The Art of Selling | Kuntal Shah | Oaklane Capital Management LLP | Co-founder - Needl.ai

ELI5/TLDR

Everyone teaches you how to buy a stock. Almost nobody teaches you how to sell one, even though selling is where most of the pain and regret lives. Kuntal Shah, a three-decade Indian value investor, spends an hour confessing that his biggest, most expensive mistakes have all been the same one: selling good companies too early because they looked “expensive” to him. The talk is part framework, part therapy session, and the punchline is that the price isn’t set by your idea of fair value but by the last buyer willing to accept a lower return than you.

The Full Story

Selling is the dark continent

There is, Shah says, an ocean of advice on what to buy, when to buy, how much to buy. On the sell side there is almost nothing. The reason is structural: a buy idea appeals to everyone, but a sell idea only matters to the handful of people who already own the thing. He points to a Bloomberg statistic from before regulators forced more balance: only 1.7% of broker reports carried a “sell” rating. Analysts who wrote sell reports lost access to the company. So the silence is not an accident.

He opens with a myth. Abhimanyu, in the Mahabharata, learned how to enter the spiral battle formation while still in the womb, but never learned how to get out. He fought bravely and died inside it. The question for the investor is the same: do you decide your exit, or does the market decide it for you?

You cannot borrow someone else’s playbook

A long stretch of the talk is dedicated to demolishing hero-worship. We quote Buffett verbatim, Shah says, while ignoring that Buffett the teacher and Buffett the practitioner are two different people.

He said diversification is a hedge against ignorance, but Berkshire is one of the most diversified conglomerates you can find. He said airlines are a bad business, but he has owned all four major airlines. He said derivatives are weapons of mass destruction, but Berkshire is one of the biggest underwriters of complex derivatives in the world.

The “20-punch card” advice — make so few decisions that you treat each like one of twenty holes you can ever punch — sits next to the fact that 60% of Buffett’s own positions were held less than a year. The point is not that Buffett is a hypocrite. It is that he runs a craft answerable only to himself, with a capital base, temperament and time horizon that an ordinary shareholder does not have. Copying the words without the context is how you get ruined.

Our relationship with risk is not fixed

A key slide: our behaviour toward risk and reward changes with time. We are not consistently rational because most investment decisions are not made on fundamentals alone — there is emotion bolted onto everything. Sometimes we behave like a sheep buying the fashionable theme; sometimes like a connoisseur holding a fine business for years; sometimes like a rabbit refusing to sell at a loss because we want to break even first. And, tellingly, most investors only pay sharp attention to their portfolio under stress.

Nothing sedates the human mind like a dose of easy money.

The pain is asymmetric. Losing a dollar hurts roughly four times more than gaining one feels good. So attention floods in only once losses arrive.

The two ways selling hurts

Selling, Shah argues, is the hardest of the three decisions because it hurts whichever way it goes. Sell a winner and price keeps climbing — seller’s remorse, the regret of leaving early. Sell a loser and you are forced to admit a mistake, you crystallise a loss, and you trigger self-doubt about whether you even know what you are doing.

Even the legends get it wrong here. Stanley Druckenmiller dumped tech stocks because they looked expensive, took a $600 million loss, then re-entered near the NASDAQ peak with $6 billion and lost half in six weeks. Isaac Newton — possibly the highest IQ of his era, and an actual Master of the Mint — sold his South Sea Company stake at a tidy profit, then watched friends get richer, couldn’t stand it, borrowed money, bought back at the top, and was nearly bankrupted.

I can calculate the motions of the heavenly bodies, but not the madness of people.

Buy and hold is not buy and forget

This is the part Shah wants to puncture. “Wealth is created in the long run” is true, but the long run is a chain of short runs you have to live through. Buy-and-hold does not mean buy at any price; it does not mean wait forever for break-even.

The horror examples are about entry price, not business quality. The “Nifty Fifty” in 1970s America were genuinely strong, moated businesses — and investors who bought them at egregious starting valuations still lost 70-80% over the following decade. Wipro was once 4% of India’s entire GDP by market cap; the company did everything right afterward and the stock still took nearly two decades to break even. A wrong entry price can mean three decades of zero real return — and three decades described on a slide feels like nothing, but lived through, it feels like eternity.

Reverse the DCF

Here is the most useful mechanical tool in the talk. In listed markets the price is handed to you every day. So instead of running a discounted-cash-flow model forward to estimate value, run it backward: take the current price and solve for the growth rate the market is already assuming. Then ask whether that assumption is sane.

At the peak, Cochin Shipyard’s valuation was effectively factoring in that India would build 13 aircraft carriers — the US today has 11.

He calls expensive valuation “telescopic”: shift the telescope by one degree and you are looking at a different galaxy, not just a different star. Small changes in heroic assumptions swing the price wildly. Indian defence companies were trading at 11x sales with 1% exports, while global peers traded at 1.8x sales with 51% exports. Accor, with 5,000 properties and 800,000 rooms, was valued like the Taj with 24,000 rooms. The future growth was already in the price.

And the iron law underneath it all:

Profit margins are mean-reverting. Valuations are mean-reverting. “This time is different” are the four most dangerous words in the market.

The mistake that took him twenty years to see

The confession at the heart of the talk. For two decades Shah valued stocks against his own yardstick of cheap and expensive — which meant against his own high required rate of return. He sold winners the moment they crossed what he considered fair.

What he missed: he is not the one who sets the price. As a company scales and becomes liquid, it attracts a different class of buyer — a sovereign fund happy with 4% annual returns, say — whose required return is far lower than his. The price is set by the last marginal buyer and seller, not by him.

The expected returns demanded by marginal investors keep dropping as a company matures. I was arrogant that I was the arbitrator of cheap and expensive. The compounding of stupidity is very cruel. Had I held those stocks, I would probably be sipping something in the Bahamas instead of giving this lecture.

