How to validate a SaaS idea in 7 days

Write one sentence in this exact shape: [specific person] currently does [specific thing] to solve [specific problem], and it costs them [time or money].

Bad: "A tool that helps freelancers manage their finances." Good: "UK freelance designers currently reconcile invoices in a spreadsheet each quarter, and it costs them a full day plus £400 in accountant corrections."

The good version can be wrong. That's what makes it worth testing. If your sentence can't be false, you have a category, not a hypothesis.

Then write your kill criteria today, before you have any emotional investment in the outcome. Something like: "If by day 7 nobody has paid or signed a dated commitment, I stop." Write it in a file, with the date. You will not be a neutral judge on day 7 — this is you, now, making the call for the version of you who won't be able to.

Finally, run the free validator below on your sentence. Twenty seconds to catch the failure modes that are obvious from the outside: already solved, unmonetisable, or incoherent. It won't tell you the idea is good. It will sometimes tell you it's dead, which saves you six days.

You need twenty of them reachable this week. Not a market — twenty individuals you could message today.

Look in this order, because it's the order of how much signal each gives you:

1. Communities with a search function. Subreddits, Discord servers, Slack groups, forums. Search your problem phrase and read complaints, not questions. A complaint is a person who has already paid an emotional cost.
2. Review sites for the tool they use now. One- and two-star reviews of the incumbent are a list of people actively unhappy with the status quo, with their reason attached.
3. Job boards. If companies are hiring someone to do this manually, it's an expensive problem with a budget line already attached.
4. Your own last two years. Who have you personally watched struggle with this?

Write twenty names in a spreadsheet with where you found them and the exact words they used about the problem. Those words are your future landing page copy. Do not paraphrase them into marketing language — the paraphrase is where the signal dies.

If you cannot find twenty in a day, that is data. Either the problem is rarer than you think, or the people who have it don't gather anywhere, which makes them very expensive to reach later. Both are worth knowing on day 2 rather than month 9.

Book five conversations. Fifteen minutes each. Your only job is to not pitch.

Ask about the last time it happened, in the past tense:

• "Walk me through the last time you dealt with this."
• "What did you do about it?"
• "What did that cost you — hours, money, or something you didn't get to do?"
• "Have you tried anything to fix it? What happened?"
• "What would have to be true for you to change how you do this?"

Never ask "would you use a tool that…". People are generous and imprecise about hypotheticals, and a yes costs them nothing. Every yes you collect this way is a false positive you'll pay for later.

The signals that matter: they've already tried to solve it (a spreadsheet, a VA, a half-built script), they can name what it costs in a unit, and they get animated. The signal that ends it: polite interest with no history of trying. Polite interest is the single most expensive thing you can mistake for demand.

After five conversations, rewrite your day-1 sentence in their words. It will have changed. That change is the most valuable thing you'll produce this week.

This is the day most founders skip, and skipping it is why validation weeks so often produce nothing.

Go back to the two or three most animated people and make a concrete offer: "I'm building this. It'll be £X/month. I'll have something working in three weeks. Would you pay for the first three months now at half price?"

Then stop talking. The silence is where the answer lives.

Three possible outcomes, and only one of them is good:

• They pay. Even £50. Real money from a stranger is the only signal that isn't self-reported. Take it, note the date, and refund it without argument if you don't ship.
• They sign a dated LOI. Second best: "I commit to trialling this at £X when it ships, by [date]." A signature with a date is a real commitment; enthusiasm is not.
• They go vague. "Definitely interested, let me know when it's ready." That is a no delivered politely. Record it as a no.

If nobody in your five will pay or sign, you don't have a business yet. You may still have a good idea — but the wedge, the price, or the person is wrong, and building would just be an expensive way to find that out.

One page. Four hours, not two days. The point is a test, not a website.

Structure:

• Headline: their problem in their words, from day 3. Not your product category.
• Subhead: the outcome after using it, in one sentence.
• Three bullets: what it does. Concrete. No adjectives you'd find on a competitor's page.
• Price: a real number. "Contact us" tests nothing.
• One CTA: an email field, or a real payment link if you can ship in weeks.
• Who it's not for: one line. It costs you nothing and makes the right person feel found.

Don't fake a product with a screenshot of something that doesn't exist. Say plainly that it's early and you're taking first customers. Honest early pages convert better than polished vague ones, because the honesty is itself a signal about who you'll be to work with.

Run the free landing page roast on it before you send traffic anywhere. The most common finding is that the headline describes the product instead of the outcome — twenty minutes to fix, and it's the largest single lever on the page.

Friends clicking your link is not a test. You need people with no relationship to you.

Cheapest options that produce a real read:

• £50–100 of ads targeted at the job title or interest from day 2. Enough for 200–500 clicks in most niches. You're buying a conversion rate, not customers.
• Post where you found them — but participate honestly, as someone who has the problem, not as someone with a link. One good comment in the right thread beats ten posts.
• Message the other fifteen from your spreadsheet individually, referencing the specific thing they said. Not a template. Fifteen personal messages take an hour.

Watch the conversion rate from visit to email. Under 2% on cold traffic means the message isn't landing — usually the headline. 5%+ means something real is happening. Between them, keep watching.

Also watch what people write in the email field's follow-up, if you ask one question. "What are you using now?" in a single optional field is the highest-information question you can ask a stranger, and roughly a third will answer it.

Open the file you wrote on day 1. Read your kill criteria before you look at any results. This order matters — reading the results first lets you renegotiate the criteria without noticing.

Then count the only two numbers that carry weight: people who paid, and people who signed something dated. Emails collected are a weak third; likes and encouragement are zero.

One or more paid. You have a business to build. Ship the smallest thing that delivers what they paid for, within the timeframe you promised.

Signatures but no money. Go back and ask for money specifically. If the signature doesn't survive a payment request, it wasn't a commitment.

Emails only. Your message works and your offer doesn't. Change the price or the wedge — same audience, different ask — and run days 4–6 again. One more week, not one more month.

Nothing. Honour the criteria. You bought a definitive answer for seven days and about £100, which is a genuinely good trade. Write down what you'd test differently, take a few days off, and start the next one with the same discipline.