By stage
Seed diligence
What to check once there is a product and early revenue — where the data is real but too thin to prove anything, and the risk is over-reading it.
Realistic time Eight to fifteen hours over three weeks
At seed there is finally something to look at: a product in use, some revenue, a handful of customers, a few months of data. The temptation is to treat that data as evidence. Mostly it is not — twelve customers and six months of history cannot establish retention, unit economics or a repeatable sales motion, and confident conclusions drawn from them are usually wrong.
What seed data can do is raise questions. A cohort chart with a sharp early drop, a customer list where one name is 60% of revenue, a sales cycle that has lengthened each quarter — none of these is conclusive and all of them are worth asking about. Diligence at this stage is about finding the questions, not the answers.
The second thing that becomes assessable at seed is execution. There is now a track record, however short: things the founders said they would do, and whether they did them. Ask for the deck from the last raise and compare it with what happened. That comparison is the most informative single document in seed diligence and almost nobody asks for it.
01
Customers and revenue
Start with the customer list and the revenue by customer. Concentration is the most common structural weakness at seed, and it is invisible in an aggregate revenue figure. A company with $400k of revenue and one customer at $250k of it is a different proposition from one with forty customers averaging $10k.
Then look at how customers were acquired. Founder-led sales are normal and healthy at seed; the question is whether anything works that does not depend on a founder in the room. A company where every customer came through the chief executive's personal network has not yet demonstrated that it can sell, only that the founder can.
Revenue quality matters more than quantity. Recurring contracted revenue, pilot revenue that will not renew, one-off services revenue and grants are all revenue, and they are worth very different amounts. Ask for the split and expect the founders to know it without looking.
Check
- Revenue by customer — what is the concentration?
- How was each of the largest customers acquired?
- What proportion of revenue is recurring, contracted and likely to renew?
- How much of it is pilots that have not converted?
- Have any customers churned, and what did they say when they left?
- What is the pricing, and has it changed? Why?
02
Retention and engagement
Retention is the metric that most nearly predicts a venture outcome, and at seed there is rarely enough history to measure it properly. What you can do is look at the shape of the earliest cohorts, however small, and ask whether usage is deepening or decaying.
Be careful about the denominator. Retention measured on paying customers looks very different from retention measured on sign-ups, and companies naturally report the flattering version. Ask which is being shown and ask to see the other.
Engagement data is often more useful than retention at this stage because there is more of it. Are people using the product weekly? Is usage per account growing? Are new features adopted? These are answerable with six months of data in a way that annual retention is not.
Check
- Cohort retention by month, on paying customers.
- Is usage per account growing, flat or declining?
- What proportion of accounts are genuinely active?
- What does the company consider its core engagement metric, and why?
- Talk to two or three customers directly if the founders will allow it.
03
Execution track record
Ask for the deck the company used at its last raise. Compare the plan with what happened: the milestones, the hiring, the revenue projections, the product roadmap. Founders who hit roughly what they said they would hit are demonstrating something that no current-quarter metric can.
Missing a plan is not disqualifying — most startups do, and the plan was a guess. What matters is whether the founders can explain the gap accurately and without defensiveness, and whether the explanation suggests they learned something.
Look at hiring too. Who have they recruited, from where, and did those people stay? A seed company that has attracted strong people from better-paid jobs is showing you something about the founders that is hard to fake.
Check
- Ask for the previous fundraising deck and compare it with reality.
- What did they say they would do with the last round, and did they?
- Who has joined, from where, and has anyone left?
- How has the burn rate tracked against plan?
- What is the current runway, precisely?
04
Money and structure
At seed the financial picture becomes real enough to examine. You want the monthly burn, the runway, the cash in the bank as of a recent date, and the plan for the round you are joining — specifically, what milestone it is meant to reach.
The structural checks from pre-seed still apply, plus a few more: the full cap table with convertibles modelled, any debt, any grants with conditions attached, and whether the company has the licences and registrations its business actually requires.
Check
- Cash in bank, monthly burn, and runway in months.
- What milestone does this round reach, and does the money cover it?
- Full cap table, fully diluted, with convertibles modelled.
- Any debt, grants with clawbacks, or government funding conditions?
- Are the required licences and registrations in place?
- Are contractor and employee IP assignments complete?
Stop and think
Red flags
- One customer representing more than half of revenue, presented as validation.
- Retention shown on sign-ups rather than paying customers, with reluctance to show the latter.
- A significant gap between the last round's plan and what happened, explained defensively.
- Every customer acquired through a founder's personal network, with no other channel tested.
- Burn that has grown faster than revenue with no corresponding hiring plan.
- Reluctance to introduce you to any customer.
- A runway of under four months at the point the round is being raised.
Take these into the room
Questions to ask the founders
- Which customer would hurt most if you lost them, and how likely is that?
- What did the last round's plan say, and where did reality diverge?
- Which metric are you most worried about?
- What have you tried that did not work in the last six months?
- Who is the best hire you have made, and how did you get them?
- What does this round have to prove for the next one to happen?
- If growth stopped tomorrow, how long could you operate?
From the decoder