His personal “maximum culpa” slide is a graveyard of companies he sold because they looked expensive by his measure, that then compounded for years.

Practical handles

On selling a winner without agonising: he now drips out in small tranches — “a foot in the door” — rather than dumping everything, so the average sell price neutralises over time. On selling a loser: if a stock falls 20-25% and he does not have the courage to buy more, that is the first warning that he doesn’t actually understand it well enough. He gives a static, non-falling laggard roughly three to four years, because cycles tend to correct in that window. On taxes: booking a temporary loss to offset capital gains is sound, but selling and buying back is “easier said than done.” On euphoria: when thematic funds are launching and TV anchors are talking about a sector, that froth is the opportunity — fortunes are made because markets are irrational, so take advantage of it rather than being swept up.

His pre-mortem habit: before selling, he imagines the stock 30-40% higher a year later, slapping him in the face, and asks how he’d feel. If the justification still holds up, he lets it go and sleeps.

Regret is inevitable. Suffering is optional. Investing is not about maximising everything; it is about sleeping well at night.

On AI and on growth

Asked about AI (he co-founded Needl.ai), he frames it as pure upside for investors: it automates grunt work, filters noise, and applies your framework consistently across mountains of data. But the model is just infrastructure available to everyone — “a Formula 1 car in my hands is useless.” The edge is in asking the right question, feeding the right data (70-80% of mistakes come from faulty inputs), and interpreting the output.

And a line worth keeping on growth:

Cancer is also growth — but it’s unwanted growth. Not all growth is value-accretive. Growth only creates value if it beats the required return on capital and there’s no dilution.

His example: Indian telecom grew vertically for years, yet of two-and-a-half surviving players, only one makes a profit. Huge market growth, almost no money made.

Key Takeaways

Selling is structurally under-discussed: only 1.7% of pre-regulation broker reports were “sell” rated, because analysts who wrote them lost access to the company.
Loss aversion is roughly 4:1 — a dollar lost hurts about four times more than a dollar gained feels good — which is why investors only scrutinise their portfolios under stress.
The arithmetic of losses is brutal and asymmetric: a 50% loss requires a 100% gain just to break even.
Buy-and-hold is not buy-at-any-price. The Nifty Fifty were great businesses bought at bad prices and still lost 70-80% over a decade. Wipro once = 4% of India’s GDP and took ~20 years to break even.
Reverse-DCF: don’t forecast value forward; take today’s price and solve for the growth rate the market already assumes, then sanity-check it (Cochin Shipyard was pricing in 13 aircraft carriers).
Expensive valuations are “telescopic” — a one-degree shift in assumptions moves you to a different galaxy of price.
The price is set by the last marginal buyer and seller, not by you. As a company scales, it attracts buyers with much lower required returns (sovereign funds at 4%), so what’s “expensive” to you can keep rising.
You cannot copy another investor’s playbook — Buffett’s public advice contradicts his own practice; the context (capital base, temperament, time horizon) is everything.
Reasons to sell are the mirror image of reasons to buy: when facts, conviction, or management change, the buy thesis no longer holds.
The “Four Bs” of selling: valuation Bizarre, business model Broken, hypothesis Busted, or Bad/Bandit management.
A promoter selling a small stake is not automatically a sell signal — they have personal and diversification reasons; revisit your thesis, don’t reflexively follow.
A 20-25% drop where you lack the conviction to buy more is the first signal you don’t understand the business well enough.
Drip-sell winners (“foot in the door”) to neutralise your valuation bias on the average exit price.
Pre-mortem journal: before selling, imagine the stock 30-40% higher in a year and check whether your reasoning still holds; if it does, accept the regret and move on.
Not all growth is value-accretive — “cancer is also growth.” Growth only adds value if it beats the required return on capital without dilution.
Profit margins and valuations both mean-revert; “this time is different” is the market’s most dangerous phrase.
Public-sector lenders trade at a structural discount for three reasons: decisions not made in shareholders’ interest, forced dilution at the worst times (recapitalised when multiples are low), and steady market-share loss.
AI is an infrastructure layer available to all; the edge is the right question + right data (70-80% of errors are bad inputs) + right interpretation.

Claude’s Take

This is the rare investing talk that is mostly confession, and it is better for it. The framing device — that buying gets all the literature while selling stays a “dark continent” — is correct and underexploited, and Shah’s willingness to put his own graveyard of premature sales on a slide gives the abstractions teeth. The single most valuable idea here is the marginal-buyer insight: that “expensive by my yardstick” is meaningless because the price is set by whoever has the lowest required return, and maturing companies recruit exactly those buyers. That reframing is genuinely load-bearing and worth more than most of the buy-side frameworks floating around.

The reverse-DCF point is old hat to professionals but cleanly explained, and the “telescopic valuation” metaphor is a keeper. Where I’d dock points: the talk leans on survivorship and hindsight — the regret over stocks he sold that later compounded is real, but for every Page Industries that ran away, an unmentioned thesis presumably broke and vindicated an early exit. He half-acknowledges this (“you can’t get all decisions right”), but the emotional weight of the talk tilts toward “I sold too early,” which is a survivor’s lament, not a symmetric audit. The pre-mortem and “sleep well at night” framing is sensible but soft, and the bull-market warning at the end is standard-issue caution.

Still, the signal-to-fluff ratio is high, the mental models are durable, and the honesty is disarming. An 8. It loses the last two points because the prescriptions stay deliberately personal (“works for me, may not work for you”) rather than resolving into anything you can act on mechanically — which is intellectually honest but limits the payload.

The Art of Selling | Kuntal Shah | Oaklane Capital Management LLP | Co-founder - Needl.ai | CFA Society India