design the least work that puts that belief in front of a customer: a landing page with a price and a button, a video, a pre-order, or the service delivered by hand to ten people. Every manual delivery is a customer conversation in disguise. The embarrassment of shipping something so small is the price of learning in a week what a full build would have taught you in a year.
- 3
Check that you have a secret, not a category
Peter Thiel's test for an idea is a question: what important truth do very few people agree with you on? An idea most people in your industry would nod at is a category, and a category is where every competitor already is. A real idea rests on a secret, something true about how customers behave that others have not seen because looking was uncomfortable. Write your secret as one sentence about customer behaviour, then check it against the past-tense facts from your conversations. If the sentence survives and is still unpopular, you have something to build; if everyone agrees with it, you have a fight.
- 4
Test how you will reach customers at the same time
Weinberg and Mares found that almost every failed startup had a product and lacked a channel, and their fifty percent rule says to spend half your time on traction from the first day, not after launch. Testing an idea includes testing whether its customers can be reached: list all nineteen channels, choose three, and run a small test on each measured in customers and cost per customer rather than in attention. An idea whose customers you cannot find for a sensible price is not yet an idea; it is a product looking for a business.
- 5
Launch the half that works and let the market answer
Fried and Hansson's rule is to build half a product, not a half-assed product, and to ship it the moment the core delivers value. The temptation while testing is to keep adding features so the test will be fair; the effect is that the test never happens. Cut everything a customer has not described needing in the past tense, launch what remains to a handful of real users, and let the market write the roadmap. A week of real use answers questions that a month of planning cannot, and the features customers ask for after using the half are the ones worth building.
Try this today — free
Today, write your idea's riskiest assumption as one sentence, then message three people who have the problem and ask each what they did the last time it happened and what it cost them. Do not describe your idea.
Questions people ask
- How do I know if my startup idea is good?
- You do not know from thinking about it; you find out from customers. A good idea rests on a real problem people have already tried to solve, a truth about them that most competitors have missed, and a first market small enough to own. The test is whether people who have the problem will commit time or money before the product exists.
- How many customer interviews do I need before building?
- Enough that the answers start repeating, which is usually somewhere between fifteen and thirty conversations run by the Mom Test rules. The number matters less than the questions: ask about specific past events, not opinions about your idea, and count commitments rather than compliments.
- What is a minimum viable product, really?
- The smallest experiment that tests your riskiest assumption with real customers. It can be a landing page with a price, a video, a pre-order form or the service delivered by hand. It is defined by what it teaches, not by how much product it contains, and it is usually much smaller than founders expect.
- Should I keep my idea secret while testing it?
- You can test an idea without pitching it at all: the Mom Test conversations are about the customer's life, not your product, and reveal nothing. Once you have found a real secret about how customers behave, tell it to the people you need to build the company and nobody else. Ideas are rarely stolen; they are usually never tested.
- What if people say they like the idea but nobody pays?
- Then you have learned the most important fact: the problem is not painful enough, or the price is wrong, or you are talking to the wrong people. Go back to the conversations, ask what they actually did and paid the last time, and look for the smaller, sharper problem inside the one you described.