A calendar, drawn in chalk, with crumpled used pages below..
We Don't Usually Think About the Far Future

Tech companies are recent. They also tend to die quickly. If you wanted your tech company to last awhile, how would you do it?

That’s one of those “selling yourself short” questions. If you wanted your company to last awhile, how would you do it? Tech companies don’t last long.

Let’s talk about what counts as “long” and why modern-style companies die much, much faster.

The oldest companies in the world have a few things in common, but the big one is being about their mission, most commonly supporting their family. Why? Well, companies constantly set short-term targets. You have to balance the short term against the long term, and companies are terrible at that. One trick to avoid that is to stop working for Q3 profitability and start working for your children and your grandchildren.

This is a scathing indictment of companies, if you want to do anything long-term.

In fact, even tech companies recognise this and try to work against it. That’s how obvious it is.


What’s long-lasting, for a company? Let’s go through the relevant Wikipedia page.

There are a number of very, very old Japanese companies. The Kongō Gumi construction company in Japan made it around 1,400 years, and is family-owned. Nishiyama Onsen Keiunkan was over 1,300 years – 52 generations of the same family. There are several more family-owned Japanese businesses in the oldest few. The Koman Hotel (46 generations in the same family,) the Hōshi Ryokan (46 generations,) Genda Shigyō (merger of two family businesses in the early 1800s.)

Looking outside Japan, Staffelter Hof in Austria was an abbey for its first 940ish years, serving God rather than family, before becoming family operated for the last 7 generations. St Peter Stiftkulinarium started as an abbey wine cellar that gradually expanded, in the 1800s, into a restaurant. Château de Goulaine has been family-owned since 1,000 AD. Pontificia Fonderia Marinelli (1000 AD) has been owned by the Marinelli family that whole time.

There are a few government institutions that are similarly old – the UK mint and Monnaie de Paris (Paris mint,) though it’s hard to call that a business.

And other than me skipping over some (also family-owned) Japanese businesses to give Europe a chance to even make the list, that’s basically it. The big exceptions are a monastery-winery which later became family owned, and an Abbey wine cellar.

Businesses that change hands and are run for simple profit rather than family or God don’t make it that long. Even among for-profit family businesses, a fair number of the oldest sell religious goods.

It’s about mission rather than simple financial opportunity. And it helps if what you sell (construction, lodging, religious goods, alcohol) has a timeless appeal.

But What’s the Problem?

Modern economics is all about incentives – people do what they’re rewarded for doing, and avoid what they’re punished for doing. A company is, obviously, built in this same spirit. Companies have salaries and payments to give people incentives, and managers to apply punishments like demotions and firing. The whole “managers managing managers” hierarchy is an incentive system of its own. People want to impress important people, and a hierarchy of executives establishes a common set of important people. Again, rewards and punishments.

Companies tend to use these incentives in a direct, primitive sort of way. Annual performance reviews, promotion to a small set of named ranks, each department has a separate “pool” of money called a budget and so on.

Decisions are made by individual executives or by formally-appointed committees. There are many, many decisions made outside of that structure because neither of those is very good, but it’s important for the company to pretend it’s all done that way, because that’s how the company incentive structure is set up. You have designated responsible individuals, who are supposed to be doing specific tasks, and they are rewarded or punished for how well they do that.

There are the top people, a CEO and a Board of Directors. And they establish a cadence of short-term goals like quarterly profit, and longer-term goals, such as up-to-five-year plans. That’s about as far in the future as you can plan with a structure like that. Individual people may individually think about terms longer than that, but the structure is set up not to be able to.

Real human systems have much blurrier boundaries. That’s because they’re established for longevity, not for profitability.

“If you want to go fast, go alone. If you want to go far, go together.”

A business can’t have more than one first priority, in the long run. If making money for next 20 quarters is the priority, somebody will find a way to make you choose between that and still existing in ten years. Indeed, tech companies are masterful at doing that to larger, older companies. Whatever your two highest priorities are, you’re going to have to choose between them sometimes.

If a company has “providing for the family” as its highest priority, it can keep doing that. If going outside the family was more profitable (often,) a modern profit-first company would do that. And in so doing, it would lose loyalty that can only be usefully measured beyond a ten-year horizon.

