A memo can run to fifteen pages and still be decided on four of them - the thesis, the bear case, the return math and the terms. Everything else is background that makes the reader feel informed without changing the answer.
11 min read · Level 4
There is a natural instinct to make a memo comprehensive - to show that every angle was considered, every risk category was named, every comparable company was reviewed - and comprehensiveness is frequently mistaken for rigor. It is closer to the opposite. A memo that gives equal space to fifteen sections has made a decision, whether or not the writer intended to, that all fifteen matter equally to the outcome, and that decision is almost never true. The four sections that actually carry weight are uncomfortable to write precisely because they require commitment: a one-sentence thesis that could be wrong, a bear case with real numbers in it, a return calculation that might come out unattractive, and terms that might not be as protective as they first appeared. It is much easier to write three good paragraphs about the market than to write one honest sentence about what has to be true - and the memo that takes the easier path will read, to almost anyone, as more thorough than the one that took the harder one.
Most memos contain a market overview, a team background, a product description, a competitive landscape and a history of the company - and almost none of a real decision is made in any of those sections. They are context. They help a reader who knows nothing about the company get oriented, and orientation is genuinely useful, but none of it is what changes a vote from no to yes.
Four sections do that work. The thesis states, in one or two sentences, the specific bet being made - not 'this is a good company' but the actual mechanism by which money is expected to be made, stated precisely enough that it could be wrong. The bear case states the mechanism by which the thesis fails, with the same specificity, and is covered in full in the next lesson. The return math turns the thesis into numbers: what is paid, for what ownership, and what that ownership is worth under a base case and a bear case. And the deal terms describe the actual structure of the investment - price, security type, and the protective provisions that determine what happens when things go wrong, which is exactly the scenario a purely narrative memo is least equipped to address.
The return math section is where vague conviction becomes a checkable number, and it is worth walking through in full because it is also where most memos quietly go soft. Suppose an investment of 2,000,000 buys 8.0% ownership at a 25,000,000 post-money valuation - 2,000,000 divided by 25,000,000 is 8.0%, so the arithmetic on entry at least is unambiguous. The base case assumes the company exits in five years at a 150,000,000 valuation, with ownership unchanged (a simplification, since further financing rounds would ordinarily dilute this - a real memo should say so explicitly rather than let the assumption pass silently). At 8.0% of 150,000,000, the position is worth 12,000,000, which is a multiple on invested capital of 12,000,000 divided by 2,000,000, or 6.0x. Over five years, that multiple corresponds to an annualised return of roughly 43.1% - the fifth root of 6.0 minus one.
The bear case runs the same arithmetic against a 40,000,000 exit: 8.0% of 40,000,000 is 3,200,000, a multiple of 1.6x on the same 2,000,000, or roughly a 9.9% annualised return over five years. Two numbers, computed the same way, tell you something a paragraph of enthusiasm cannot: the range of outcomes the memo is actually underwriting, and how much of the case rests on hitting the higher figure rather than the lower one.
The deal terms section then asks what protects that return if the company underperforms even the bear case. A 1x non-participating liquidation preference on the preferred stock means the position gets its money back before common shareholders see anything, in a sale below the value at which the preferred converted - a real, if partial, floor under the 1.6x scenario. Pro-rata rights preserve the option to maintain the 8.0% stake in future rounds rather than being diluted out of it. None of that is decoration; it is the difference between a return calculation that assumes nothing goes wrong and one that accounts for what happens when something does.
Because a reader - and a writer - can only hold a few numbers in their head at once, and a fifteen-page memo that treats every section as equally important guarantees that the four sections which actually decide the vote get the same attention as the two paragraphs on the competitive landscape that nobody will remember by the following week.
And because the sections that are missing when a deal goes wrong are almost always the same four. A company that fails rarely fails because the market overview was inaccurate; it fails because the bear case never named the actual risk, the return math never showed how thin the margin for error was, or the deal terms provided no protection at all when the downside case arrived. Concentrating attention on the four sections that decide the vote is not a stylistic preference - it is where the analysis that would have caught the problem was supposed to live.
