How to Validate a Startup Idea in One Weekend (Before Writing Any Code)

You can get a kill-or-advance verdict on a startup idea in one weekend, for under $100, without writing a line of code. The method: Friday night you define exactly who has the problem, Saturday you have 5 honest conversations with those people, Sunday you run a fake-door test on strangers, and Sunday night you make the call with a decision matrix.

To set expectations honestly: a weekend cannot prove an idea will work. Nothing can. What it can do, reliably, is kill bad ideas early and promote promising ones to a real 30-day effort. Killing a doomed idea on Sunday instead of month four is worth more than most accelerators.

Here is the playbook, hour by hour.

Friday night: how do you define the persona and problem?

Write two sentences before you do anything else, because every Saturday and Sunday activity depends on them:

1. The persona sentence. Name a specific human: role, context, and the situation that triggers the problem. "Operations manager at a 20-to-80-person HVAC company who schedules 6 field techs by group text" is a persona. "Busy professionals" is a horoscope.
2. The problem sentence. State what they are trying to do and what it currently costs them, in time or money. "Rescheduling one sick tech's day takes her 45 minutes of texting and two angry customers."

If you cannot fill in both sentences with that level of specificity, you are not ready for Saturday, and that is a useful discovery on its own. Vague persona plus vague problem is the number one reason validation weekends produce mush: you end up interviewing random people about a feeling.

Worth noting: a well-formed startup brief already has this shape. This is exactly why Ignition structures every daily idea the same way, with a named persona, the exact problem stated in one line, and the competitive landscape mapped. If your idea came from an Ignition brief, Friday night is mostly done: the brief is the input format the weekend expects, and your job is confirming it against live humans. If your idea came from your own head, spend the full Friday evening getting it into that format.

Last Friday task, 20 minutes: list the three closest existing solutions, including "spreadsheet" and "do nothing." You need these Sunday, and knowing them keeps you honest Saturday.

Saturday: where do you find 10 real people to talk to?

You need 10 candidates to get 5 conversations, and where you find them depends entirely on the persona. Use whichever row matches:

Persona type	Where to find them	The move
B2B operator (ops, finance, agency)	LinkedIn search by title, your second-degree network	DM asking for 15 minutes of advice on how they handle X today. Offer a $25 gift card
Tradesperson or SMB owner	Local Facebook groups, supplier counters, just calling businesses	Call at off-peak hours, say you are researching how shops handle X, ask for 10 minutes
Prosumer or hobbyist	Subreddits, Discord servers, niche forums	Post asking how people currently handle X. Interview the commenters
Developer	Twitter/X, Hacker News, relevant GitHub issues	Comment threads about the pain point, then DM
Consumer (broad)	Your own extended network, community groups	Screen hard for people who actually have the trigger situation

Two rules. First, only talk to people who match the persona sentence; five perfect-fit conversations beat twenty sloppy ones. Second, pay attention to how hard recruiting is. If you cannot find 10 people who have the problem in one Saturday morning, finding 10,000 customers later will not be easier.

Saturday: what questions should you ask (without leading the witness)?

Ask about their past behavior and never mention your idea. That is the core of Rob Fitzpatrick's The Mom Test, and it exists because people are kind: describe your idea and they will compliment it, and compliments are worthless data. Past behavior cannot flatter you.

The five questions, roughly in order:

1. "Walk me through the last time you dealt with X. What did you actually do?"
2. "What have you already tried to fix this? What did that cost you?"
3. "What happens if you just do nothing about it?"
4. "Where does this rank on your list of problems this month?"
5. "Who else has this worse than you?"

Question 2 is the money question. If they have never tried anything, they may complain, and they do not care; the problem is a papercut. If they duct-taped a spreadsheet, hired a VA, or pay for a tool they hate, you are looking at real demand with a price tag already attached.

Take notes on their exact words. "I lose half a day every Friday to this" is a quote you will reuse in your landing page copy tomorrow and your cold outreach for months.

Five honest conversations and a $50 fake-door test will teach you more about your idea than 50 hours of building it.

Sunday: how do you run a fake-door test?

A fake-door test shows strangers a product that does not exist yet and measures whether they act. Conversations test whether the problem is real; the fake door tests whether the promise pulls.

The build, 2 to 3 hours:

1. Landing page. One page on Carrd, Framer, or plain HTML. Headline states the outcome in the persona's words (steal a Saturday quote). Three bullets on what it does, one screenshot mockup if you can manage it, and a single call to action: "Get early access" with an email field. Optionally add a pricing hint ("early users lock in $29/month") because willingness to see a price is itself signal.
2. Traffic. Two options. Spend $50 to $75 on narrowly-targeted ads (Meta or Google, targeting the persona's job titles or search terms), aiming for 100 to 200 visitors. Or, for niche B2B personas ads cannot target, post where they gather: the subreddit, the Facebook group, the Discord from Saturday. Frame it honestly: "I'm building this because of X, would you use it?"
3. Measure. Visitors, email signups, and any unprompted replies or DMs.

Benchmarks, as rough rules of thumb: under 2% of visitors leaving an email is weak, 5% or better from cold traffic is genuinely promising, and anyone who emails asking "when can I have this" counts triple.

Sunday night: how do you decide to kill or commit?

Score the weekend against this matrix. Be mechanical about it; you are tired and attached, which is exactly when founders talk themselves into things.

Signal	Kill	Weak (park it)	Commit to 30 days
Problem mentioned unprompted in conversations	0 or 1 of 5	2 of 5	3+ of 5
People already tried or paid for a workaround	None	1 person	2+ people
Fake-door email conversion	Under 2%	2 to 5%	Over 5%
Inbound follow-up ("can I try it?")	None	1	2+
Your own energy after 5 conversations	Drained	Neutral	More curious than Friday

Three or more cells in the Commit column: give it 30 days and start building the smallest thing that delivers the outcome. Mostly Weak: park it with a written note on what would change your mind, and go score your next candidate (the ranking system in how to pick one idea exists for exactly this loop). Mostly Kill: kill it Sunday night, feel the sting for ten minutes, and be glad it cost a weekend instead of a year.

The weekend checklist

• Friday, 2 hours: persona sentence, problem sentence, three existing alternatives.
• Saturday morning: recruit 10 people from the matching channel.
• Saturday afternoon: 5 conversations, five questions, exact-words notes.
• Sunday morning: landing page live, $50 to $75 of traffic or one honest community post.
• Sunday night: fill in the matrix, make the call, tell someone.

One more thing. If the idea dies, the weekend was not wasted: you now have a validated recruiting channel, interview reps, and usually one better idea that surfaced in conversation three. Ideas are cheap and evidence compounds. And if you find yourself with no idea worth a weekend, fix the top of the funnel first with a real idea sourcing system.