Early Stage Investing Due Diligence Checklist: 7 Must-Have Steps

You've got the term sheet drafted, the founder is responsive, the deal feels right. Then someone on your investment committee asks a simple question about the cap table, and nobody in the room can answer it. I've watched a $150K check stall for three weeks over exactly that — not because the deal was bad, but because the diligence process had no order to it. Founders got asked for the same document twice. Reference calls happened after the IC memo was already written. It was chaos dressed up as rigor.

So here's the thing: an early stage investing due diligence checklist isn't about cramming a pre-seed or seed deal into an enterprise procurement template. It's about knowing which questions actually move the needle at this stage, in what order, and what you can safely skip. I'll share the checklist I've refined across roughly 40 seed-stage deals, including the three I got badly wrong.

What actually matters in an early stage due diligence checklist

The first time I ran diligence on a seed deal, I spent four days building a 60-tab Excel spreadsheet with a full three-statement financial model. The startup had $8K in monthly revenue. I couldn't fill in half the tabs, and the two that mattered — founder-market fit and cohort retention — I barely touched. Total waste of time.

Early stage diligence is a different animal from growth-stage diligence. At Series A and beyond, you're verifying numbers. At seed, the numbers barely exist, so you're verifying people and hypotheses. The weight shifts hard.

Here's the split I now use, and I'll defend it:

Diligence area	Weight at pre-seed/seed	Weight at Series A+
Team & founder references	~35%	~15%
Market & competitive position	~25%	~15%
Product & technology	~15%	~20%
Traction & metrics	~15%	~30%
Legal & cap table	~10%	~20%

Notice legal is only 10% at seed. That doesn't mean skip it. It means a missing IP assignment is a fixable red flag; a founder who's been quietly misrepresenting traction is not.

What you can genuinely skip at this stage

Full financial audits. Customer concentration analysis (there are no customers yet). Detailed unit economics models that assume a sales motion that doesn't exist. I've seen investors demand a five-year P&L from a company with a prototype and two founders in a co-working space. That request tells the founder more about you than your questions tell you about them.

The 90-minute rule

One practical constraint I impose on myself: the first pass of a seed checklist should take about 90 minutes of reading plus 3 reference calls. If it's taking six hours, you're either over-engineering or avoiding the real question — do I believe in these people?

The sequence: how early stage due diligence actually runs

Most checklists list categories. Almost none give you an order. Order is where the time savings live, and where I've lost the most deals to drift.

Days 1–3: the screen and the first data request

You should be able to make a go/no-go on continuing diligence within three days of a founder meeting. At this point, request exactly four things:

• The cap table, as it stands, with all option grants and convertible notes
• Incorporation documents and confirmation all founders have signed IP assignment agreements
• Whatever traction artifacts exist — Stripe dashboard export, an analytics screenshot, a list of signed LOIs
• A one-paragraph answer to "what's the biggest risk to this business?"

That last one is my favorite filter. I've found founders who answer it honestly and precisely are, on average, materially better operators than the ones who deliver a pitch-deck paragraph dressed up as a risk.

Days 4–8: deep-dive and references

This is where you read the data room properly and run reference calls. Here's the part almost every template gets wrong: you should be doing references before you write the IC memo, not after. Sounds obvious. I've violated it exactly three times, and in two of those cases the reference call surfaced a co-founder departure that killed the deal after we'd already verbally committed. Embarrassing.

Days 9–10: the IC memo and term sheet

By day ten, either you're writing an investment memo or you're sending a clean pass. The memo should be one page. If you can't compress the thesis into a page, you don't have a thesis.

The checklist itself, in the order I actually use it

Team and founder verification

This is where you win or lose. Beyond the standard "tell me about yourself," the questions I actually run:

1. How long have you known your co-founder, and how did you meet? (If the answer is "we met at a hackathon six months ago," you know your equity split risk is real.)
2. What's the last hard decision you made together, and who disagreed?
3. Who on your team could replace you if you got hit by a bus tomorrow?
4. What would make you shut this down?

