Operating-Memory Diligence: What to Ask Before You Fund

You are not investing in an idea. You are investing in execution capacity, and the pitch deck tells you almost nothing about it.

Mission-driven projects fail in the operating layer far more often than in the thesis. The vision was sound. Decisions disappeared, governance stayed fuzzy, founder dependency became structural, reporting turned into storytelling, and by the time it showed up in a board update it had been true for a year.

Financial diligence is mature and you already do it. Operating-memory diligence barely exists, and it is where the risk actually lives.

Nine questions

Ask them of the team, together, and watch how they answer as much as what they answer.

1. "Show me the last significant decision you made and the reasoning behind it."

The single most informative question here. You are looking for a written record with the alternatives that were rejected. What you usually get is a confident verbal account from one person.

Concerning: only the founder can answer. Good: someone else finds it and reads it out.

2. "What are you contractually committed to, and when does each end?"

Ask for the list. Most early projects do not have one anywhere. This is where the surprises live: an auto-renewing lease, a grant with reporting obligations nobody tracked, a supplier arrangement on terms nobody remembers agreeing.

3. "If the founder were unavailable for a month, what would stop?"

Push past "we'd manage". Ask which specific decisions would queue. Name three and ask who would make them.

4. "What have you decided twice?"

Every organization has re-decided something because nobody could find the first decision. A team that can name an example is being honest and has noticed the problem. A team that says "nothing" has either not noticed or is managing you.

5. "Who can spend money without asking, and up to what?"

If the answer is nobody, every purchase queues behind the founder and you have found the bottleneck. If the answer is vague, the control is vague.

6. "What did you decide NOT to do in the last year, and why?"

Tests whether purpose is being used to make decisions or only to raise money. A project that has turned nothing down has no boundary, and a project with no boundary drifts. See Purpose Drift.

7. "How does someone new find out why things are the way they are?"

If the answer is "they ask", the organization has no memory outside its people, and every departure is a data loss event.

8. "What is your runway, and who else knows that number?"

Ask the founder privately, then ask two others. The spread tells you how much of the team is making decisions with information. See Open Books.

9. "What would you tell me if you were not raising?"

Asked last, and asked plainly. The answer is sometimes the most valuable thing in the meeting.

What the answers mean

A well-run project can produce documents. Not polished ones. A decision record with a date, a commitments list, a runway figure that three people give consistently. The existence of the artifact matters more than its quality.

Watch who answers. If the founder answers all nine, you have measured founder dependency directly, whatever they said in question three.

Distinguish young from fragile. A two-person project should not have formal governance, and demanding it is unhelpful. What it should have is the founder able to say what they would build first as they grow. Ask that instead.

After the investment

Diligence answers a question once. These are worth revisiting, because operating memory decays quietly.

Ask for the decision record with the quarterly report. Not more reporting — the same reporting, source-backed. A board update that cites what was decided and when is checkable. One that narrates progress is not.

Watch the concentration numbers. Client concentration, and how many decisions still route through one person.

Notice when reporting gets smoother. A report that stops containing bad news is not evidence that bad news stopped.

Where it goes wrong

Requiring a system you have chosen. Mandating a specific tool creates compliance, not capacity. Require the artifacts and let the team pick how.

Diligence as interrogation. These questions work in a conversation and fail as a questionnaire, because the informative part is the hesitation and who fields it.

Confusing polish with capability. The most impressive deck often comes from the project with the least operating discipline, because that is where the effort went.

Treating a "no" as failure. A project that says "we do not have that written down" is telling you the truth and has told you exactly what your first support dollar should buy.

What good looks like

You leave with three documents and a clear picture of what would break, and the team leaves knowing what you will ask for next quarter.

And when something does go wrong eighteen months later, you find out from the operating record rather than from a resignation.

Try this week

Take a project already in your portfolio and ask questions one and four.

If nobody can produce a decision record, you have not found a failing project. You have found the cheapest, highest-leverage support you can offer one.