How to Evaluate a Startup Before Joining: A Due Diligence Guide

A startup offer is also a risk decision. You do not need to predict the next unicorn, but you do need enough evidence to judge whether the company can pay you, whether customers want the product, and whether the job will still make sense when plans change. This guide gives you a research sprint, exact questions, and a scorecard you can use before saying yes.

Decide what you are trying to learn

The goal is not to find a startup with zero risk. That company does not exist. Your goal is to decide whether the risk is visible, understood, and acceptable for your finances and career.

Separate every finding into three buckets. Verified means you saw a filing, product, customer reference, offer document, or consistent evidence from more than one source. Reported means a founder, hiring manager, database, or company webpage said it. Inferred means it is your conclusion. Keep those labels in a simple note, because they stop a polished pitch from turning into a fact in your head.

Use two questions as your filter: “What would have to go right for this role to work out?” and “What would happen to me if it does not?” A seed-stage role may still be worth taking if you will learn from a strong manager, ship work you can show later, and can absorb a job search if the company shuts down. A bigger round alone does not answer those questions.

Run a 45-minute research sprint before the next interview

Give yourself 45 focused minutes before a founder call or before you accept an offer. You are looking for contradictions, not trying to produce an investor memo.

1. Minutes 0–10: understand the product. Read the homepage, pricing page, product documentation, and recent releases. Write one sentence: “This company helps [specific customer] solve [specific problem] by [specific mechanism].” If you cannot write that sentence, ask for a demo instead of pretending you understand the business.
2. Minutes 10–20: build a funding timeline. Search the company name, founders, legal entity name from the offer letter, “funding,” and “raised.” Start with the company’s press releases and an investor’s portfolio page. Record the date, amount, named source, and a link.
3. Minutes 20–30: inspect the team. Read founders’ profiles, company bios, and recent announcements. Look for relevant operating experience, not prestige alone. Then list who would set your priorities, review your work, and unblock you.
4. Minutes 30–40: look for customer evidence. Try the product, read recent app-store or marketplace reviews when they fit the business, search for customer case studies, and look for a customer or partner discussing the product on its own channel.
5. Minutes 40–45: prepare your questions. Turn gaps into short questions. You can borrow additional prompts from our guide to questions to ask in a startup interview, but ask the same core question of more than one interviewer when possible.

Use Crunchbase and PitchBook as starting indexes, not final proof. Crunchbase says its data includes venture-partner submissions, contributor updates, government filings, news, and manual verification. PitchBook says its research begins with public and self-reported information, then adds secondary sources, validation, and direct contacts. A database record can tell you what to verify; it cannot show you the company’s current bank balance or explain a round’s terms.

How to find a startup’s funding history

Build a timeline before you ask about money. A press release may tell you the round date, amount, and investors. An investor’s portfolio page can confirm that the firm names the company. A database entry can surface the lead. Mark a round “confirmed” only when you have a primary source or matching independent evidence.

For a U.S. company that may have raised under Regulation D, open the SEC’s EDGAR Full Text Search. Enter the legal company name first, then try the founder’s name or a prior company name if needed. Choose the “Exempt offerings” category or form type “D,” use a sensible date range, and open the newest filing and any amendments. The search supports company name, ticker, CIK, filing category, form type, date, and location filters.

A Form D is a public notice of an exempt securities offering filed on EDGAR. In a result, read the issuer identity, date of first sale, related people, revenue range if disclosed, total offering amount, amount sold, amount remaining, and investor count. Those are fields on the official Form D. The form gives you a dated view of an offering. It does not give you a current cash balance, a full investor roster, or a forecast.

Treat a Form D as evidence of a reported offering, not a stamp of quality. The SEC says a Form D is a notice and does not represent SEC registration or approval. That distinction matters when an interviewer mentions a filing as if it settles the company’s prospects.

For a Canadian company, check the legal name in Corporations Canada’s federal database. It confirms corporations created under federal corporate law, but it does not cover provincial, territorial, or foreign corporations. If the federal search has no result, follow the official provincial or territorial registry path rather than treating the absence as a red flag.

Ask about runway without making it awkward

Runway is the period a company can keep operating under a stated cash and spending plan. Your job is not to demand a spreadsheet. Your job is to learn whether leadership has a specific operating plan and can explain its assumptions.

