How to Check if a Startup Is Legit Before You Buy In

Before you commit your data, your team's time, or your budget to an early product, you want to know one thing: is this real? The landing page cannot answer that - it is a claim machine, designed and copywritten, identical in quality whether the startup has ten paying customers or none. The signals that do answer it all live outside the startup's own website. Here is where to look.

The rule that cuts through everything

The strongest question you can ask of any stated number - users, customers, revenue - is: how is this known? A number with no provenance is a claim. A number with provenance is a fact, and the provenance is the part worth reading.

The cleanest formulation of this discipline is the evidence-tier idea: every number wears its source - the founder's own report (shown, but labelled as their word), what an independent machine observed (a payment that arrived signature-verified, a join that was email-confirmed), or what members themselves say when asked (R-22 on this board; the same logic works anywhere). A startup that attests to big numbers is fine. A startup that hides which numbers are attested is telling you something.

The five checks, cheapest first

1. Does the team exist outside the product? Named founders, a domain with history, an account that posted before the launch week. Not fame -just existence. A startup whose entire footprint began the day of its launch page is either brand new (fine, if honest) or fictional (not fine).

2. Do the claims survive being specific? "Trusted by hundreds of teams" dies under one follow-up: "which teams, for what?" You don't need the answer to be a logo wall - you need the founder not to flinch. Public boards make this structural: to graduate, a startup on this board needs one accepted claim from a customer organization independent of the founder, on published terms, with a confirmed email and company address behind it (R-11) - the bar for "real customer" is checkable, not a vibe.

3. Is there history the startup can't edit? Landing pages rewrite silently. Audit trails don't. A board that logs every state change with its rule - when terms were published, when the clock started, what happened and why (the public trail is readable by anyone) - gives you a history that began before you arrived and can't be rearranged for your visit.

4. How are early users actually treated? The best predictor of how you'll be treated is how the previous person was. On boards with public claims, the answer window is on a clock (R-10): every claim gets a response in five days, in public. A sales process can promise this. A public record can prove it.

5. Do the incentives point at you? The final check is structural: does the startup's position depend on how it treats you? A seat held on tenure, won in public, kept by serving customers - the founder's standing is literally downstream of your experience (see how the seat machine works). That is rare, and worth preferring when you find it.

The red-flag glossary

Not every gap is a scam - early companies are made of gaps. But three patterns are close to disqualifying on their own: numbers presented without their source (the "how is this known" question refused), terms that can't be quoted (no written scope/duration/cost - see the buyer's checklist), and no reverse gear (no way to leave, export, or delete). A startup can be tiny, unfinished, and honest. It cannot be unspecific, unquotable, and un-leavable, and also honest.

Where to run these checks cheaply

You can run all five in under ten minutes on the B2B board - the startup's seat, terms, claims window, and proof ladder with every number's tier are all public on one page. For the classier version of the same idea, see what counts as a verified pilot customer; for what a real offer should contain, the buyer's checklist; and if you're hunting rather than vetting, start with finding startups to pilot with.