Writing an investment memo that stands up to scrutiny

An investment memo is the argument for putting money into something, written down so it can be examined. It is read by people who were not in your meetings and who will pull on the weakest thread first. A partner deciding whether to fund a round, an investment committee weighing one deal against ten others, a corporate development team sponsoring an acquisition. None of them will give you the benefit of the doubt. The memo has to carry the case on its own.

Most memos read fine until a reader stops on a number and asks where it came from. That single question decides everything. If the answer is a clean line back to a source, the rest of the document gets trust. If the answer is a shrug or a round number nobody can trace, the reader starts re-checking every other claim, and you have lost the room. Writing a memo that stands up is mostly about making sure no number ever produces that shrug.

Who reads it and what they are looking for

Understand the reader before you write a word. An investment committee is not reading for inspiration. They are reading for reasons to say no, because saying no is cheap and saying yes is where careers get made or ended. They want to know the thesis fast, see that you have done the work and find the risks already named so they do not have to dig them out themselves.

The best readers are testing one thing above all: have you been rigorous, or are you selling? A memo that sells gets discounted on sight. A memo that lays out the bull case and the bear case, with the evidence for each, gets taken seriously. You are not trying to win the argument by hiding the weak points. You are trying to show you have already thought harder about the weak points than the reader will.

The structure that works

A memo that holds up follows a predictable shape, because predictability lets the reader navigate and check. Surprise is for the contents, not the layout. Use this order:

1. 1.Thesis in one line. The whole bet in a single sentence a reader can repeat. If you cannot compress it, you do not have it yet.
2. 2.Market and why now. The size of the opportunity and the reason it is moving today and not five years ago. Why now is the part most memos skip and most committees probe.
3. 3.The business and the model. What the company does, how it makes money, who pays and what the customer actually buys.
4. 4.Unit economics. What it costs to acquire a customer, what that customer is worth and whether the gap between the two is real and durable.
5. 5.Traction and evidence. The proof the thesis is working, with the underlying numbers, not the headline ones.
6. 6.Risks named plainly. The things that could break the case, stated in your own words before someone states them for you.
7. 7.Returns and sensitivities. The outcome under a base case, an upside and a downside, never a single number.
8. 8.The ask and use of funds. How much, at what terms and exactly where the money goes.

Each section should answer the obvious follow-up question inside the section, not push it to an appendix. If you claim a market is large, the next line should show how you sized it. If you claim strong retention, the next line should show the cohort behind it.

What separates a credible memo from a flimsy one

Four habits do most of the work. They are not glamorous and that is the point.

• Every claim is tied to a source. A number with no provenance is an opinion wearing a suit. Whether it comes from the data room, the model or a named external study, the reader should be able to follow it back.
• Sensitivities are shown, not a single point estimate. One number is a guess pretending to be a fact. A range with the assumptions behind it tells the reader you understand what drives the outcome.
• Risks are surfaced, not buried. The risk section should contain the things you are genuinely worried about. A reader who finds a real risk you left out assumes there are more you also left out.
• The market is built bottom up, not top down. Counting real customers, real prices and a realistic capture beats taking a giant industry figure and assuming you win a slice of it.

The test for any figure in a memo is simple: if a reader points at it and asks how you got it, can you answer in one sentence with a real source? If not, that figure is a liability, not an asset.

The failure modes that sink memos

The same mistakes recur across thousands of memos. Knowing them lets you delete them before a reader finds them.

• Unsupported TAM. A market size lifted from a top-down report with no path showing how this company captures any of it. Experienced readers ignore the number and quietly downgrade the whole memo.
• A hockey-stick with no basis. A revenue curve that bends sharply upward with no mechanism explaining the bend. If the growth has no named driver, it is a wish.
• Cherry-picked metrics. Showing the months that look good and omitting the ones that do not, or quoting the metric that flatters while the relevant one is worse. Readers notice what is missing.
• Vanity numbers. Registered users instead of paying users, downloads instead of retention, pipeline instead of closed revenue. Big numbers that do not connect to value are a tell that the real numbers are smaller.

Each of these is a shortcut, and a memo is the wrong place for shortcuts because the reader's entire job is to find them. The fix is the same in every case: use the number that survives a hard question, even when it is smaller and less exciting than the one you wish you could use.

Where Nexlyr AI fits

The hard part of a memo is not the prose. It is making sure the figures are right, traceable and stress-tested before anyone else sees them. That is the work Nexlyr AI is built for. You give it your model and your data room, and it reads the data and computes on it directly, writing and running real queries against your actual numbers rather than estimating. So the revenue, the unit economics and the market figures come out of your source, not out of a guess.

It will not invent a number. If the data room does not support a figure, Nexlyr AI does not show it. That single behaviour kills the most common cause of a memo falling apart, because the thing a reader is hunting for, the unsupported figure, is removed before it can be written. This is what operational trust means in practice: you can rely on the numbers because they were drawn from your own data, not generated to look good.

On sensitivities, it runs the what-ifs for you. Flex an assumption, a churn rate, a price, a conversion step and see the outcome move, so the memo carries a base, an upside and a downside built on the model instead of a single optimistic line. Where a standard analytical framework fits the material, it uses one, and where it does not, it leaves it out rather than forcing it.

Then a post-build pass reviews the finished memo the way a sceptical reader would. It pressure-tests the figures, raises the questions a committee is likely to ask and flags the risks the draft glossed over, so you can answer them in the memo rather than in the meeting. The output is a fully editable deck you take in and present, so you keep control of the final word while the figures underneath it hold up.

A short checklist before you send

Before the memo leaves your hands, walk it once as the reader will:

1. 1.Pick five numbers at random and trace each to a source in one sentence.
2. 2.Find the single most aggressive assumption and check it has a downside case beside it.
3. 3.Read the risk section and ask whether the thing that actually keeps you up at night is in there.
4. 4.Replace any vanity metric with the one that ties to revenue or retention.
5. 5.Confirm the market figure is built from customers and prices, not from a slice of a giant total.

If the memo passes that walk, it will mostly pass the committee's. The goal is not a document nobody can question. It is a document that has good answers ready for every question worth asking.

Keep exploring

• Market-entry assessment: which frameworks, in what order
• How to run a quarterly business review that changes what people do next
• Investment memo