The Complete Guide to Startup Due Diligence

Due diligence is the deep investigation you run before committing to a startup. Where screening ranks many companies quickly, due diligence goes deep on the few finalists, verifies what the decision rests on, and turns the findings into a memo a committee can act on.

Screening gets you to a shortlist. Due diligence is what stands between that shortlist and a check. It is the difference between "this looked good in the ranking" and "we know what we are backing and where the risks are." Skipped or rushed, it is also where the avoidable losses come from: the inflated number nobody re-checked, the cap table surprise, the dispute that was public the whole time.

This guide covers when to run due diligence, the themes a thorough process covers, a checklist you can reuse, why you should re-verify the doubtful rather than everything, what belongs in the memo, and how the process ends in a defensible verdict.

In this guide

• When to run it
• The themes a thorough process covers
• The due diligence checklist
• Re-verify the doubtful, not everything
• Red flags are findings, not footnotes
• The memo and the verdict
• Common mistakes
• FAQ

When to run it

Due diligence is expensive in time and attention, so you run it on finalists, not on the field. Screening has already done the coarse filtering and told you which companies clear the bar. Due diligence spends real effort only on the handful you are seriously considering.

The transition matters. Everything screening surfaced, the scores, the claims, the flags, is the starting point for due diligence, not something to redo from scratch. Good process carries that context forward. It does not throw it away and start the reading again. A finalist that arrived with two flagged claims and a strong thesis fit tells diligence exactly where to dig first.

The themes a thorough process covers

Due diligence is broad by design, because the thing that sinks a deal is usually the thing no one looked at. A thorough process covers, at minimum:

• Claims and traction. The numbers the decision rests on, verified against evidence rather than accepted from the deck.
• Financials. Revenue quality, unit economics, burn and runway, and whether growth is healthy or subsidized.
• Cap table. Ownership, option pool, prior terms, and anything that complicates the next round.
• Legal standing and disputes. Litigation, regulatory exposure, and corporate standing in the official registers.
• Intellectual property. What is actually filed and owned, checked against USPTO, EPO or INPI, versus what a slide claims.
• Team and references. Track record, founder-market depth, and what backchannel references actually say.
• Compliance. Sector-specific obligations and, for regulated evaluation, the governance the deal will have to satisfy.

The list is not the work. The judgment is knowing which themes carry the most risk for this specific company and going deepest there. A deep-tech company lives or dies on the IP theme; a marketplace on unit economics; a regulated fintech on compliance.

The due diligence checklist

A reusable starting checklist, to adapt to the deal:

Claims

• Every headline claim extracted and given a verdict against public sources
• Named partnerships and customers confirmed with the third party, not just the deck
• Market-size figures checked against analyst sources and the reachable segment

Financials

• Revenue recognition and quality of revenue understood
• Burn, runway and the assumptions behind the next-round timing
• Unit economics that hold before scale, not only after

Structure and legal

• Cap table, option pool and any unusual prior terms
• Litigation, liens and regulatory actions in the registers
• Corporate standing and incorporation confirmed

Team

• Founder track record and founder-market fit
• Backchannel references beyond the ones offered
• Key-person risk and retention

IP and compliance

• Patents and IP ownership filed and assigned correctly
• Sector obligations and, for candidate evaluation, AI Act posture

The point of a checklist is not to tick boxes uniformly. It is to make sure the theme that would have sunk the deal was at least looked at.

Re-verify the doubtful, not everything

A naive due diligence process re-checks every fact from zero. That is slow, and it wastes effort on claims that were already solidly verified upstream. A sharper process does the opposite: it reuses what is already proven and concentrates the deep, expensive work on what is doubtful, contradicted, or simply unchecked.

This is where the verdict statuses from screening earn their keep. A claim already marked verified, from concordant public sources, does not need re-litigating. A claim marked qualified, contradicted, or to-confirm is exactly where diligence should spend its hours. Diligence is triage before it is investigation.

The result is depth where it matters and speed where it does not, instead of uniform slowness that runs out of time before it reaches the hard questions.

Red flags are findings, not footnotes

Due diligence exists to surface the things that should change or stop a decision. A red flag is any finding serious enough to weigh against the deal on its own: a contradicted claim about traction, an undisclosed dispute, a cap table that will not survive the next round, a compliance gap in a regulated market.

The discipline is to treat a red flag as a first-class output, not a line buried on page nine. Surface it, state what it is, and let it carry its real weight against an otherwise strong case. The whole point of diligence is that a serious problem cannot hide behind an impressive average.

A useful test: if a finding would change the decision and it is not on the first page of the memo, the memo is written wrong.

The memo and the verdict

Due diligence ends in an artifact, not a feeling. The memo lays out the company, the verified facts, the risks, and the open questions in a form the committee can read and argue with. It gives everyone the same basis for the decision, the same way a scoring basis does for screening.

And it ends in a verdict: a clear position, backed by the evidence, that the committee can accept, challenge, or reject. The verdict is not the machine's decision. It is the case, assembled and sourced, handed to the people whose job is to decide. Under the EU AI Act, evaluating people and companies for high-stakes outcomes calls for human oversight, and a diligence process that hands the committee a defensible case, not an automated ruling, is built for exactly that.

Common mistakes

• Diligence on the field instead of the finalists. Depth is expensive. Spend it where the decision is live.
• Re-verifying everything. Reuse what is proven. Concentrate on the doubtful.
• Starting over from the deck. Carry the screening context forward or repeat the work.
• Burying red flags. A finding that does not change the score was not treated as a finding.
• A verdict without a memo. A conclusion no one can audit is an opinion, not diligence.
• Uniform depth. Going equally deep on every theme runs out of time before the one that matters.

Frequently asked questions

What is startup due diligence? Due diligence is the deep investigation an investor or program runs before committing to a startup. It verifies the claims, checks financials, cap table, legal standing, IP and risks, and produces a memo and a verdict a committee can act on.

How is due diligence different from screening? Screening ranks many companies quickly into a shortlist. Due diligence goes deep on the few finalists. Screening is breadth; diligence is depth. Good diligence reuses what screening already verified rather than starting over.

What should a due diligence memo contain? The company, the verified facts, the risks and red flags on the first page, the open questions, and a clear verdict backed by evidence. Anyone on the committee should be able to audit how the conclusion was reached.

How long does due diligence take? As long as the risk requires and no longer. Concentrating on the doubtful rather than re-checking everything is what keeps it from expanding to fill all available time before it reaches the hard questions.

Can due diligence be automated? Parts of it: extraction, claim verification, gathering the public record. The judgment and the decision stay human, which is also what the EU AI Act expects for high-stakes evaluation.

The bottom line

Due diligence is where a shortlist becomes a decision you can stand behind. Run it on finalists, carry the screening context forward, re-verify the doubtful rather than everything, put the red flags on the first page, and end in a sourced memo and a clear verdict that a human owns.

How this shows up in Deckwise

Deckwise's Due diligence pillar runs on the finalists a screen produces. It reuses what screening already proved and re-checks only what is doubtful, then covers the diligence themes above into a memo and a verdict ready for the committee. The risks are surfaced before the decision, not after it. The benefit, stated plainly: you commit with conviction, and you can defend every part of the case that got you there.

Back to The Deckwise Method, or start upstream with The Complete Guide to Startup Sourcing.