Good Business Is Sustainable Altruism

In the long run, you get what you incentivize. I didn’t come up with this, but it took me a long time to really digest it. It applies to business, and also to policy design, team management and education, even to how you relate to yourself. Any system, run long enough, ends up looking like whatever its incentive structure rewards.

Put the other way around, if your goals run against people’s long-term incentives, then no matter how strong the system or how strict the rules, all you end up doing is creating friction. Force might hold things together in the short run, but in the long run you will lose to structures that go with human nature.

Only later did I understand why Buffett and Munger treat the business model as the core thing to examine in an investment. It tells you whether the company can make money (plenty of bad models make money in the short run), but more importantly, it tells you what the company is being trained to become in order to make that money.

When most people talk about business models, they talk about revenue sources, margins and pricing strategy, and all of that is right. But as I see it, a business model is also an incentive structure for the whole organization. It decides what is most easily rewarded for everyone in the company, and the direction of that reward shapes the organization’s habits day after day.

Imagine two companies. The first makes most of its profit from long-term users renewing. Its sales, product and support people wake up every morning thinking: “How do we make users feel it’s worth it, so they want to stay?” Nobody needs to write that on the wall, because it is simply the easiest way to make money.

The second makes most of its profit from new customers’ first purchases, so the organization’s energy naturally flows toward marketing, sales scripts, channels and packaging. Product matters too, but it is treated more as packaging material for getting that first purchase, and how users feel after they buy doesn’t directly affect the most important number on the income statement.

Both companies may make money, and in the short run the second may even make more. But over time, the way each one makes its money trains it into a completely different organization: the first keeps getting better at delivering value, the second at building up expectations.

Munger said something to the effect of “Show me the incentive and I’ll show you the outcome.” I’d add: show me what a company has turned into five years later, and I can work backward to what its business model was rewarding.

Costco makes its money on membership fees

If you don’t know Costco well, let me quickly explain what makes it special. Costco is an American warehouse-style membership retailer. You have to pay an annual fee (65 dollars for a basic membership, 130 for the executive tier) before you can shop there. Once inside, everything is priced very low, because Costco strictly caps its markup on goods at about 14%, while a typical supermarket’s gross margin is usually 25% to 35%. In other words, Costco’s markup barely covers its operating costs, and the company’s real profit comes from membership fees.

This simple structural change has consequences far deeper than you’d expect. The most direct one is that the profit anchor moves. When a company’s main profit shifts from “how much we make on each item” to “how many people are willing to renew,” the whole organization’s behavior gets rewritten. The number Costco cares about most is its membership renewal rate (currently steady above 90%). To get nine out of ten members to pay again voluntarily every year, the only sustainable way is to make them really feel it’s worth it.

The low-margin constraint has also become the way Costco trains itself. Since it can’t coast on fat markups, it is forced to keep getting better at supply chain efficiency, selection discipline and inventory turnover. It carries only about 4,000 SKUs (a typical supermarket has 30,000 to 50,000), with just a few carefully chosen options in each category. In effect, it does the filtering for you and backs it with its own reputation.

Costco’s way of keeping users happy is also extremely simple, which is why it lasts. Plenty of companies want to please users too, but their idea of “pleasing” is far too broad: subsidies, gimmicks, over-the-top service and emotional marketing all count. These things are expensive and hard to keep up, and in the end they turn into a bottomless pit. Costco doesn’t play that game. It does three things: low prices, reliable quality and a consistent experience. Every time you go to Costco, you can check that it hasn’t lied to you: prices really are lower than elsewhere, and the stuff really is solid. Certainty you can check for yourself, repeated a hundred times, becomes trust.

