Selling is a big blind spot for investors. For every thousand articles about when to buy a company, we’re lucky to find one about when to sell. Yet selling has a huge impact on our long-term performance. And the evidence is clear: most people are no good at it. But this weakness also provides the opportunity for an enduring competitive advantage.  

The Problem

It’s well known that a lot of individual investors suffer from selling biases like loss aversion. Recent research shows that the professionals struggle too. An impressive study found that professional fund managers, although pretty good at buying, are terrible at selling. It was a landmark study, so let’s unpack the findings.

“We document a striking pattern: while the investors display clear skill in buying, their selling decisions under-perform substantially…selling decisions not only fail to beat a no-skill strategy of selling another randomly chosen asset from the portfolio, they consistently under-perform it by substantial amounts.”

The sell decisions of professional investors are so bad that they performed worse than chance. That’s rough. But the buy decisions did add value, so it’s not that fund managers have no idea about investing. Instead the authors found a flawed process where selling doesn’t receive adequate attention:

“We present evidence consistent with the discrepancy in performance between buy and sell decisions being driven by an asymmetric allocation of cognitive resources, particularly attention…We conjecture that PMs in our sample focus primarily on finding the next great idea to add to their portfolio and view selling largely as a way to raise cash for purchases”

“PMs in our sample have substantially greater propensities to sell positions with extreme returns: both the worst and best performing assets in the portfolio are sold at rates more than 50% percent higher than assets that just under or over performed. Importantly, no such pattern is found on the buying side – unlike with selling, buying behavior correlates little with past returns and other observables”

Most professionals have no good process for selling. They focus on buying, and only think about selling when they are fumbling around to free up cash. When they do sell, it is with little research. So they sell whatever sticks out the most: the biggest gainer or the biggest loser. It’s not logical, and they don’t think about the future prospects. The result is decisions that are so bad, they would have been better off throwing darts at a board to pick a position to sell.

This focus on past returns is also mirrored in much of the folk wisdom that floats around about selling:

  • ‘Sell when a stock doubles (or is up 20%, or 30%, or whatever arbitrary number)’
  • ‘Never sell at a loss’
  • ‘Nobody ever went broke taking a profit’
  • ‘Sell anything that falls by 10%’
  • ‘Water your flowers and trim your weeds (sell companies whose share price has fallen)

Intelligent investing is supposed to be forward-looking. It’s the future that counts. Yet all these rules of thumb, and the trading of the average fund manager, is based on past price movements.

The core problem is that most people’s sell process abandons all that is good about their buy process:


It’s a big problem. But there is a better way. My style of fundamental growth investing, led me to adopt two core selling principles:

Principle #1: Sell quickly when a thesis is broken

This principle is key to generating high annualised returns by 'winning big, and losing small'. Selling quickly when a thesis is broken allows you to ‘lose small’ when you have made a mistake. More importantly, it allows you to quickly re-deploy your precious capital in to a new high-conviction idea.

Like running a marathon, it is simple, but not easy.

There are five steps:

  1. Identify your thesis (actually write it out)
  2. In that initial thesis, identify what would cause you to sell
  3. Continuously monitor the company, its competitors, customers etc.
  4. Update your valuation estimate and avoid thesis creep
  5. Sell quickly if the thesis is broken

Let’s work through those steps with a real-world example: Class (ASX:CL1). Class provides a SaaS software product that helps Self-Managed Super Funds (SMSFs) manage their accounts. It’s a sizable industry in Australia, with over 500,000 SMSFs. It’s a growing market: each year more people’s retirement balances hit a size where it becomes worthwhile to consider managing it themselves.

Step 1: Identify your thesis (actually write it out)

Few investors clarify their thinking on why precisely they are buying a company. Even fewer take the time to actually write down that investment thesis. If you do, you will be ahead of the pack.

We must have a clear idea of what our logic is in the first place, so that we can know when that thesis is broken.

We first bought Class shares shortly after the IPO in early 2016. It was a simple thesis. Class was disrupting traditional desktop software – an inevitable shift to the cloud was underway. At the time of purchase, Class was dominating the new cloud-based market, and winning over two thirds of new cloud customers.

Even better, Class’ major competitor, the incumbent BGL Super, had stumbled with their first launch of a cloud product. We initiated a position and over the next two years Class’ share price rose over 80%.

It looked like the ideal investment, a scalable software business that was dominating a sticky niche, while a sluggish incumbent failed to adapt.

But that would change.

Step 2: Identify what new evidence would cause you to sell

The moment before you buy a stock is the last time you will be thinking objectively. It is crucial that you use this moment to write down precisely what new evidence would cause you to sell in future.

We identified multiple risks that could have befallen Class. The government could have changed the rules around SMSFs. A major security breach could have broken client’s trust in Class’ cloud service. Neither of those came to pass.

Another risk was that a competitor could somehow crack the market and start stealing share. Class would not be so lucky on this count.

Step 3: Continuous monitoring

Eternal vigilance is the price of superior returns.

