When to Sell a Stock: Five Sell Rules to Write Before You Buy

Why selling is harder than buying

Most investors have a process for buying: screens, reading the 10-K, a valuation. Few have one for selling, so sell decisions get made under pressure, after a sharp move, when judgment is at its worst. Behavioral finance has a name for one common result, the disposition effect: the tendency to sell winners quickly to lock in a gain and hold losers in the hope of getting back to even.

Sell discipline is the fix: a short set of rules, written at purchase, that say which evidence would make you sell and which would not. It builds on a written thesis with kill criteria, covered in the investment thesis guide. This article goes further into each sell trigger, what to ignore, and how to size a partial exit.

When to sell a stock: five rules that hold up

1. The thesis is broken

A kill criterion you wrote before buying has been met: revenue growth below your floor for two quarters, a margin target withdrawn, a balance-sheet limit breached. Act on it whatever the price has done. If the stock is up, the gain came from something other than your reasoning, and luck is not a plan for the next year.

2. The price is above a reasonable value range

Estimate value as a range, not a point, and update it when results change. When the price moves above the top of that range, ask the only question that matters: would I buy this today, at this price? A reverse DCF shows the growth the new price requires; if that requires results you consider unlikely, the upside you bought may be gone. The priced-for-perfection guide covers the warning signs.

3. A better opportunity needs the money

Switching is justified when a new idea offers a clearly higher expected return for similar risk, judged on the same terms as the holding: a written thesis, a value range and a downside case. Set the bar well above “slightly better”, because every switch has costs and the new idea has not yet been tested by a quarterly report.

4. The position has grown too large

Set a maximum portfolio weight at purchase. When a winner grows past it, one company’s bad quarter can do outsized damage to the whole portfolio. Trimming back to the maximum is risk control, not a view that the stock will fall.

5. You find a mistake in your original analysis

Being wrong about the future is normal. Being wrong about a fact you had at the time is different: you treated a one-off gain as recurring, missed a debt maturity, or misread how the company earns revenue. Rebuild the thesis with the corrected facts, at today’s price. If it still stands, keep it; if it does not, sell, even if the stock has not moved yet.

Selling rules: what should not trigger a sale

Good selling rules also list what to ignore. None of the following, alone, is evidence about the business:

• A price drop. Price is noisy over weeks and months. If every condition in your thesis still holds, a lower price means the same business costs less.
• Getting back to even. Your purchase price is irrelevant to what the stock is worth now.
• Headlines and analyst target changes that do not touch one of your conditions.
• One soft quarter that misses consensus but stays within your conditions.
• Macro forecasts about rates or recessions, unless your thesis depends on one explicitly.
• Boredom. A thesis that takes three years to play out will look idle for most of them.

One caution: a sharp drop without obvious news sometimes means someone knows something you do not. Before calling it noise, check for a new filing, a guidance change or a sector event; the news feed explains why individual stocks moved.

A sell discipline decision table

Use this to match what happened with the evidence to check and a typical rule-based response. The responses are examples of rules an investor might write, not recommendations.

Trigger	Evidence to check	Typical rule	Common trap
Kill criterion met	The 10-Q or 10-K figure behind it	Sell in full	Rewriting the criterion after the fact
A condition slipped once	Cause: timing, one-off or trend	Watch, with a deadline	Waiting indefinitely
Price above top of value range	Updated range; growth the price implies	Trim toward a smaller weight	Raising the range to justify holding
Price far above range	Same, plus what the market may know	Sell in full	Anchoring on the recent peak
Position above maximum weight	Current portfolio weight	Trim to the maximum	Letting a winner become half the portfolio
Better opportunity	Both theses on equal terms	Switch only on a wide margin	Comparing a tested holding with an untested idea
Error in original analysis	Thesis rebuilt with corrected facts	Hold if it survives, sell if not	Defending the first conclusion
Price fell, conditions intact	News, filings, sector events	Re-check; no sale	Selling to stop the discomfort

Trimming vs selling

Trimming and selling answer different questions. Trimming says: the thesis still holds, but this position carries more risk than I planned, because it is too large or priced near the top of my range. Selling says: the reason to own this is gone.

• Trim when the position exceeds its maximum weight, or when the price sits above your value range by a modest margin you defined in advance.
• Sell in full when a kill criterion fires, when an analysis error removes the thesis, or when the price is so far above your range that no reasonable scenario supports it.
• Stage the exit if you genuinely cannot decide: sell a fixed fraction now and set a dated review for the rest. Write down the date, or staging becomes a way of never deciding.

Watch for one pattern: trimming a stock whose thesis has broken. It feels prudent, but it keeps money in an idea you no longer believe.

Worked example: one holding, three sell decisions

Illustrative only. Quillmoor Instruments is a fictional company, and every figure is a round, made-up number chosen to show the method, not a real security or a recommendation.

An investor with a $100,000 portfolio buys 250 shares at $40, a $10,000 position. The written rules: value range $55–$65; maximum weight 15%; trim to the maximum if the price exceeds the top of the range, exit in full if it exceeds it by more than 20%; kill criterion, gross margin below 40% for two consecutive quarters. For simplicity, the rest of the portfolio stays flat.

Checkpoint	Price	Position value	Portfolio weight	What the rules say
Purchase	$40	$10,000	10.0%	Buy; value range $55–$65
Month 4: price down 25%, conditions met	$30	$7,500	7.7%	Re-check the evidence; no sale
Month 20: above value range, overweight	$90	$22,500	20.0%	Trim 63 shares (about $5,670) back to 15.0%
Month 26: kill criterion triggered	$84	$15,708	n/a	Sell the remaining 187 shares

Month 4. The shares fall 25% on a market sell-off. Gross margin is 44% and every condition holds, so the rules call for a re-check, not a sale.

Month 20. After two strong years the investor raises the value range to $70–$85, but the price is $90, 5.9% above the top: enough to trim, not enough to exit. The position is now 20.0% of a $112,500 portfolio, so the investor sells 63 shares and keeps 187, a 15.0% weight.

Month 26. Gross margin prints 39% and then 38%. The kill criterion fires while the stock still trades at $84, 110% above cost. The rules say sell the remaining 187 shares, and the gain does not change that.

Taxes, costs and keeping a record

Selling at a gain may create a tax bill, and the rules differ by country, account type and holding period. Treat tax as one cost among several, alongside trading costs and the time a replacement idea needs to prove itself. It can reasonably change the timing or size of a sale; it should rarely keep you in a stock whose thesis is broken. For your own situation, ask a qualified tax professional.

Record every sale, trim and deliberate hold: date, price, the rule that triggered it and what you expected. An investment journal turns those records into feedback on your sell discipline. To avoid checking every condition by hand, VALYSTR’s Thesis Guardian monitors the conditions you write for each thesis, and the investment thesis template helps you set the conditions in the first place.