Two versions of the same memo can describe the same deal and do opposite jobs - one is written to get to yes, the other is written to find out whether yes is correct.
9 min read · Level 4
The received wisdom is that a good memo is a well-argued one, and that framing quietly trains people to become better lawyers for their own conclusions rather than better investors. A well-argued case for a bad investment is a more dangerous document than a poorly-argued case for the same investment, because it survives scrutiny that should have killed it. The uncomfortable implication is that the quality of a memo should be judged partly by how much it argues against itself - a memo that spends real space on the bear case and on what has not been checked is doing more of the actual work than one that reads more persuasively, even though it will look, superficially, like the weaker document to an audience trained to reward confidence. Writing this way over years produces a specific and unglamorous track record: a folder full of memos that concluded 'pass,' most of which nobody but the writer ever saw, and which are collectively worth more than any single memo that got approved.
An investment memo is a written case for a specific investment, produced before the money is committed and read by people who can say no. That description sounds neutral, but the same document can be built to do two entirely different jobs, and most people who write one have never decided which job theirs is doing.
The first job is advocacy. The memo exists to move a deal through a process - past a committee, past a partner, past your own hesitation - and it is written the way a lawyer writes a brief: gather the facts that support the conclusion, present them persuasively, and manage the weak points rather than dwell on them. A memo built this way reads well. It is also, structurally, incapable of changing the writer's mind, because changing the writer's mind was never one of its jobs.
The second job is discovery. The memo exists to find out whether the investment is actually good, and it is written the way you would write it if nobody but you would ever read it - stating the claim plainly, stating what would have to be true for the claim to hold, and stating what has not yet been checked. A memo built this way sometimes concludes 'do not do this,' which is exactly the outcome the first version is structurally unable to reach.
Most real memos are a blend, and the blend usually tilts toward advocacy without anyone deciding that it should, because advocacy is what gets rewarded in the room. A memo that surfaces its own weaknesses looks less confident than one that doesn't, and confidence is mistaken for quality far more often than it should be. The discipline this subject is built around is writing the discovery version first, in private, and only then adapting it for the room - never skipping straight to the version built to persuade.
The test for which version you have written is simple. Read your own memo and ask whether it could plausibly end in 'pass.' If every sentence in it points toward yes, you have not tested anything yet - you have documented a decision you had already made before you started writing.
Because the two versions cost the same amount of effort and produce completely different amounts of protection. A memo built for advocacy takes exactly as long to write as one built for discovery, uses the same sections, and can look identical from the outside. The difference is entirely in whether the writer was willing to let the document conclude something other than yes - and that willingness is invisible to anyone reading the finished product, including, months later, the person who wrote it.
And because the market rewards the wrong version in the short run. A confident, well-argued advocacy memo gets a deal approved faster and makes the person who wrote it look more decisive. A memo that spends a page on what could be wrong reads, to an inexperienced audience, as uncertainty. The investors who are best at this over long periods are disproportionately the ones who accepted looking less certain in the room in exchange for a document that actually tested the thesis - a trade that pays out on a timescale long enough that almost nobody notices the investors who made the opposite trade until much later.
Educational, not personalised financial advice. Nothing in the Investor Academy is a recommendation to buy, sell or hold anything — what is right for you depends on your own circumstances and the rules where you live.
A memo for a logistics-software company opens with the market slide it inherited from the founder's deck: 'a $4.2 billion global market, growing 35% a year.' That sentence is true, cited from a market-research report, and it is doing no work at all - the company does not compete for $4.2 billion, it competes for a specific, addressable slice of it, and the slide's job is to make the opportunity feel large rather than to size it.
The discovery version of the same memo does the arithmetic instead of quoting the slide. The company sells to mid-market logistics operators in its target geography; there are roughly 3,800 such firms, and the company's own pricing puts an average annual contract at 82,000. That gives a bottoms-up addressable market of:
3,800 x 82,000 = 311,600,000
Which is 311,600,000 divided by 4,200,000,000, or 7.4% of the market-slide figure. Both numbers are true. The 4.2 billion describes a global category the company will never fully address; the 311.6 million describes what the company is actually competing for today, in the segment its product and pricing are built for.
Neither figure alone tells you whether to invest - a 311.6 million addressable market can still be an excellent business if the company can win a meaningful share of it, and a company that later expands its product can grow that number. But a memo that quotes 4.2 billion and stops has not sized the opportunity; it has borrowed a number that was computed for a different purpose and let it stand in for one it cannot actually answer.
The specific wrong move is writing the memo after the decision has already been made socially - after a partner has said 'I love this' in a meeting, after a term sheet has effectively been agreed in conversation, after everyone involved already knows which way this is going. At that point the memo becomes a record of a decision rather than a test of one, and no amount of a template's rigor changes what job the document is actually doing.
The second mistake is treating polish as evidence of quality. A memo with clean formatting, a confident tone and no loose threads is pleasant to read and tells you nothing about whether the thesis survives contact with a hostile reading. The correlation between how good a memo feels to read and how correct its conclusion is turns out to be weak, and weakest exactly where it matters most - at the moment before the mistake becomes expensive.
The third is borrowing numbers instead of computing them, which is what the market-slide example above actually is. A cited figure ('a $4.2 billion market') carries the authority of its source without carrying the relevance of the question being asked, and it is the single easiest way to make a memo look rigorous while it does none of the actual work.
No audience, no formatting, no slide inherited from the founder's deck. State the claim, state what would have to be true, state what you have not checked. This version is allowed to end in 'pass,' and if it never does across every deal you write one for, that is itself a finding about your process.
A number sourced from someone else's report describes their question, not yours. Build your own version from the units the company actually sells and the price it actually charges, and treat any gap between the two as something the memo has to explain rather than something to quietly drop.
If every sentence points toward yes, the document has not tested anything - it has documented a decision made before the writing started. A memo that never once reaches 'pass' across many attempts is evidence about the process, not evidence that every deal was excellent.
Once the discovery version has survived your own hostile reading, adapt it for the audience that needs to approve it. The adaptation is allowed to be more confident in tone - the underlying claims should not have changed at all in the process.
Before any private investment where you are the one deciding, or recommending a decision to others - the discipline applies at any check size, because the failure mode (writing to persuade rather than to find out) does not scale with the amount of money involved.
A one-page discovery memo is overkill for a routine follow-on into a position you already track closely and where nothing material has changed since the last review - in that case the existing monitoring record does the job a fresh memo would only repeat.
Across private investing, a shape recurs in which the memo is drafted in the days after a partner has already said, in conversation, that they want to do the deal. The document that results is technically complete - it has a thesis section, a risk section, a return calculation - and it is functionally a transcript of a decision that had already been reached socially. The tell is usually in the risk section, which lists categories rather than numbers, because naming a specific number that could kill the deal is no longer a live possibility once the room has already agreed.
Notice when a memo is being written to explain a decision rather than to test one. The giveaway is usually timing - it was started only after informal agreement already existed - and the fix is writing the discovery version before that conversation happens, not after.
Create a free account to read the rest and apply it to your own investing.