Pieter Slegers — Our Shopping List (Part I)

3i Group is a listed company whose job is to buy stakes in private businesses, improve them, and sell them at a profit. In practice it is one investment wearing a company's clothes: about 90% of its private-equity returns come from a single holding, the European discount chain Action.

Action is the thing to understand. It sells cheap non-food goods, has grown revenue at 26% a year since 2011, and — the key number — recovers the roughly €500,000 it costs to open a store in under twelve months. A business that gets its money back that fast can pay for its own expansion out of profits, which is how it manages to open roughly a store a day without borrowing. It advertises almost nothing; low prices do the marketing. And it has barely started in Britain or Scandinavia.

Because 3i is a holding company, the way to value it is not a profit multiple but net asset value — what the things it owns are worth, added up. At the end of 2025 that was 3,017 pence a share, and the shares trade at 2,859 pence, so you can buy the assets for slightly less than they are reckoned to be worth. The chief executive owns £585 million of stock, 865 times his salary, which means he is paid by the same thing you are.

Adyen is the plumbing behind card payments for large retailers. When you pay, the money has to travel from your bank through the card network to the shop, and normally several middlemen each take a slice. Adyen built one system that does the whole journey, so it removes those middlemen and keeps a fee on every transaction.

Its advantage compounds quietly. Every payment it handles teaches its software a little more about which transactions are fraud and which are genuine, so it declines fewer real customers than rivals do. Better approval rates win bigger merchants, which means more payments, which improves the software again. Competitors, mostly built by bolting old systems together, cannot easily copy that.

The verdict here is quality yes, price no. The shares are at the cheapest valuation in the company's history — 25 times next year's profits — and the stated buying level is 20 times, about €765 against a market price of €970. So it goes on the list, not into the portfolio.

Alphabet is Google: search advertising and YouTube, plus a cloud computing business and the Play Store. Advertising is still about three-quarters of what it earns, and it shares that market with Meta in what is effectively a two-firm arrangement.

Nothing about the business is in question here. More users produce more data, better data makes the ads work better, better ads bring more advertisers and more money for search — a loop that is very hard to break into.

The objection is purely the price. The shares have risen 123% in a year and now cost 29 times next year's profits. The stated willingness to buy is at 18 times, which works out at $210 against a market price of $336 — a fall of nearly 40%. That is not a prediction; it is a refusal to pay up, written down in advance so it can be checked later.

ASML makes the machines that make advanced computer chips, and it is the only company in the world that can. Its market share is about 90%, its machines weigh as much as two blue whales, and China has spent years failing to copy one. Every leading chipmaker, from TSMC to Intel, buys from it.

Two features make it more than a one-off equipment sale. First, once a chip factory is built around ASML machines, switching would cost billions, so nobody does. Second, ASML then collects service fees on each machine for twenty to thirty years — a long annuity attached to every sale. And because it supplies everyone, it makes money regardless of which AI company ends up winning.

Again the answer is not now. The shares cost about 37 times next year's profits against a ten-year average of 32. The stated entry level is 25 times, or roughly €745 against a market price of €1,245.

Cintas rents work uniforms to businesses, then collects, launders and redelivers them every week, and sells the same customers first-aid kits, fire-safety checks and hygiene supplies while the van is already there.

The economics come from the delivery route. Once a van is driving a street, adding another customer on it costs almost nothing, so each new client is more profitable than the last — and a competitor with fewer customers per route can never match the cost. Customers rarely leave: the uniforms are cheap relative to the hassle of switching, and more than 95% renew every year.

The growth argument is a penetration one. Cintas serves about a million businesses; there are roughly sixteen million that could use it, and most still wash their own. The shares cost 34 times earnings; the stated buying level is 25 times, about $132 against $179.

When an insurer writes off your car, it has a wreck it does not want and no interest in selling. Copart takes it, lists it on an online auction, and charges both the insurer and the buyer a fee. Think of it as eBay for written-off cars.

Two things make it hard to attack. It owns 19,000 acres of land across the United States, which matters enormously when a hurricane produces 90,000 wrecks in a night — most competitors lease their yards and simply run out of space. And it has a million registered buyers worldwide against roughly 150,000 at its nearest rival, so its auctions fetch higher prices, which brings more sellers, which brings more buyers.

The long-run driver is that cars are now full of sensors and software, so repairs cost more, so insurers write off more of them rather than fixing them. That has been true for forty years. Right now the shares have halved from their peak on falling volumes, though prices per car still rose 7% last quarter. The company also holds net cash worth 15% of its market value — strip that out and it costs 18 times profits instead of 21. The stated buying level is 18 times including the cash, about $28.70 against $33.80, which is the stricter of the two ways of counting.

FICO owns the credit score every American lender uses. It licenses the formula to the three credit bureaus and collects a fee each time a score is pulled for a mortgage, a car loan or a credit card. About 95% of the market runs on it.

What real pricing power looks like: mortgage-score revenue rose 52% in a year, not because more houses were sold but because FICO simply charged more, and nobody could go elsewhere. Roughly thirty cents of every dollar of revenue ends up as profit.

The valuation lesson here is the useful part. On the headline figure the shares are at their cheapest in a decade, 24 times next year's profits. But FICO pays its staff heavily in shares, and that cost — equal to a quarter of reported profit — is not in the headline number. Adjust for it and the real multiple is 30 times, not 24. The stated buying level is 25 times after that adjustment, about $901 against $1,058. Two weeks later a US regulator approved a rival score for mortgages and the shares fell hard, which is why FICO reappears on the May Best Buys list.

Action is a European chain of discount stores selling cheap non-food goods — the sort of shop where nothing costs much and people leave with a basket full of things they did not plan to buy. It is not listed; the way to own it is through 3i Group, which holds the controlling stake.

Its growth has been extraordinary and steady: revenue compounding at 26% a year since 2011, roughly one new store opening every day. The reason it can grow that fast without raising money is that a new store costs about €500,000 and pays that back within a year, so profits from existing stores fund the next batch.

The competitive loop is simple and self-reinforcing: buying in enormous volume lets it undercut everyone, low prices spread by word of mouth so it barely advertises, and the savings go back into lower prices. There is a lot of room left — several European countries, including the UK and the Nordics, have no Action stores at all.