Thesis Drift: How the Reasons You Bought a Stock Quietly Disappear

Ask most investors why they own a stock and you get a clean answer. "I own it because their margins keep expanding." "Because the government revenue is sticky." "Because the AI capex cycle has years left to run." Good answers. Real reasons.

Now ask them the harder question: is that still true?

Most people go quiet. Not because they were wrong to buy. Because somewhere between the purchase and today, the reason quietly stopped being checked. The position is still in the account. The case behind it is gone. That gap has a name, and it is the single most expensive habit in self-directed investing.

What thesis drift actually is

Thesis drift is the slow erosion of the reasons you bought a stock. You buy for a specific case. Over the following quarters, the facts move. A contract gets cancelled. A margin trend reverses. A moat narrows because a competitor caught up. None of it shows up as a flashing red light in your brokerage app. The price might even be up. So you keep holding a position whose original justification no longer exists, and you call it conviction.

It is not conviction. Conviction is holding a reason through volatility. Thesis drift is holding a stock after the reason left.

Why it is so easy to miss

The problem is structural. Your brokerage shows you price, day change, and percent allocation. It does not show you why you own anything, because it never asked. Your reasons live in your head, or in a note you wrote eighteen months ago and never reopened.

So the information that would tell you the thesis broke (a 10-Q, an 8-K, an earnings call, a contract loss) arrives through a completely different channel than the one you watch. You watch the price. The evidence is in the filing. The two only get reconciled when you sit down for a quarterly review, which is usually after the stock has already moved against you.

A concrete example

Say you own Palantir on two pillars: government revenue stays sticky, and operating margin keeps expanding. Those are good reasons. Now France ends an intelligence contract, and the UK opens a review of a large NHS deal. Both are public, both are dated, both land on the government-revenue pillar specifically. Not the margin pillar. The government one.

If you are only watching the price, you might not connect those headlines to your actual reason for owning the stock. The price could even shrug them off for a while. But your thesis just took real damage on one of its two legs, and you would not know to re-underwrite the position until it was obvious to everyone.

That is the moment thesis monitoring is built for. You can read the full breakdown of how it works on our thesis monitoring page.

How to catch drift before it costs you

You do not need software to start. You need discipline. Three habits do most of the work.

Write the pillars down before you buy. Not a paragraph. Two or three specific, falsifiable reasons. "Margins expand from X to Y." "Government segment keeps growing double digits." Specific enough that a filing could prove them wrong.

Define the break in advance. For each pillar, write the single fact that would invalidate it. This is a pre-mortem. Deciding what would change your mind while you are calm is worth ten times more than deciding it while the stock is gapping down.

Check the pillars against primary sources, not the price. When earnings or a filing lands, do not ask "did the stock go up." Ask "did this confirm or contradict my reasons." Those are different questions with different answers, and only one of them protects you.

Where this leaves you

Thesis drift is not a knowledge problem. Everyone reading this knows you should sell when the reasons change. It is an attention problem. The reasons and the evidence live in two different places, and nobody reconciles them until the quarterly review.

Closing that gap is the entire idea behind Helm. You write the reasons you own each position, and we read the filings, earnings, and news against them, then tell you, with the dated source, when one starts to break. Not the price. The reason.