How to write an investment thesis (in ten minutes, with a template)

Why the note matters more than the analysis

Memory doesn't preserve reasoning. It preserves outcomes and then reconstructs a reason that fits.

So when a stock you bought is down 40%, your memory obligingly supplies a thesis you're still holding for. When one is up 80%, it supplies a thesis you were brilliant to have. Neither is likely to be what you actually thought at the time, and both feel completely genuine.

A written thesis breaks that loop. It's the only way to answer the two questions that determine whether you improve:

• Was I right for the reason I thought I'd be right? A win from an unexpected direction is luck, not skill, and treating it as skill is how people scale up a process that doesn't work.
• Is this still the investment I made? Companies change. If the thing you bought has become a different thing, holding is a new decision, and it should be made deliberately.

Without a written record, you can't ask either question honestly. You'll get an answer — it just won't be true.

The template

Four sentences. Do not write more than a page; length is procrastination wearing a lab coat.

1. What it is.

One sentence, plain language: what the business does and how it makes money.

If you can't do this without jargon, you don't understand the company well enough to own it individually. That's not a moral failing — it's a signal to put the money in the index core instead.

2. What I believe that the price doesn't.

The actual thesis. What do you think is true that the market apparently doesn't agree with?

This is the sentence people skip, and it's the whole thing. "It's a great company" is not a thesis — great companies trade at prices that reflect their greatness. A thesis needs a gap: the market thinks growth is slowing and I think this quarter was a one-off; the market is pricing the legacy business and ignoring the new segment; the market thinks this is cyclical and I think it's structural.

If you can't name a gap, you're not investing on a view. You're buying something familiar, which is fine — just buy the index and skip the maintenance.

3. What has to happen.

The observable events that would confirm you, and roughly when.

"Margins expand back above 30% within four quarters." "The new segment gets to 20% of revenue by next year." Specific enough that in twelve months you can check yes or no without arguing with yourself.

4. What would prove me wrong.

One falsifying condition, observable, with a number in it if possible.

This is the sentence that saves you money. It's also the one people write vaguely on purpose — "if the story changes" is unfalsifiable, which is exactly why it feels comfortable. Make it something a spreadsheet could check: "two consecutive quarters of revenue growth under 10%," "the CEO leaves," "they lose the contract that's 30% of revenue."

Then add two logistics lines:

Size: what percent of the portfolio, and why that number. Review: a date, roughly a quarter out.

A worked example

Deliberately ordinary, because the point is that this is quick, not clever:

What it is. A payments processor that takes a small cut of each transaction on its network; revenue scales with volume, costs mostly don't.

What I believe that the price doesn't. The market is treating last quarter's slowdown as the new growth rate. I think it was one large customer's inventory correction, and volume normalizes in two to three quarters.

What has to happen. Transaction volume growth back above 12% year-over-year by Q2, and no further customer concentration losses.

What would prove me wrong. Two more quarters of sub-8% volume growth, or the loss of another top-five customer. Either means the slowdown is demand, not one customer.

Size: 4% — this is a single-customer-concentration risk, so it doesn't get a full position. Review: after Q2 earnings.

Notice what it isn't. No discounted cash flow model, no page of industry background, no price target. A price target is usually false precision; the falsifying condition does the work a target pretends to do, and does it honestly.

The four failure modes

Thesis drift. You bought a growth story, growth stalled, and now it's a "value play." That's not an update, it's a retrofit — and it's how people hold broken positions for years. If the reason to own something changes, that's a new decision at today's price, not a continuation.

Unfalsifiable theses. If nothing could prove you wrong, you're not making a claim, and you'll never learn anything from the outcome either way.

The thesis written afterward. Buying first and justifying later is the most common version of this, and it produces a document with all the form of a thesis and none of the value. Write it before the order, always.

No review. A thesis you never check is just a nicer-looking impulse. Put the review date somewhere you'll actually see it.

Where the thesis pays off

Two places, and neither is at purchase.

When it drops. A decline is the moment your judgment is worst and your urge to act is strongest. Having the falsifying condition written down converts an emotional question — "should I get out?" — into a factual one: did the thing I said would prove me wrong happen? Often it hasn't, and the price is just moving. Sometimes it has, and then you have permission to sell without it feeling like an admission. Either way you're checking a decision instead of making one under pressure. That's the same machinery behind when to sell a stock and the main defense against panic selling.

When you review a year of them. Ten written theses are a dataset about you. Read them together and patterns appear that no single position reveals: you're consistently right on the business and wrong on timing, you always size too big when you're excited, your losses cluster in one sector, you never actually check the falsifying conditions.

That's the feedback loop most investors never build, because they never wrote anything down.

Practice writing them where it's cheap

Theses are a skill, and the first ten are bad. They're vague, they're unfalsifiable, they're secretly written after the fact. Better that those ten cost nothing.

That's a good use of a paper account — not to see whether you can pick winners, which a small sample can't tell you, but to accumulate written theses and then grade yourself against them. Sydnical is built around that loop: every position takes a reason, and the feedback is on the decisions rather than a balance that's mostly noise at this sample size. The historical crash scenarios are the stress test — a written thesis is easy to keep in a calm market and revealing in a falling one.

FAQ

What is an investment thesis?

An investment thesis is a short written statement, made before you buy, explaining what the business does, what you believe that the current price doesn't reflect, what would confirm you're right, and what would prove you wrong. Its main purpose is accountability: it lets you distinguish later between an idea that's working and one you've unconsciously rewritten.

How long should an investment thesis be?

Four sentences, plus a target position size and a review date — under a page. Length isn't rigor. A long thesis is usually harder to falsify, which defeats the point; the value is in being specific enough to check, not thorough enough to impress.

What should an investment thesis include?

What the business does in plain language, the specific gap between your view and the market's, the observable events that would confirm you and roughly when, and one falsifying condition with a number in it. Add the intended position size and a review date so the thesis has consequences.

How do I know when my thesis is broken?

You wrote the answer at purchase — that's what the falsifying condition is for. Check it against what actually happened, not against the stock price. A secondary test: would you buy this today at today's price knowing what you now know? If not, you're holding for a reason other than the thesis.

Do I need an investment thesis for index funds?

Not in the same form. An index fund's thesis is structural — you're buying the market's long-run return at low cost, and no company-specific event invalidates it. What's worth writing down instead is your contribution plan and what you'll do during a decline, because that's where index investing actually goes wrong.

---

The ten minutes before you buy are worth more than the ten hours after. Start writing them where it costs nothing →