The point I think matters most is that the exit door is always open. Costco members can stop renewing at any time. There’s no two-year contract, no cancellation fee, and no customer service maze where you have to call three times before you can cancel. The user’s veto has never been taken away. In effect, Costco faces a referendum every year, in which members decide, by renewing or not, whether it is still worth the money. That’s where its moat is: users can leave, but they don’t want to. On the surface it seems to have “no barriers,” but precisely because it never uses force to keep people, it pushes itself to keep treating “keeping users happy” as the easiest way to make money.

Making money from denials and lock-in

With Costco in mind, let’s look at two industries where the incentive structure has gone badly wrong. American health insurance is the most typical example. The insurer’s business model is simple: collect premiums and pay out as little as possible. You pay thousands, even tens of thousands of dollars a year in premiums, but when you actually get sick and need the money, the insurer’s profit depends precisely on how many claims it can turn down.

So the most profitable skill in the whole industry became managing claim denials. Insurers employ large numbers of people to review claims and look for reasons to deny them, and they design complicated terms, so that when you most need help, you can’t understand your own rights. According to OECD data for 2022, administrative costs make up 8% of total spending in the US health system, and administrative spending per person is over 1,000 dollars, far higher than in other developed countries. A good part of those costs goes to working out “how to legally not pay you.”

The payer and the beneficiary are clearly the same person, yet the incentives point in opposite directions: you paid, hoping to be taken care of when you get sick; the company collected, hoping you won’t come to it when you do. The whole money-making mechanism systematically rewards “not helping the customer.” Within this structure, an insurer that sincerely and actively pays claims will have worse margins than its peers and actually lose out in competition.

Enterprise software is another example. The traditional enterprise software model that Oracle and SAP represent may be the most refined version of the “handcuff moat”: huge implementation costs (a large ERP project can easily cost millions of dollars and take years), custom development embedded deep in business processes, systems so complex they need specialized consultants just to maintain, and regular license audits. An audit is really a form of compliance intimidation: you may have violated the license terms without knowing it, and after the audit you owe a large sum. Once a customer is on the system, it’s very hard to switch, because switching costs too much, which has little to do with whether the system is any good. So the easiest way for the company to make money becomes raising switching costs.

The rise of SaaS was originally a rebellion against this incentive structure. The early idea was pure: pay monthly, no long-term contract, users can leave at any time, and the company is forced to prove it’s worth it every month. That’s the same logic as Costco’s. The irony is that many SaaS companies, once they got big, went back down the old road. Take Salesforce. It now pushes 3-to-5-year contracts, early cancellation means paying the full remaining balance, auto-renewal clauses catch many small and midsize businesses off guard, and data migration costs are so high that people don’t dare touch anything. The old promise of “leave anytime” has turned into another kind of handcuffs.

So the pull of an incentive structure is real. When “making money from friction” is easier than “making money from value,” even companies that started out under an “anti-lock-in” banner eventually get pulled back, and very few companies can resist. That’s why I think Costco’s design is especially worth studying: its structure blocks the “make money from friction” road, so the company never even gets a chance to try.

Big shops bully customers

Beyond the misalignment built into business models, there’s a subtler problem: even a company that starts out with a good incentive structure can lose it as it gets big. There’s an old Chinese saying: when the shop is big, it bullies the customer; when the customer is big, they bully the shop. Applied to companies today, it means that once a company is big enough that its users’ veto weakens, it starts testing the limits without even realizing it.

In the early days of a startup, you really do need each particular user. When a user says “if I’m unhappy, I’m leaving,” you lose sleep over it. You naturally think of yourself as the vendor, the side that has to win the customer over, and you do everything you can to serve every single person well.

But once you’re big enough and have enough users, your mindset quietly shifts: losing one user doesn’t matter, because others will come anyway. You’re confident in your growth, so you stop caring about each particular person. If a user is unhappy, you look at the overall numbers, see they’re fine, and move on. Slower support doesn’t affect the quarterly report, and if the terms get a bit more complicated, most people won’t read them anyway.

