BackStartups

The desk

Diligence, sized for an angel cheque

Funds spend six weeks and a legal budget on diligence. You have a few evenings and no budget at all. These 17 guides cover what is worth doing at that scale — what can genuinely be checked at each stage, what only matters in certain sectors, and the handful of questions that consistently earn their time.

3
Stage guides
7
Sector guides
7
Craft guides

Start with these three

If you have never run diligence on a startup, the order that works is: read the stage guide that matches the round, make three reference calls, then read the term sheet from your own position rather than the lead's. Everything else is refinement.

7 guides

By sector

The questions that only matter in one kind of company, and the metrics that mean something different in each.

7 guides

The craft

Reference calls, data rooms, financial models, red flags and the first term sheet — the parts of diligence that are the same everywhere.

Angel diligence: common questions

How much diligence should an angel actually do?
Proportionate to the cheque. Four to eight hours at pre-seed, eight to fifteen at seed, and rather less at Series A where an institutional lead is doing the heavy work. Beyond that you are generating precision the available information cannot support.
What is the highest-yield hour in angel diligence?
Reference calls on the founders. They are the only source of information about how someone behaves over years rather than over a pitch, they consistently surface things no founder meeting would, and they are the step most often skipped because asking feels awkward.
Can I do useful diligence when a fund is leading the round?
Yes, on different questions. The lead assesses whether the company suits their fund on their terms. Nobody is assessing what the round means for a small cheque taking a different share class with no board seat — that part is yours, and it takes about an hour.
Does diligence actually improve angel returns?
It cannot prevent the failures that come from a market not existing or a team falling apart, which is most of them. What it prevents is the specific, visible failures: unassigned IP, a departed co-founder holding a third of the company, a preference stack that makes your shares worthless in any realistic exit.

Next

Diligence tells you about one company

It does not tell you how many companies you need to back, how large each cheque should be, or how much to hold back for follow-ons. That is a different kind of arithmetic.