Due diligence is where a fund moves from what a founder says to what the fund can stand behind. It is also where many early-stage funds lose time: requests go out by email, documents arrive in five places, the same question is asked twice, and by the time the committee meets nobody is sure which version of the revenue number is current.
A good VC due diligence process is not about checking everything. It is about checking the few things that would change your decision, recording what you found and where, and doing it without wearing out a founder who may be talking to several funds at once. The sections below cover the stages, the workstreams, how to scale the work by stage, and the mistakes worth avoiding.
Where diligence sits in the deal process
Survey research gives a useful picture of the overall funnel. Gompers, Gornall, Kaplan and Strebulaev surveyed 885 venture capitalists at 681 firms. For every deal a firm closed, it considered about 101 opportunities, met the management of about 28, reviewed about 10 at a partners meeting, ran due diligence on about 4.8 and offered about 1.7 term sheets. Diligence comes late, after a partner has already decided the company is worth serious attention.
The same study reported that the average deal took 83 days to close, that the average firm spent 118 hours on due diligence over that period, and that firms called 10 references on average. Early-stage firms called fewer references (8 on average) than late-stage firms (13), and spent less time overall. These figures are from a 2016 working paper and describe averages across many kinds of funds, so treat them as context rather than a target.
The VC due diligence process in seven stages
Funds organise the work differently, but most follow a sequence like this one. Writing it down, with an owner for each stage, is the single biggest improvement a small team can make.
- 1Decide to start. A partner sponsors the deal and agrees to spend real time on it. Write down why, in two or three sentences.
- 2Agree the key questions. List the three to five things that would change your decision, such as whether revenue is recurring or whether the founder can hire engineers. Everything else is secondary.
- 3Request information once. Send one organised request list with clear owners and due dates, rather than questions spread across calls and emails. Our diligence information request list is a starting point.
- 4Work the evidence. Review documents, call customers the founder names and some they do not, call references, and test the numbers against the deck.
- 5Record findings as you go. Every finding should say what was checked, against which source, and what it showed.
- 6Resolve or escalate open issues. Some questions will not be answered. Decide which are acceptable risks and which block the deal.
- 7Write it up for the decision. Summarise what holds, what was disputed and what remains unknown, so the people deciding can see the evidence behind each point. Our guide on how to write an investment memo covers the write-up.
What to check: the main workstreams
Most diligence falls into a handful of workstreams. The table shows what each covers, the evidence that answers it, and how much depth an early-stage deal usually needs.
| Workstream | Key question | Typical evidence | Depth at seed |
|---|---|---|---|
| Team | Can these founders build this company? | Reference calls, past work, how they answer hard questions | High: often the main focus |
| Customers and traction | Are the reported customers and revenue real and as defined? | Customer calls, invoices or billing exports, usage data | High for any number in the deck |
| Market | Is the market large enough and reachable from where they start? | Bottom-up sizing, customer interviews, industry experts | Medium |
| Product and technology | Does the product work and can the team keep building it? | Product demo, technical call, architecture overview | Medium, higher for deep technology |
| Financials | Does the plan make sense and does the round fund it? | Historical accounts if any, budget, burn and runway | Light to medium |
| Legal and cap table | Does the company own what it sells, and who owns the company? | Incorporation documents, cap table, IP assignments, key contracts | Focused on blockers |
| Security and data | Is customer data handled responsibly? | Security questionnaire, policies, past incidents | Light unless data is central to the product |
The legal workstream usually connects to closing documents. In the United States, many financings start from the NVCA's model legal documents, such as the stock purchase agreement, investors' rights agreement and voting agreement, which the NVCA updates from time to time. Your counsel leads that work, but diligence findings, such as an IP assignment that was never signed, often feed directly into it.
Scale the work to the stage and the cheque
A pre-seed cheque into a two-person company and a lead investment in a growth round need very different processes. The mistake is to run the later-stage process on an early company, or the early-stage process on a larger cheque.
- Pre-seed and seed: focus on the team, the first customers and whether any traction claims hold. Ask for a short list of documents. Most funds rely heavily on reference and customer calls.
- Series A and later: add a revenue quality review, cohort data, a deeper financial model, a legal review by counsel and often a security review.
- Following, not leading: if another investor is leading, ask what they have checked and what they found, but still check the points that matter to your own decision.
- Exceptions to your thesis: when a deal falls outside your usual rules, spend extra time on the specific reason it is an exception.
Founders talk to each other about how funds run diligence. Asking for everything at once, repeating questions they have already answered, or disappearing for weeks does more damage than a quick, clear decline. Our guide on what goes in a startup data room shows the process from the founder's side.
Common due diligence mistakes
- No key questions. Without them, diligence becomes a checklist exercise that produces a long document and no clearer decision.
- Checking only what the founder offers. Call at least one customer or reference you found yourself.
- Losing the definitions. "Revenue" in the deck, the model and the data room may mean three different things. Write the definition next to every number you check.
- Overwriting the past. When a founder updates a number, keep the earlier one. A change between versions is itself evidence.
- Scattered requests. Questions in calls, emails and messages produce duplicate work for the founder and gaps for you.
- Confirmation bias. Once a partner is excited, diligence can drift towards finding support. Assign someone to write the strongest case against the deal.
Our VC due diligence checklist lists items by workstream, and our guide on how to verify startup traction goes deeper on revenue and customer claims.
Where Delsivo fits
Delsivo is designed to keep diligence in the same record as the rest of the deal. In the pilot release, analysts send founders follow-up forms and scoped file requests through the fund's branded portal. Requests move through clear states, from Sent to Submitted to Complete, and reminders go only for items still outstanding. Uploads are scanned for malware before anyone opens them.
Every material claim from the deck, forms and documents sits in a claim tracker with its definition, period, units and source, labelled founder reported, corroborated, disputed or unresolved. When a fact changes, the earlier statement is kept. Data rooms with folders, permissions, guest access, watermarks and activity logs are planned for the first release. Committee packets and decision records come in a later release. AI suggests claims and questions and cites its sources. Your team checks the evidence and makes every decision. See data rooms for the planned design.
Sources
- Gompers, Gornall, Kaplan and Strebulaev: How Do Venture Capitalists Make Decisions? (NBER Working Paper 22587, September 2016) (checked 6 November 2026)
- NVCA: Model legal documents (checked 6 November 2026)
Delsivo publishes this guide and makes software for venture funds, so we have a stake in the topic. Facts about other companies, research and regulations come from the sources listed, on the dates shown, and can change. Corrections are welcome through our contact page.