How to Run Due Diligence on an Early-Stage Startup: My First-Cheque Checklist

Most first cheques are written on conviction and a deck. I have sat on both sides of that table, raising capital and deploying it, and the pattern is consistent: the investor who loses money rarely missed a red flag, they simply never looked for one. Due diligence on an early-stage company is not a document review, it is a short, disciplined investigation with a fixed scope and a deadline.

Here is the process I use at IRIS Vision Capital before a first cheque goes out, written so an aspiring angel can run it the first time and an experienced investor can steal the parts that tighten a weak stage.

Start with your own mandate, not their deck

Diligence without a mandate is just curiosity. Before you open a data room, write three lines you can show someone: the cheque size you will write, the stage and sector you will write it in, and the outcome that makes the investment work. Angels investing through groups have historically returned around 2.6x over 3.5 years, roughly a 27% IRR, according to the Wiltbank and Boeker research on returns to angel investors in groups. That average hides a portfolio where most positions return nothing. So your mandate has to assume concentration of returns, which means every company you back needs a credible path to a large outcome, not a comfortable one.

Write the mandate down once and reuse it. It turns a vague feeling of unease into a specific disqualification you can say out loud: this is a good business, it is not my business.

Diligence does not tell you whether a company is good. It tells you whether what the founder believes about the company is true.

Week one: the four questions that end most deals

Do not begin with legal documents. Begin with the four questions that have the highest chance of ending the process, because the cheapest diligence is the diligence you never have to finish.

1. Is the problem expensive for the buyer? Ask the founder which line item in the customer's budget this comes out of. A founder who cannot name the budget line is selling a nice-to-have.
2. Is there repeat, unprompted usage? Pilots prove interest. Second purchases, renewals and unbudgeted expansions prove need.
3. Can this team reach the next milestone with this round? Not the vision. The next funding milestone, with the money actually in the bank.
4. Is the market structurally open? Ask who the incumbent is and why they have not solved it. If the answer is that incumbents are lazy, keep digging. If the answer is a structural constraint, a regulation, a channel conflict, a cost base they cannot cut, you have something.

If any of these fails badly, stop and tell the founder why in plain language. You will get better deal flow from founders you declined honestly than from founders you ghosted politely.

Verify the numbers yourself, line by line

Every deck I see reports growth. The job is to find out which growth. Ask for the raw export, not the summary slide, and rebuild three things in your own spreadsheet.

• Revenue by customer by month, for the last 18 months if it exists. You are looking for concentration. If one logo is more than 30% of revenue, the company has one customer and a hobby.
• A monthly cohort retention table. Take every customer who started in a given month and track whether they are still paying 1, 3, 6 and 12 months later. Churn hidden by new sales is the most common distortion in early-stage reporting.
• Bank statements against the P&L. Three months is enough. You are checking that revenue recognised is cash collected, and that payroll and burn match the story.

Then compute two numbers the founder may not have: net revenue retention, and months of runway at current net burn with the new round included. Ask what the plan is at six months of runway, not at zero. Founders who have thought about the bridge are easier to back than founders who have only thought about the raise.

On the fundraising environment, be realistic about what the next round demands. Carta's data shows that of companies raising a seed round in 2022, only around 17% reached Series A within two years, against a historical norm closer to 25% to 30%. That is the base rate your investment is priced against. If the plan needs a Series A in 18 months to survive, you are underwriting the market more than the company.

Diligence the founder, not the narrative

I have never seen a pitch fail because the founder was inarticulate. I have seen many fail because the founder could not be questioned. Use the meeting to test how they handle pressure on their own data.

Three questions that reveal more than an hour of pitching:

• What did you believe 12 months ago that you no longer believe, and what changed your mind?
• Who is the best person you tried to hire and failed to close, and why did they say no?
• Walk me through the last customer you lost, from first contact to the day they left.

Then take references, and take them properly. Two founder-selected references tell you nothing. Ask for one former colleague who no longer works with them, one customer who churned, and one investor from a previous round. Call them yourself. The single most useful question on a reference call is: what should I know that they would not tell me? The silence before the answer is often the answer.

As a Christian I take the stewardship side of this seriously. Capital I deploy is not only mine to lose, it represents other people's work and other people's trust. Honest diligence is a form of respect, both for the person whose money it is and for the founder who deserves a clear yes or a clear no rather than a slow fade.

The legal and cap table pass

Ownership

Get the full cap table as a spreadsheet with every instrument listed: shares, options, SAFEs, convertible notes, advisor grants. Model the dilution post-round including conversion of every instrument. I have seen founders discover during diligence that they owned 20 points less than they thought. If the founding team holds under 50% at seed, the next round becomes structurally hard to raise.

Rights and obligations

Read for the things that bind the company's future: intellectual property assigned by every contributor including contractors, customer contracts with unusual termination or exclusivity clauses, debt and personal guarantees, any side letter promising information or veto rights to an earlier investor. Founder vesting matters more than most angels think. A co-founder with fully vested shares and no role is a liability on every future round.

Price the deal and do the exit maths

Valuation is where enthusiasm does the most damage. The discipline is simple: work backwards from a plausible exit, not forwards from a hoped-for multiple.

Take an exit value you can defend by pointing at a comparable transaction in that sector. Apply dilution across the rounds the company will need, 20% to 25% per round is a reasonable working assumption. Then ask what your position is worth at exit. If your realistic case does not return the whole value of your intended portfolio, the entry price is wrong, however much you like the team.

You cannot fix entry price later. Every other mistake in early-stage investing has a remedy. That one does not.

The failure modes, and what to do instead

• Diligence by social proof. A respected name on the cap table replaces your judgement with theirs, and you have no idea what their mandate was. Do your own work, then look at the syndicate.
• Open-ended process. Diligence that drifts for three months kills the deal and your reputation. Give yourself a fixed window, two to three weeks, and a named decision date.
• Confusing a nice founder with a backable one. Warmth is not a signal. Clarity under questioning is.
• Skipping the churned customer. It is the most uncomfortable call and the most informative one.
• Negotiating terms before the facts are in. Terms discussed early become anchors you cannot move once the numbers come back worse than the deck.

If you want to see this from the other side of the table, the founder-facing version of the same checklist is in investor readiness is a data room problem, and more on capital and deal flow sits in the Investment collection.

What to do this week

Write your mandate: cheque size, stage, sector, and the outcome that makes it work. Three lines, nothing more. Then take the next deal in your inbox and run only the four week-one questions against it before you read the deck a second time. You will decline faster, you will decline for reasons you can defend, and the deals that survive will be worth the full process.