How to Evaluate a Startup: Problem, Team, Market, Cash, and Terms

A good startup pitch makes the future feel obvious. Due diligence asks what evidence would have to be true for that future to happen.

This is an educational checklist, not personalized investment advice. Early-stage investments can be illiquid and can lose all of their value; Investor.gov explicitly warns investors to be prepared for that possibility.

1. Is the problem expensive enough to matter?

Ask who experiences the problem, how often it occurs, what it costs them, and what they do today. A clever product is not the same as a painful problem. Evidence can include paid pilots, repeated usage, signed contracts, customer interviews, or a measurable replacement of an existing spend.

2. Is the team learning faster than the problem changes?

Credentials can be useful, but execution evidence is more useful. Look for the team's ability to ship, talk to customers, correct bad assumptions, recruit missing skills, and explain hard tradeoffs without hiding them.

Do not invent a heroic founder narrative. Ask what the founders actually did and what changed because of it.

3. Is the market defined from the buyer upward?

Large industry forecasts often make a pitch look safer than it is. Start with the actual reachable buyer. How many such customers exist in the first practical segment? What do they spend on the current alternative? How long does a sale take? What prevents a larger competitor from bundling the same feature?

4. Does growth improve or worsen the economics?

Revenue growth can hide poor unit economics. Review gross margin, customer concentration, acquisition cost, support burden, infrastructure cost, burn rate, runway, and the amount of future capital the plan assumes.

For AI products, include inference and data costs rather than treating model usage as free.

5. What exactly are you buying?

Equity, SAFEs, convertible instruments, and crowdfunding securities can produce very different outcomes. Read valuation/cap mechanics, dilution, liquidation preferences, voting or information rights, transfer restrictions, platform fees, and the position of other investors.

A strong company can still be a poor investment at unfavorable terms.

6. What are the three most credible failure modes?

Write them down before the upside case. Examples might include a customer-acquisition channel disappearing, a regulated approval taking longer than expected, a key supplier becoming unavailable, gross margin never improving, or the product becoming a feature in a larger platform.

Then ask what evidence would tell you that each failure mode is becoming more likely.

7. What would change your mind?

If no negative evidence can change the conclusion, you are not doing due diligence. Define disconfirming evidence before committing: churn above a threshold, customer concentration that is too high, insufficient runway, weak security terms, or repeated inability to convert pilots into paid contracts.

A one-page startup review

Area	Question
Problem	Who pays to make this problem go away?
Product	What evidence shows the product works in a real workflow?
Team	What have the founders learned and shipped?
Market	What first segment can actually be reached?
Economics	Does serving the next customer improve the business?
Cash	How long until the company needs more financing?
Terms	What security and rights are being offered?
Downside	How can this fail, and can the investor tolerate that outcome?

The useful discipline is simple: separate what the pitch says from what the evidence supports.