How to validate a startup idea in five days

The trap most founders fall into#

Most "validation" is procrastination dressed up as research.

You read a Lenny's Newsletter article. You make a Notion. You add a column for competitors. You spend three weeks polishing a Figma. You ship nothing, talk to nobody, and convince yourself you're being rigorous.

Real validation is uncomfortable on day one and clarifying by day five. Here is the version of it that actually works — built around a simple constraint that the 2026 startup validation literature converges on: you do not learn anything new by reading.

Day 1 — Write the one-sentence promise#

Write the smallest sentence that fully describes what you want to ship:

For [user] who [pain], we provide [solution] that [unique benefit].

Examples:

• For solo SaaS founders who lose hours to invoicing, we provide an AI bookkeeper that closes their books in under an hour.
• For boutique law firms that draft VC term sheets, we provide a clause library that auto-flags non-standard terms.

If you can't write the sentence in one sitting, the problem isn't tight enough. Sharpen it. Pick a smaller user, a more specific pain, a narrower solution. You can always widen later. Most founders go the other way and never come back.

Day 2 — Find ten of your user#

Not a hundred. Ten.

Why ten? Because the 2026 customer-discovery consensus puts the pattern-emergence threshold at 8-10 interviews, with diminishing returns after about 15. Ten is the smallest number that gives you statistical signal without becoming a sub-project of its own. The Founder Institute curriculum uses the same range. You can scale up later if the signal is ambiguous; you cannot scale down once you've burned three weeks on outreach.

Where ten of your user lives varies by niche. Some places that actually work:

• Reddit niche subs. Search for posts complaining about your pain — those are your users.
• LinkedIn. Filter by job title and post a question. Reach out to the people who reply substantively.
• IndieHackers / Hacker News. If you're building for makers, they're already there.
• Communities. Slack and Discord servers for the role. Most have a #intros channel.

Ten people you can actually message. Don't move on until you have all ten.

Day 3 — Run five problem interviews#

Half an hour each. Voice or video. The script:

1. Tell me about the last time you ran into [the pain].
2. What did you do about it? How long did it take?
3. Have you tried any tools to fix it? Why did you stop using them?
4. If a tool solved this, what would have to be true for you to pay for it?

That's the whole script. Do not pitch your solution. Ask, listen, take notes verbatim.

After five interviews you'll know one of three things:

• The pain is real and frequent → keep going.
• The pain is real but rare → kill the idea, look for an adjacent one.
• The pain isn't real to your target user → you picked the wrong user. Re-segment.

About 40% of would-be founders give up here. That's the system working. The remaining five interviews you keep banked for day 5; they become your launch list.

Day 4 — Build the smallest possible offer#

Not a product. Not even an MVP. An offer.

The cheapest offer that proves intent:

• A landing page with a price and a "Buy now" button that takes pre-orders.
• A Calendly link that books a 30-minute paid consultation.
• A Google Form that collects email + a $50 deposit refund-on-demand.

The bar isn't "is it polished" — the bar is "would I be embarrassed to put this in front of a stranger?". If yes, ship it.

This step matches the validation pattern advocated by LeanFoundry's 90-day framework and IdeaFloat's 7-day framework: both compress to "test demand with a real offer before you build". A landing page with a price is the modern version of the smoke test that has worked since the lean-startup era.

Day 5 — Send it to your ten#

Direct messages, not a megaphone.

Hi [name]. I remember you mentioned [specific pain] last [day]. I built [tiny thing] that addresses it. The price is [X]. Worth a look?

Three outcomes:

Five days. Maybe ten if you're being honest with yourself. Cheaper than three weeks of Figma every time.

What we're building to make this easier#

The day-1 sentence and the day-3 interview script can be done with a notebook. The day-2 outreach and the day-5 follow-ups are exactly the moments where founders get stuck.

That's where speedrunfounder's idea lab and validation lab come in. The AI co-pilot suggests your ten users, drafts the interview questions, and tracks responses against your hypothesis so you don't fool yourself with selective memory.

But — this matters — none of that replaces day 3. You still have to talk to humans. The tool just makes sure you've talked to enough of them, and that you're not lying to yourself about what they said.

That's the whole game.

— Artur