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 completed return-math table for the logistics company from the previous lesson, entered at 2,000,000 for 8.0% at a 25,000,000 post-money valuation.
Base case: exit at 150,000,000 in year 5 Proceeds = 0.08 x 150,000,000 = 12,000,000 MOIC = 12,000,000 / 2,000,000 = 6.0x IRR = 6.0^(1/5) - 1 = approximately 43.1%
Bear case: exit at 40,000,000 in year 5 Proceeds = 0.08 x 40,000,000 = 3,200,000 MOIC = 3,200,000 / 2,000,000 = 1.6x IRR = 1.6^(1/5) - 1 = approximately 9.9%
Deal terms: preferred stock, 1x non-participating liquidation preference, pro-rata rights on subsequent rounds, priced at the 25,000,000 post-money above.
The ask, stated plainly at the end of the memo: approve up to 2,000,000 on these terms, with a board observer seat rather than a full board seat given the size of the position relative to the round.
Notice what the table does that a paragraph could not. It shows that the base case requires a 6.0x outcome and the bear case still returns a positive, if modest, 9.9% annualised - which means the position does not require the base case to be roughly acceptable, only to avoid an outcome worse than 40,000,000. That is a genuinely different risk profile from a deal whose bear case is a total loss, and the only way to see the difference is to have actually run both numbers rather than described the company as 'attractive at this valuation.'
The specific wrong move is spending the most words on the sections that carry the least decision weight - three pages on the founding story and the market landscape, four sentences on the return math, because the founding story is easy and pleasant to write and the return math requires committing to numbers that could turn out to be wrong.
The second is running only a base case. A single number dressed up as a forecast looks precise and hides the actual range of outcomes being underwritten; the bear case is not pessimism, it is the other half of the same arithmetic, and a memo that shows only the 6.0x without the 1.6x has quietly decided the reader does not need to see how the deal performs if things go worse than expected.
The third is treating deal terms as an appendix rather than a section that carries weight. A generous liquidation preference or a lack of pro-rata rights changes the actual return distribution as much as a change in the exit multiple does, and burying it in a term sheet attachment rather than in the memo's own return math means the committee is voting on numbers that do not reflect what they are actually approving.
Not 'a strong team in a large market' but the actual mechanism: what the company must do, and what must be true in the world, for the investment to work. If the sentence survives being read back with 'we believe' removed from the front of it, it is specific enough.
Ownership times exit value divided by capital invested, for both scenarios, using the same entry price and the same ownership percentage. State the annualised return alongside the multiple; a 6.0x that takes five years reads differently once it is also a 43.1% IRR.
Unchanged ownership across five years almost never happens - dilution from future rounds is close to certain. Say so in the memo rather than letting a convenient assumption pass as fact, and note in one line what a more realistic dilution path would do to both multiples.
A liquidation preference changes what happens in the bear case specifically; state what it is worth in that scenario rather than describing it only in legal language a reader has to translate themselves.
The market overview, team background and competitive landscape earn a paragraph each, not a page - enough for a reader with no context to orient, and no more. If cutting them feels uncomfortable, that discomfort is worth noticing: it usually means those sections are doing work the four that matter should be doing instead.
For any memo above the size where a one-line thesis statement genuinely captures the decision - which is most private investments large enough to need a written case at all.
A very small position, sized so that even a total loss is immaterial to the portfolio, does not need a full return-math table with base and bear cases - a shorter note stating the thesis and the size rationale is proportionate, and building the full structure every time trains people to skip it when it actually matters.
A shape that recurs across private investment memos of considerable length is an extensive treatment of the market, the team and the competitive set, followed by a brief paragraph asserting the deal is 'attractively priced' with no supporting arithmetic. The length of the document creates an impression of thoroughness that the actual decision-relevant content does not support, and a committee reading quickly is more likely to be reassured by volume than to notice that the one section which would let them check the claim was skipped.
Judge a memo's thoroughness by the depth of its four load-bearing sections, not by its page count. A short memo with a real base case, bear case, return table and terms section has done more work than a long one without them.
Create a free account to read the rest and apply it to your own investing.