We must continuously monitor for thesis-breaking evidence. That means keeping tabs on the company, its staff morale, its new products, customers, competitors, regulators, suppliers, etc. There are many tools that can help this along: Glassdoor, Google Trends, product forums, Google Alerts, investing forums. But that should be just the beginning. Superior returns require superior portfolio monitoring.

In the example of Class there was one obvious source of intelligence that most of the market somehow missed. Remember that big incumbent BGL? Well BGL would regularly release announcements about how their new cloud based product was progressing.

For a long time these press releases were mostly hot air. Every company claims that their products are market-leading, next-generation, cutting-edge. But in early 2017 it became apparent that BGL’s new cloud products were gaining traction. Talking to BGL’s customers, it seemed that the incumbent may have started to get its act together.

No company goes without serious competition forever though. So we were careful to avoid a knee-jerk response.

That all changed in October 2017.

Step 4: Re-evaluate and avoid thesis creep

Thesis creep is one of the great traps that ensnare investors, particularly value investors. The company reports some bad news, and rather than recognise the mistake and sell, the investor holds on.

The share price has usually fallen by this stage, which can allow the original thesis to sneakily creep its way to something new:

“Sure, we originally thought the company would do XYZ, and it clearly hasn’t, but it’s just so darn cheap now, we couldn’t sell at this price”.

There were two new pieces of information that were released on the 5th of October. First, Class’ reported its latest quarterly update. It showed that the company’s net new account additions had fallen, dramatically.

In the comparable quarter a year earlier Class had added 11,880 new SMSF accounts. The same number in 2017 showed just 6,232 new accounts. A fall of 47%. And this was after a soft June quarter, which the company had guided would quickly rebound. To make matters worse, the chart the company usually reported which would have shown this fall clearly was no longer included.

Something had changed.

Later that day it was confirmed. BGL announced that it had now surpassed 100,000 accounts on its own cloud-based product. Worse still (for Class), their biggest competitor had added 23,402 accounts during the latest quarter.

When we first purchased shares, Class was winning approximately 66% of new cloud accounts. Now it appeared to be winning just 20%. That’s a huge swing in competitive position. We updated our intrinsic value estimate with the new information. Lower growth and higher acquisition costs meant the shares were significantly overvalued.

It was time to face a tough truth.

Step 5: Sell quickly when the thesis is broken.  

Class had been a star of our portfolio. I had even interviewed the CEO in a fireside chat at a client event. It wasn’t easy to reverse course and admit that we were wrong. But when a thesis is broken, we must be decisive.

We reached our sell decision on the same day the news broke. Although the shares were already down slightly, it would take many months for the market to fully absorb the new competitive paradigm. We were able to exit our position for a 67% gain.

It worked out well. Today, almost 18 months later, the shares are now over -50% below where we sold.

Sell quickly when a thesis is broken.

Principle #2: Sell if you would not be buying today

Holding is an active decision, not a passive one. It just doesn’t feel like it.

Each day the market offers us the opportunity to buy or sell our shares. Every day that we hold a position we are effectively choosing to ‘re-purchase’ it at today’s prices, and in today’s position size. If we don’t think that the current position size is the best possible allocation of our precious capital, we should sell.

This principle – to sell if you would not be buying today – includes those situations mentioned in the first principle. But it also adds the hard edge of a valuation-based sell. If the share price has risen so much that you would not be buying the shares today, it is time to sell. No business is so great that its share price can’t rise high enough to render it an unattractive investment.

It sounds simple, but again the execution takes work. It requires maintaining an accurate estimate of the company’s intrinsic value, and being willing to trim the position, or even sell out entirely, when share prices rises too far above intrinsic value.

My personal approach to investing in high growth businesses is to sell slowly when the motivation is purely based on valuation. This is to reflect the ability of truly superior businesses to consistently outperform even the most optimistic estimates. It is both an art and a science, but ultimately we must be disciplined: sell if you would not be buying today.


Selling is a big blind spot for most investors. But that weakness means we have a huge opportunity to improve. If you adopt a sound selling process, based on the future and not the past, you will gain a massive competitive advantage over other investors. 

Simon James

Thanks for this, this is easily the hardest part I find in investing, particularly with a buy and hold mentality being drummed into most investors from the core books. I am now much more focused on having my own method of valuation and using that to do both the purchase and to decide when to sell. I also give it a margin of safety, so if things are going great, I might allow X% extra FOMO upside onto my sell valuation. So when it shoots past this, I take your point about "would I buy right now with greed" and "is the current and X% of future performance already baked into the price"...

Matt Joass

Thanks Simon, great points. I also think it is a great idea about setting a margin of safety on the upside as well. It depends on the style of investing you are undertaking. But for fast-growing businesses with a lot of optionality, fundamentals can often surprise to the upside. We don't want to be shaken out in these cases, while also maintaining valuation discipline. Trimming a position can also be a prudent approach to position size management. We don't need to be 'all or nothing' with our sells any more than we do with our buys.

