July 18, 2026 · IC memos, Underwriting
How to write an IC memo (the structure disciplined funds use)
By Emily Buckley, founder of SamCIO
The investment committee memo is the most honest artifact a fund produces. The deck is the founder's story. The model is a set of assumptions wearing math. The memo is the only document where the firm says, in its own voice, what it believes and why.
Which is exactly why bad memos are so expensive. A memo that just summarizes the deck adds nothing. A memo that varies wildly by author makes deals incomparable. And a memo written after the decision, to justify it, is worse than no memo at all.
Here is the structure disciplined funds converge on, and the standards that separate a defensible memo from a decorative one.
The structure
1. Recommendation, first
Open with the call: Pursue, Diligence, or Pass, and the two sentences of reasoning that carry the most weight. Committees read the first paragraph carefully and skim the rest under time pressure. If your recommendation is buried on page six, the memo is not driving the decision, the meeting is.
2. Thesis fit
Before the market and the metrics: does this deal belong in this fund? State the thesis criteria it matches, the criteria it misses, and anything that touches the anti-thesis. A great company that does not fit the mandate is still a pass, and writing that down keeps the portfolio coherent across years of temptation.
3. Market
Size the opportunity honestly, and separate the deck's claims from your own view. If the founder says the market is $40B, the memo should say whether you believe it, and on what basis. Cite external evidence where it exists and say "unverifiable" where it does not. The committee should never have to guess which numbers were checked.
4. Team and execution
Specific signals, not adjectives. What has this team shipped, sold, or survived that predicts execution here? What did references and your own calls actually surface? If a concern came up in a conversation, it belongs in the memo with its provenance, not in the hallway afterward.
5. Traction and terms
The numbers, with periods attached. Revenue without a time period is decoration. State what is real (contracted, recognized), what is forward-looking (pipeline, LOIs), and what the round terms imply at entry. "Not disclosed" is a finding, and it should lower confidence rather than disappear.
6. Risks and mitigants
The section committees actually reread. List the ways this deal fails, ranked, each with a mitigant or an honest "none." Include risks raised in calls and emails, not just risks visible in the deck. A memo with no serious risks listed is not a strong deal, it is a weak memo.
7. Return case
Entry ownership, realistic exit scenarios, MOIC and IRR under each. One base case with a hockey stick is not a return analysis. Low, base, high, with the assumptions that drive each, so the committee can argue about the assumptions instead of the arithmetic.
The standards that make it defensible
Same structure, every deal. The value of a memo compounds only if deal #40 is comparable to deal #4. When each partner has a personal format, the archive is a pile of essays, not a record.
Claims get verdicts. External claims (market size, growth rates, competitor stats) are checkable. Check them, and record supported, contradicted, or unverifiable. Internal metrics can't be verified externally, so critique them instead: consistency, plausibility, gaps.
Evidence includes conversations. The most decision-relevant information is often verbal: a founder's answer on churn, a customer's hesitation on a reference call. If it changed your view, it goes in the memo with a source and a date.
Versions are kept. When new information arrives, write a new version next to the old one. The delta between memo v1 and v3 is some of the best diligence reading a committee ever gets.
The vote attaches to the memo. Record who voted, on which version, with dissents. Two years later, "why did we pass?" should have an answer on file, not a debate about who remembers what.
The uncomfortable math
A memo at this standard takes an analyst most of a week per deal. That is why memo quality collapses exactly when it matters most, in the weeks when deal flow spikes and the committee is busiest.
That gap is what we built SamCIO for: Sam drafts the memo at this standard, from the full record (deck, model, data room, your team's emails and call transcripts), through your firm's weighted framework, in about a minute. You edit, the committee decides, and every version and vote stays on the record. The structure above is not just advice we like; it is the structure Sam writes in.
But the structure works with or without the AI. Write the recommendation first, attach periods to numbers, give risks their own ranked section, and keep the versions. Your future IC will thank you, and so will your LPs when they ask how you decide.