Then the reference calls. Insist on at least two names you choose yourself, ideally former colleagues or managers from the founder's prior roles. Expect 20-30% of references to go suspiciously flat — polite, vague, no specifics. That's a signal. A genuinely strong reference will offer unprompted detail: "She shipped the product that saved our Series B" or "He was the only person who stayed calm during the outage."

Market analysis

You are not going to out-analyze a founder on their own market in a week. What you can do is check whether they're being honest about the size of the opportunity. If a founder presents a $50B TAM from a top-down report, ask them to walk you to it bottom-up: "If you win, how many customers pay how much per year?"

In my experience, roughly a third of founders cannot answer this without a spreadsheet they didn't bring. Not fatal — but it tells you where the thinking is shallow.

Product and traction metrics

The metrics that matter at seed are narrow. I look for four numbers:

• Week-4 retention for consumer or self-serve products — anything below ~20% at seed is a warning sign, and I've seen it hold across dozens of deals
• Logo retention and NRR for B2B — a seed-stage NRR above 110% is rare and worth a premium
• Burn multiple (net burn divided by net new ARR) — anything above 3 at seed means you're buying revenue rather than finding fit
• Founder hours per deal closed — if the founder is doing all the sales and it's working, that's fine; if it's not working and they're doing all the sales, that's a problem

Ask for cohort retention charts, not a single retention number. A single number can be sliced to say anything. Cohorts can't.

Legal, cap table, and the boring stuff

Four documents I will not close without seeing:

• Cap table, fully diluted, including every convertible instrument and its conversion terms
• Founder IP assignment agreements, signed
• Any prior employment agreements that might contain non-competes or IP clauses that conflict
• Confirmation no litigation is pending or threatened

If you're using a free template or an Excel version of an early stage investing due diligence checklist, this is the section to double-check against a founder's actual documents. Templates tend to list the categories; the actual conflict usually hides in a specific convertible note issued eighteen months ago with a 20% discount and no cap.

The questions you'll get asked about this checklist

How long should early stage due diligence take?

Ten working days, from first real meeting to term sheet, is the target I aim for. Seed deals that drag longer than three weeks tend to die from momentum loss rather than from a discovered problem. If you can't decide in ten days, the honest answer is usually that you don't believe the thesis enough to defend it at IC.

What should a founder prepare before diligence?

A data room with five folders (corporate, financials, product, customers, hiring) is enough. Founders who've never done this often over-prepare. That's fine. What founders get wrong is the reverse: sending documents in a slow drip over two weeks. Send everything at once. The founder who says "I'll get you the cap table by Friday" and actually does — that's a data point in their favor.

Should I use a template or build my own?

Start with a template, then cut it in half. Any published checklist — Cooley's, an Excel template from a VC blog, a free download — will be over-engineered for seed. Take the categories you actually use and drop the rest. I built my own after three years of template use because I found myself ignoring 40% of the rows every single time. A checklist you actually fill in beats a comprehensive one you don't.

What I got wrong, so you don't have to

Three deals I still think about. The first: a B2B SaaS company where I ran full diligence, closed, and discovered six months in that the "co-founder CTO" had actually never signed his IP assignment. The fix took four months and a lawyer. Now I ask for signed assignments in the first fifteen minutes of the data room review.

The second: I skipped references because the founder was so impressive I didn't want to hear anything that might pop the balloon. That's not diligence. That's confirmation bias with a spreadsheet attached.

The third: I ran diligence on a company for six weeks, kept finding small issues, kept going. By the time I passed, the founder had moved on to another lead investor, and I'd burned six weeks of my own attention. The lesson: diligence should confirm or refute a thesis, not hunt for excuses to keep talking to a founder you like.

What I've learned is that the checklist isn't the thing. The discipline is. A founder who tells you upfront "we've got a messy cap table from a bridge round two years ago" is nine times out of ten a better bet than one whose documents are spotless and whose references are vague. The point of diligence isn't to find the perfect deal. It's to know exactly what you're buying before you write the check — and to have the spine to walk if the answer is no.