Two weeks sounds fast. What am I giving up?

Precision you couldn't use anyway. Two weeks is enough to learn whether strangers act. That's the decision you're really making. It isn't enough to size a market to the nearest million. It won't tell you what to charge either. Those questions matter later. They get cheaper once you have customers instead of guesses.

How many conversations before I can trust the pattern?

Stop when you can predict the answer before they give it. That tends to land at ten to fifteen people of the same type. Still surprised at twenty? Then your segment is too broad. Your sample isn't too small. A bigger sample won't fix a problem with where you drew the line.

What if nobody replies to anything?

That is a result, and it arrives faster than any other. Silence from a channel you chose means one of two things. Either the channel is wrong or the promise is. Change one of them, name which one, and re-run the same week. Change both at once and you learn nothing from the second try either.

Can I run this while working a full-time job?

Yes, and the calendar hardly moves. The work is 10–20 hours. But most of it is waiting. It spreads across evenings, and the two weeks still hold. Book the conversations first. They're the part that needs other people's diaries.

When does validation legitimately take longer?

When it takes weeks just to reach one buyer. Think of industries with a lot of rules. Or buying inside a big company. Or doctors. Or anything at all with a gatekeeper in the way. Someone stands between you and the person with the problem. The tests don't change and the pass marks don't change. Only the calendar stretches. It stretches because of access, not because you need to be more sure.

How long does it take to validate a startup idea?

Where the two weeks actually go

Almost none of it is work. Other people’s response times set the schedule. That’s why you can’t cram it into a weekend. It’s also why it doesn’t need a quarter.

	What you do	Hours	What you’re waiting on
Days 1–2	Write the pass marks. Name one buyer and three places you can reach them.	3	Nothing. Do this before anything else.
Days 2–5	Send the asks. Read what your buyers already posted in public about the problem.	4	Replies. Expect a third of them.
Days 4–9	Have the conversations. Ask about the last time it happened, never about what they’d do.	5	Their calendars.
Days 6–8	Put up one page that makes one promise and asks for one commitment.	3	Nothing.
Days 8–14	Send cold traffic at it. Count.	2	Enough visitors for the number to mean something.

Add it up and it’s a long weekend of effort spread across fourteen days. The two weeks are the delivery time, not the labour.

The pass mark is what makes it two weeks

Say you haven’t written down what counts as success. Then you can go on forever. Every result invites one more test. Eight of twelve said the problem was real — is that good? Without a number written down first, it’s whatever you need it to be. You decide that when you read the result.

That’s the real reason a timeline blows out. Not because you’re being thorough. Because nothing was pinned down. Decide in advance what the numbers mean. Seven of ten people describe the problem with no prompt? Go. Five? Stop. Then it ends on the day of the tenth conversation. We wrote the full version of this, gate by gate. It’s in the four-gate framework.

Month three is not validation

If your plan runs long, look at what’s filling the extra weeks. In our experience it’s one of four things, and none of them are tests:

Building a demo “so people have something to react to.”
Fixing the page again instead of sending more people to it.
Waiting for a bigger sample. The pattern was already clear at twelve.
Reading up on the market rather than talking to it.

Every one of those feels like good work. The tell is that none of them can fail. A test you can’t fail isn’t giving you information. It’s giving you the feeling of progress. That feeling is what four-month builds are made of.

When two weeks is genuinely wrong

Access is the honest exception. Say your buyer is a committee that does the buying at a hospital. Or a bank that the regulators watch. You will not be having ten conversations in five days. Pretend otherwise and you end up with the wrong people. You talk to the ones who answer, not to the ones who matter.

The fix is to be patient about the calendar. And hard about the bar. Same gates, same pass marks, longer wait. What you must not do is lower that bar. Not because the right people are slow. Swap real buyers for the people you can reach. Now you’ve validated the wrong market.

Start today, decide on the 14th

The best hour you will spend is the first one. It needs no research at all. Open a document. Write down the four things that would have to be true. Then write the number next to each one. Everything after that is scheduling.

Don’t want to spend the first days hunting for where your buyers talk? That’s the part we automated. Doing it by hand is covered in our guide to validating a startup idea. And how much it costs covers the other half of this question.