The 10-Point Startup Idea Validation Checklist

An idea is validated when ten specific things are true. Until then, you're guessing.

Why a checklist beats vibes

Most founders validate by feel. They talk to a few friendly people, hear "that sounds cool," and read it as permission to build. Six months later they have working software and no buyers. The problem isn't effort — it's that enthusiasm and validation feel identical from the inside.

A checklist breaks the loop that fools you. Each item below is binary: you either pass it or you don't. There's no partial credit for "kind of." Run your idea through all ten before you write a line of code — the ones you fail tell you exactly what to test next.

The 10-point checklist

1. Problem stated in one sentence

Write the problem your customer has — not the solution you want to build. No jargon, no product language, no "platform that leverages." If you can't say it in one plain sentence, you don't understand it yet, and neither will anyone you pitch.

Passing looks like: "Solo lawyers waste hours every week chasing clients for documents."
Failing looks like: "An AI-powered workflow automation layer for legal operations." (That's a solution wearing a problem's clothes.)

2. Specific target segment

"Small businesses" is not a segment — it's a continent. A real segment is narrow enough that you can name where these people gather, what tools they already use, and roughly how many exist. Specificity is what makes the other nine checks possible.

Passing looks like: "Solo dentists in cities under 200k people."
Failing looks like: "Healthcare professionals" or "anyone who hates admin work."

3. You've spoken to 20+ people in that segment

Surveys measure what people will click; conversations reveal what they actually do. Twenty is enough to hear patterns repeat and stop being surprised. Talk to them — on a call, in person, voice-to-voice — not through a form.

Passing looks like: Twenty recorded or noted conversations where you mostly listened.
Failing looks like: "I got 200 survey responses." Surveys are a complement, never a substitute. See our guide to market validation for how to run these conversations well.

4. They describe the problem unprompted

The strongest signal in any interview is the customer raising the pain before you mention it. If you have to remind people the problem exists, it isn't urgent enough to fund a company. Watch for the moment they lean in and tell you how bad it is.

Passing looks like: "Don't even get me started on how long invoicing takes" — said before you brought it up.
Failing looks like: "Yeah, I guess that could be annoying" — only after you described the problem in detail.

5. They've tried existing solutions

Real pain leaves a trail of workarounds. When people have already cobbled together spreadsheets, hired help, or paid for a clunky tool, you know the problem is worth money to them. A pristine, solution-free space usually means nobody cares enough to try.

Passing looks like: Customers list two or three things they've already tried and why each fell short.
Failing looks like: "No, I've never really looked for anything to fix that." Mapping these alternatives is also the start of real competitor analysis.

6. They can quantify the cost

Pain you can't measure is pain people will live with. Ask what the problem costs in hours, dollars, or deals lost, and listen for a number rather than an adjective. The size of that number sets the ceiling on what you can charge.

Passing looks like: "It costs me about six hours and two lost deals a month."
Failing looks like: "It's just really frustrating." Frustration isn't a budget line.

7. You've identified the buyer

The person with the problem is frequently not the person with the budget. In companies, an end user can love your product while procurement quietly kills it. You need to know who signs off, what they care about, and whether they even feel the pain.

Passing looks like: "The office manager feels the pain; the practice owner writes the check, and she cares about staff retention."
Failing looks like: "Everyone in the company would want this." Everyone means no one owns the decision.

8. A landing page converts cold traffic

Warm intros from your network don't count — they're being polite. Put a clear page in front of strangers who've never met you and see whether they raise their hand: sign-ups, demo requests, or a waitlist. Cold conversion is the first proof that demand exists beyond your friends.

Passing looks like: A meaningful share of cold visitors take the action you asked for.
Failing looks like: Heavy traffic and a flat conversion line. A sharp product vision document helps you write a page that actually lands the promise.

What "meaningful" actually means (rule of thumb, not a law): for a low-friction email signup or waitlist from genuinely cold paid traffic, founders often treat 2–5% as a live signal and below ~1% as a flat line. For a "request a demo" or "talk to sales" action the bar is much lower — even 0.5–1% can be strong, because the intent is heavier. These numbers swing wildly with your price and model: a $50k ACV enterprise tool converting cold visitors at 0.3% to a booked call may be doing better than a $9/mo app converting at 4%. Don't import someone else's benchmark — judge the trend (does a sharper page move the number?) more than the absolute.

9. At least 3 people pre-paid

Cash beats every other signal because it's the only one that costs the customer something. A pre-order, a deposit, a signed letter of intent with money attached — these separate real intent from polite interest. Three is a small number on purpose; if you can't find three, the demand isn't there yet.

Passing looks like: Three customers have put money down before the product exists.
Failing looks like: "Forty people said they'd definitely buy." Free yeses are worth roughly nothing.

Why three, and when three is the wrong floor: three is deliberately small — it's the minimum that proves someone who isn't your friend will part with money for a thing that doesn't exist yet. It scales with deal size. For a $20/mo consumer app, three pre-payments tells you almost nothing about a market that needs thousands of buyers; you'd want ten-plus, or a paid waitlist with real conversion. For a $40k/year B2B contract, one signed order with a deposit is a louder signal than fifty consumer pre-orders — the buyer did procurement, looped in stakeholders, and committed budget. Treat "3 pre-paid" as the consumer/SMB rule of thumb and let ACV move it: higher price, fewer pre-sales needed; lower price, many more.