This is the incentive structure at work. Once users’ veto weakens (few alternatives, high switching costs, deep ties to an ecosystem, so that even unhappy users can’t be bothered to switch), the short-term cost of mistreating them drops, and the organization naturally drifts that way. Users go from “a specific person” to “a data point in DAU,” and “keeping users happy” goes from an instinct to a task that only gets done once it’s written into OKRs. The company also starts serving many masters at once: users, shareholders, regulators, partners, internal processes. It talks about customer obsession, but when quarter-end revenue targets clash with user experience, the user never wins.

So the vendor mindset can’t just be early-stage enthusiasm. It has to become an institution, which concretely means keeping the user’s veto real, always. As long as the veto exists, the company has to stay humble. Once it’s gone, humility can only be kept up by the founder’s personal discipline, and humility that depends on personal discipline is hard to scale.

Handcuffs and magnets

Put these threads together and you see a rather ironic paradox: many companies work desperately to build moats that lock users in, and it is exactly this lock-in that pushes users into learned helplessness. There’s a classic psychology experiment in which dogs are given repeated shocks they can’t escape, and in the end, even when the cage door is open, they no longer try to leave. Users locked in by a moat are the same: they tried, found they couldn’t leave, and eventually gave up struggling.

You may have been through this yourself: you’ve used some piece of software for years and have been unhappy with it for a long time, but all your data is on it, your workflow has been built around it, and switching means learning something new, so you keep putting up with it. You’re still paying, but you’ve become a “trapped payer.”

When users end up like this, the company changes too, because its biggest source of outside pressure, “users might leave,” has disappeared. Without outside pressure, an organization slowly goes dull. Since existing users can’t leave anyway, the drive to innovate weakens, and since people who complain have no alternative, service standards slip. The smartest people get moved over to retention mechanics and upselling and stop working on the product. The moat looks deeper, but in fact the company has wiped out the outside force that pushed it to get better.

So there are two completely different kinds of moat. One is the handcuff kind, which keeps users through friction, switching costs and the hassle of cancelling: users stay because leaving is too much trouble. It makes the short-term financials look good, but in the long run it dulls the company, because what it rewards is “making it harder for users to leave,” and the company stops thinking all the time about how to please users. The other is the magnetic kind. The door is always open and users can leave whenever they like; what keeps them is lasting value and trust. Every time users pass the city gate, they feel the city is more worth staying in, so they choose to stay.

Costco is a magnetic moat. Members can skip renewing every year, and nothing stops them from leaving. But precisely because of this “leave anytime” design, Costco’s organizational culture has been pushed to keep making users feel it’s worth it, and its renewal rate of over 90% has been won that way, year after year.

Sustainable altruism

To close, here’s a personal belief I’ve gradually come to: good business is a kind of sustainable altruism. That may sound too idealistic, and I’m not saying good companies should do charity. If a company’s way of making money structurally requires it to hurt users (raising switching costs, creating information asymmetry, profiting from users’ inertia or confusion), it is bound to be short-lived. You may be raising both the user’s costs and the user’s pain at the same time. That bill doesn’t show up in the short run, but in the long run it will always land on you.

The reverse also holds. The best business models are ones where the company’s structure naturally points toward satisfying end users, so that “keeping users happy” becomes the simplest way to make money. That doesn’t depend on the founder’s kindness or on rallying calls from the company culture. It depends on structure, on incentive mechanisms that make “doing the right thing” the default and make “taking advantage of users” very hard.

As long as “keeping users happy” is the easiest way to make money, an organization accumulates trust, process discipline and scale efficiency over the long run, and none of that is easy for anyone else to copy quickly. Making money, having a moat and having satisfied users are often really the same thing, seen at different points in time. As for big shops bullying their customers, I think it’s mostly because the company’s structure already lets it, since users couldn’t leave even if they wanted to; it has little to do with the company turning bad. So to me, whether a company can stay good for the long run depends on whether its structure keeps giving users the power to say no. A Costco member deciding every year whether to renew is exactly that power.