Mark Dawson

I also struggle to find the best time to sell. I used to go off percentages earned but now I put in more effort. Quiet often I check to see when a company is going to release it's earning results. If I believe that they're going to beat market expectations then I hold onto the stock and may sell the other side of the announcement. However on the flip side if I think they're going to release a raft of bad news then I'll immediately sell to lock in any profit. Another good option is to sell in the lead up to dividend payments. As long as your not chasing dividends then sometimes you can make a few dollars as the stock price ramps just before going ex-dividend. Another method I use is to move my stops up as the share price increases. That way I prevent any loss and keep trying to creep my profits. Always hard to know how much of a percentage below to place your stops. It generally depends on the type of stock and price range movements.

Nathan Thomas

Brilliant article. A great reminder. I am absolutely terrible in selling my duds on the basis they’re much cheaper now.

Matt Joass

Interesting approaches Mark. The key for each investor is to find a considered selling process that suits their overall style. "Another method I use is to move my stops up as the share price increases. That way I prevent any loss and keep trying to creep my profits." - John Dorfman of Thunderstorm Capital outlines a similar approach to this in the book 'The Art of Value Investing'.

Matt Joass

Thanks Nathan! It's a challenge we all face. The worst is when investors deploy more capital in to a sinking thesis: investing quicksand. It's important to celebrate a smart sell at least as much as a brilliant buy.

Mr T

a great article, thanks. (love to see a series of articles from various people on 'how we make the sell decision' james...). there is no question the selling decision is poor from pro's and punters alike, and there is remarkably little academic/professional work done on this topic. i looked at the original kelly paper to see if it could be used as the basis for making smarter selling decisions - am still waiting for someone to realise that publishing on the 'reverse kelly' would be a great finance phd that would make them a shoe-in to work in the industry!! anyway, after looking at this topic a fair bit, it seems to me that there are a few 'axiomatic' reasons to sell (eg thesis rupture). But i am not aware of any convincing reasoning on when/how much to sell on valuation grounds (pro's and punters alike all are really just selling on opinion, either by 'adjusting stop losses' (but where and how and WHY to set the stop loss?), or by selling when 'the valuation hit our trigger in the model' (but come on - we can all make a model say whatever we want, and the model is essentially just a DCF or valuation multiple etc opinion on when to sell). no 'answer' from me, but this topic is a huge hole in modern portfolio mgt theory imho.

Matt Joass

Thanks Mr T (great username). Great thoughts, and I agree that there is a big gap for further research. Regarding the search for some rules about the relationship between intrinsic value and an appropriate sell limit, my take is that it should be heavily influenced by investing style. If your approach is to buy deeply undervalued cigar butts then all your returns will come from a one-time closing of the gap to intrinsic value. That deep value strategy should sell as soon as the intrinsic value is hit, or even when the discount narrows significantly. On the other hand if you are buying undervalued growth businesses, it is very likely that the intrinsic valuation itself will increase significantly over time (if you are making accurate selections), in which case my approach is to be slower to sell purely based on valuation.

FC Choong

One of the best articles I have read so far. So can easily find articles on buying but much less for selling.

Eddie Jones

Hi Matt, Maybe I am having foggy day. But would appreciate a plain English interpretation of the attached jargon from your article. Cheers EJ decisions being driven by an asymmetric allocation of cognitive resources,

Matt Joass

Thanks FC Choong!

Trevor Ridgway

Hmmm ! WHAT if "one" is merely HOLDING to receive dividends AS INCOME and the markets are see-sawing from day to day ? Surely you don't advocate selling and re-buying on a continuous basis ? Sometimes a HOLD is the only option !

Matt Joass

Hi Eddie, thanks for the question. "decisions being driven by an asymmetric allocation of cognitive resources" This is essentially saying that buy decisions are given a lot of thought by fund managers and that helps them out-perform. However sell decisions are rushed, and not given enough thought i.e. they are not allocated enough 'cognitive resources'. This leads the sell decisions to under-perform.

Dylan Jobson

Fantastic article Matt . Much appreciated. I’m still learning to adapt both my buying -> Thesis and selling processes. Admittedly I’ve gotten some really good luck of set by some horrendous capital explosions.

Matt Joass

Hi Trevor, thanks for the question. The question 'if I didn't already own this company, would I be buying it today' is a great clarifier. We don't want to over-trade around our estimate of intrinsic value due to tax and transaction costs. There should be some middle zone there. But excluding taxation impacts and transaction costs, I haven't heard a great example of why you should hold an asset at a certain price if you shouldn't buy it at that same price.

Matt Joass

Thanks Dylan!

David Morris

Finally someone writing about SELLING. I have been investing in the market for only 6yrs & it has bothered me that there's very little on the SELL subject. Been thinking about it since I started, basically I use 1. What's is the targeted Return on ones capital. 2. As / If the target is achieved & still going up, I have a method that signals SELL...via Trend lines or Sell Stops whichever works on that particular STOCK. Once again a HUGH THANKS , on covering a subject that most Professionals wont touch.

Matt Joass

Thanks David! Great to hear you enjoyed the article.

John Hyde

Does that mean you would sell CSL if you had it today?

Matt Joass

Hi John, thanks for your question. I don't personally own CSL today so just like any company I don't currently own, I would sell it if I did somehow have it today.