Ask a founder or future manager after you have shown that you understand the product:

“I’m excited by the role and I want to understand the operating plan behind it. Based on the current headcount and plan, how many months of runway do you have? What assumptions about revenue or the next financing does that depend on? If either moves later than expected, what changes first?”

Then ask, “What are the next two milestones the company needs to hit before it raises again or reaches the next plan?” A useful answer includes a range, a baseline, a milestone, and a contingency. “We are raising soon” is not a runway answer. It may be a reason to ask a follow-up.

Example: hypothetical math. Suppose a company has $3,000,000 available and expects to use $250,000 per month after revenue. That is a starting estimate of 12 months. It is not a promise. Hiring, collections, spending, and fundraising can change the calculation. Use the figure to frame your questions, not to create false certainty.

Some leaders will not share a cash number with a candidate. That is normal. You can still ask about the planning horizon, the next milestone, planned hiring, and what would change if the milestone slips. The warning sign is not confidentiality. It is evasive answers, answers that conflict across interviewers, or pressure to stop asking.

Test traction in the right way for the business

A pre-revenue developer tool and a consumer subscription app should not be judged by the same evidence. Start with the stage and business model, then ask for proof that matches it.

Company-controlled material is still useful, but label it as a claim. An independent customer mention, a functioning product, or a specific answer from two people adds weight. If the company displays a well-known customer logo, look for a case study or the customer’s own announcement before you treat it as active use.

Do not demand a revenue number if the company cannot disclose one. Ask instead which metric the team uses to know the product is working, what changed in the last quarter, and what is blocking growth. A strong team can explain the metric and its limits in plain language. A weak answer substitutes market size or press coverage for customer behavior.

Check the founders, team, and job you would actually have

Look for evidence that the people running the company can make decisions, learn, and give you usable direction. Review their work history, but do not make a famous employer your shortcut for judgment. More useful questions are: What problem do they know closely? What have they changed their minds about? Who owns the hard decisions when the founders disagree?

Next, investigate the role. Ask your future manager what you will own in the first 30 days, what will count as progress by month three, and who will review your work. If you are the first person in a function, find out whether you are being hired to explore a real need or to cover several unplanned jobs at once.

Ask a future peer for the unglamorous details: “What gets reprioritized most often?” “How are decisions documented?” “When was the last time someone disagreed with a founder, and what happened?” Compare the answers with the manager’s answer. You are assessing how the company works when information is incomplete, which is where startup work differs from a structured corporate job.

If the offer includes options or other equity, treat that as a separate document review. Read startup equity explained before you place a large value on the grant. The share count is not enough to evaluate it.

Treat these startup red flags as prompts to verify

One concern does not automatically mean “decline.” Startups are messy by design. A pattern of unclear answers, missing documentation, and pressure is different.

An offer should tell you what you are accepting, including cash compensation, title, reporting relationship, start date, and equity terms when applicable. If you need to push for clarity or adjust terms, use these scripts for negotiating a startup offer. A company can be early and still be clear.

Score the opportunity before you compare offers

Use this scorecard after the interviews. Give each row a score from 0 to 3: 0 means no evidence or a direct contradiction; 1 means mostly claims; 2 means a specific, consistent answer; 3 means direct evidence plus consistent answers. Multiply by the weight. The maximum is 54.

Example: hypothetical score. A candidate gives money a 2, customers a 2, founder transparency a 3, role a 3, work practices a 2, offer terms a 2, and personal buffer a 3. With the weights above, the total is 44 out of 54. That is a reason to compare the weak categories, not permission to ignore a serious problem. Set two gates before adding the score: do not accept a 0 for money and runway, and do not accept a 0 for role clarity.

The scorecard also makes a startup-versus-established-company choice more honest. Use it alongside our startup vs. big tech comparison rather than assuming one path is always better.

Next steps

1. Open a note for the company and label every finding verified, reported, or inferred.
2. Run the 45-minute research sprint before your next interview, then write three gaps you still need answered.
3. Ask the runway script and the role-clarity questions to at least two people.
4. Score the offer only after you have the written compensation and equity terms.
5. If the answers remain inconsistent, choose the risk you can explain to yourself, not the story you most want to be true.