10. You'd bet your next 12 months on it

The final check is honest, not analytical. With everything you now know, would you commit the next year of your life and savings to this? If the answer is a flinch, treat it as data: something in the first nine checks is weaker than you're admitting.

Passing looks like: A clear, unhesitating yes you'd repeat to a skeptical partner.
Failing looks like: "Probably, if a few things work out." That's a hypothesis, not a commitment.

How to use this checklist

Go in order. The early items are cheap and shape everything after them — a fuzzy problem statement (1) makes a clean landing page (8) impossible. Treat each check as a gate: don't spend money on traffic until items 1 through 7 pass.

Run the list as a living scorecard, not a one-time exam. Re-score it every couple of weeks while you're in discovery, and write the evidence beside each pass so a co-founder or investor can challenge it. A structuring tool like the God of Startups workspace can keep these checks, interviews, and assumptions linked in one place, but a shared doc works fine to start. The discipline matters more than the format.

When you fail an item, that's the win — it's the next experiment, named precisely. A failed item 5 sends you back to interviews; a failed item 9 means your pricing or segment is off. Validation isn't passing on the first try; it's knowing exactly where you stand.

Where this checklist legitimately breaks

A checklist that false-negatives your most interesting ideas is worse than no checklist. Two business models break the list for good reasons — know them before you kill a strong idea on a technicality.

Two-sided marketplaces need both sides validated, separately. A marketplace has no value until both supply and demand show up, so a single segment (item 2) and a single buyer (item 7) don't describe it. Run the checklist twice — once for each side — and add a chicken-and-egg gate the standard list ignores: which side can you concentrate first, and does the other side show up once you do? Three pre-paid restaurants mean nothing if no diners book; ten eager diners mean nothing with no restaurants listed. Validate that you can solve cold-start in one narrow geography or vertical before you trust any single-side signal.

Genuinely novel categories fail item 5 — and that's the point. Item 5 ("they've tried existing solutions") assumes the problem already has workarounds. For a true new category, customers won't have tried anything, because the problem was invisible or accepted as unfixable. A pristine, solution-free space is usually a death signal, but occasionally it's a real wedge. The distinction: in a dead space, people don't even recognize the problem when you name it (fails item 4 too). In a real new category, the pain is acute and unprompted (item 4 passes hard) but unaddressed because no one believed it was solvable. If item 5 fails while item 4 screams, don't kill the idea — replace item 5 with: "Have they hacked together a workaround, even a bad one, or are they suffering in silence because they think nothing can help?" The latter is rare and valuable; the former is the normal pass.

In short: don't let the checklist false-negative a marketplace because one side is thin, or a novel category because no incumbent exists yet. Adapt the failing item, don't ignore the discipline.

The false-positive trap: how each check gets gamed

The more dangerous failure isn't a false negative — it's fooling yourself into a pass. Every check has a way to game it, usually without realizing you're doing it. Audit your own evidence against these.

Leading the interview (items 3–6). You describe the problem in loving detail, the customer nods, and you log a pass. You didn't validate demand — you validated that people are polite. Fix: shut up first. If you mentioned the pain before they did, item 4 did not pass.
Cherry-picking the segment (item 2). You quietly redefine "my segment" as "the five people who loved it." A segment defined by who agreed with you isn't a segment — it's a survivorship bias. Fix: define the segment before the conversations, by an external attribute (role, company size, industry), not by enthusiasm.
Refundable deposits counted as "pre-paid" (item 9). A fully-refundable $1 deposit, a "reserve your spot for free," or a card on file with no charge are not money down — they cost the customer nothing, so they prove nothing more than a free yes. Fix: the payment must be non-trivial and at-risk. If you'd refund it without a fight, it doesn't count.
Vanity cold traffic (item 8). Sending warm audiences (your Twitter, a friendly newsletter) and calling it "cold," or counting low-intent clicks as conversions. Fix: cold means strangers who've never heard of you, and the conversion must be the real action, not a bounce.
Quantifying the cost for them (item 6). You compute "this costs you six hours a week" and they shrug agreement. A number you supplied isn't their number. Fix: the figure has to come out of their mouth, unprompted.

A clean ten-for-ten that you can't defend line-by-line to a skeptic is the most expensive false positive in startups — it buys you six months of building before reality corrects you. Write the evidence beside each pass, not just the checkmark, and have someone who wants you to fail try to knock each one down.

FAQ

How long should idea validation take?

There's no fixed number, and rushing defeats the point. Most of the time goes into items 3 through 6 — the conversations — because patterns only emerge after enough of them. Move fast on the cheap checks and let the expensive ones (pre-payment, cold traffic) take the time they need.

Do I need to pass all ten before building?

You need to pass the first seven before you build anything serious. Items 8 and 9 can run in parallel with a thin prototype, and a pre-sale often requires something to show. Item 10 is your gut reading the other nine — never override it with optimism.

What if I can't reach 20 people in my segment?

That's a finding, not a roadblock. A segment you can't reach is a segment you can't sell to, so either your targeting is too narrow or you haven't found where these people actually are. Either way, fix it before building — distribution problems don't disappear once you have a product.

Is pre-payment really necessary for every business?

For most early-stage products, yes — money is the cleanest signal you'll get. The form can flex: B2B might use a signed LOI, consumer might use a paid waitlist. If your model genuinely can't charge before launch, replace item 9 with the strongest costly commitment available, and weight it just as heavily. Before you commit, it's also worth scoping the smallest version you can build to test demand.

If you fail any of these, you have a hypothesis — not a validated idea.