And – hey look – that means only companies that do not do that survive for a truly long time.

You’ll notice that God makes a fine stand-in for family for this entire comparison. You’d need worshippers to choose God over profits. But not infrequently they do, so it works.

How Would Family Businesses Be Better?

It would be reasonable to say, “corporations kill an awful lot of family businesses. I don’t see how family businesses are better.” And that’s true.

Corporations are a simplistic, explosive structure, designed to grow rapidly with people who don’t trust each other a lot. Look back at that description: formal roles and rules, simple but spelled-out boundaries, rules that people can make up quickly and change quickly on the spur of the moment. That’s what you do when you don’t already have a lot of trust, because you’re sacrificing your people’s ability to make blurrier, more interesting decisions. And it’s what you do when you don’t have a lot of tradition or precedent, because you’re eliminating “what has historically worked” for “what we can make up easily and rapidly.”

Admittedly, there’s an extensive, expensive tradition of executive coaching to replay the successful patterns of corporate structure that we’ve found since (depending on the coach) the 1970s or even the 1870s. But you’ll notice that if the topic is “getting a big group of people to do something well for a long time,” neither of those times is long at all. There are a significant number of companies that have made it for more than 1,000 years. Having a tradition that dates back, at best, to the late 1800s doesn’t do that. There are simply no old examples to emulate.

What about the East India Company, arguably the oldest example of the modern corporation? Terrible for longevity. Also dead since 1874.

Corporate structure is descended from those which, like the East India Company, exist on a time-limited basis to exploit some specific opportunity, such as the conquest of North America. They run short-term because their mission is short-term. The structure reflects that, which is why you have so many modern financial visionaries opining that corporate governance is so short-sighted. It’s designed that way. Its incentives are set that way. It is intentionally copied from structures that work that way. If you want a different result, you’ll have to emulate something else.

Emulate Something Else

Here’s something weird: some tech companies do try that. It’s too early to say that no tech company will last a long time, since they haven’t had the opportunity yet. Depending on how you define “tech company,” they’ve had somewhere between 50 and 100 years to try, and many have survived for a noticeable fraction of that. The tech industry isn’t good at thinking long-term, though.

But when I say “try that,” I mean that some tech companies try to do an end-run around the long-term destructive tendency of corporations to prioritise their incentives. I don’t mean founding tech companies around God and family. I’ll quote Tobi Lutke of Shopify for a second. He describes this as the “shared priorities of some of the best companies in the world:”

Priority #1: Go build the best product you can possibly build
Priority #2: Make some revenue so that you can do more of Priority #1
Priority #3: Never do Priority #2 at the expense of Priority #1

What’s interesting there is that he’s declaring it to be a mission-driven company rather than a profit-driven company. I’m not going to say “this will clearly work,” because the really old mission-driven companies have family or God as their mission, and those are very powerful missions.

But it’s clearly an attempt to get around the usual corporate priorities.

And the main reason to try to get around them is to look at their track record. I don’t just mean “not creating long-lived companies,” though of course that’s true. The companies you get with a “normal” corporation are exploitative, short-sighted things, prone to strip-mining and destruction. You don’t have to be a big believer in communism or non-market economies (I’m not) to see the problems there.

There are always opportunities to eat the seed corn, to drive all year without changing your car’s oil, and to cannibalise your product for short-term profit. A company has to resist those temptations constantly. And, frankly, they aren’t good at it. A new CEO who gets a giant chunk of stock is well aware that he can lay off vital employees, stop new development, generally cut all the long-term projects of all kinds, and then sell off all his stock when the market rewards all this behaviour. One reason to overpay so highly for CEOs is that they are in a position to simply destroy your whole company for their own profit.

And, surprise of surprises, companies aren’t good at choosing leaders every 3-10 years without making a bad mistake. And so the modern company tends to drive itself into the ground in a few decades, or around a century. Statistically it happens much faster than that, of course, but I mean specifically the long-lived ones. IBM has made it for 110-ish years so far. It might conceivably make it another hundred before it chooses disastrously bad leadership that properly destroys it.

If it could make it to five times that, it could really be a contender.