The Startup Due Diligence Checklist: What to Check Before You Write the Check
A working startup due diligence checklist for early-stage investors: team, market, product, financials, and the engineering signals public GitHub data adds before the data room opens.
Key Takeaway
Due diligence fails most often not because investors check the wrong things, but because they check them in the wrong order and skip the engineering layer entirely. This post lays out a four-workstream checklist that front-loads the cheap, public checks before the expensive legal work, with public GitHub activity as a fifth stream that reveals how fast a team is actually shipping.
Most seed-stage diligence failures do not come from checking the wrong things. They come from checking the right things in the wrong order, and from skipping the engineering layer that is the cheapest, most honest signal available before a data room exists.
This is a working checklist, ordered cheapest-first so the expensive legal work happens only on deals that already survived the public layer.
Why Order Matters in Due Diligence#
Diligence is a funnel, not a list. Every hour spent on legal review of a company you will pass on is an hour you did not spend sourcing the next deal. Front-loading the free, public checks filters the funnel before you pay for the expensive ones.
The public layer answers three questions no reference call can answer as cleanly: is the team real, is the product real, and is the team shipping. Reference calls tell you what people say; the public record tells you what they do.
The Four Workstreams#
**Team.** Who are the founders, what have they built before, and does their stated expertise match their track record. Check LinkedIn, GitHub, and prior employers. For technical founders, public commit history is a direct record of capability.
**Market.** Is the market real and growing, or is this a solution looking for a problem. Check comparable companies, funding in the category, and whether the wedge is credible.
**Product.** Does the product exist beyond a deck. Check for a live site, a working demo, real users, and public repositories that corroborate the product story.
**Financials and legal.** Cap table, incorporation documents, prior financing terms, and any outstanding obligations. This is where the NVCA model legal documents come in for priced rounds [3].
Where Engineering Data Fits#
Public GitHub activity sits between the team and product workstreams and strengthens both. The GitDealFlow methodology reads commit velocity, contributor growth, and repository expansion across a weekly panel of 350+ startup GitHub orgs [2]. The core finding, backtested against 219 fundraises, is that a composite of commit velocity and contributor growth precedes Series A announcements by 21 to 47 days with a 3.4x lift [1].
In diligence terms, that means you can read whether a team is accelerating or stalling before they ever send you a data room link.
The Five Things Most Investors Skip#
- Independent shipping verification, as covered above.
- Contributor breadth. A single founder committing to a single repository is a very different company from a team of eight shipping across five repositories, even at the same raw commit count.
- Repository segmentation. Are the active repositories product code or configuration and policy? Compliance and infra work can inflate activity without advancing the product.
- The founder's own GitHub history. Starring and commit patterns reveal what the founder actually pays attention to.
- Change over time. A snapshot tells you where the company is; the trajectory tells you where it is going.
A Working Checklist#
Run these in order, and stop as soon as a deal fails two checks in a row. Public presence and team reality first, market second, product third, shipping trajectory fourth, financials and legal last. Each check should end with a written one-line verdict so the decision is reviewable later.
Engineering acceleration is not a substitute for any of the four workstreams. It is the fifth stream that connects team and product, and the one that costs nothing to check before the data room opens.