Industries are essentially games. Each industry has players competing to take similar inputs and produce similar outputs - faster and cheaper than everyone else. The inputs, rules, and outputs differ within each but the essence is the same.

Win the game and you’ll be rewarded handsomely.

As with all games people end up finding the best ways to win. The longer a game has been played with the same inputs, rules, and outputs, the more likely it is that the best ways to win have been found. Once everyone knows the best ways to win, everyone will begin playing like that.

-

Industry’s measurement of this is EBITDA margin. Individual industries have an uncanny way of grouping around a margin range and, just like with games, the more mature and stable an industry the tighter and more well-defined the ‘ways to win’ are.

-

Now what happens when the game changes?

Look at airlines. In the 1960s airliner technology was essentially linear. Want fly 5x more people? You’ll need to buy 5x more planes.

Airlines could compete on route selection, in-flight services, and safety reputation but the biggest contributor to their game was fixed - the aircrafts.

Then, in 1968, Boeing introduced a new technology to the game: the 747.

Now all of a sudden airlines could fly people at 66% of the cost they previously could’ve. Early 747 adopter’s margins expanded and they began reaping the rewards of their new game strategy.

Like any good advantage it could be manipulated and extended. Airlines with 747s transmuted their margin expansion into offering lower prices while maintaining historic margin levels. Lower prices stimulated demand and more people flew that previously couldn’t afford to.

747-equipped airlines now had cheaper prices and a growing share of the market. Other airlines were forced to match their economics and soon everyone was buying 747s.¹

Both the inputs and expected outputs of the airline game had changed.

-

So being first to harness new innovations can mean a significant EBITDA margin advantage over other players, but how long does it last?

What’s surprising is how quickly, across all industries, advantages disappear.

They are self-levelling.

As soon as one airline proved that 747s meant outputting cheaper airfares at the same margins it was clear they were winning. If other airlines wanted to maintain their positions in the market they had better start also buying 747s.

When a new advantage enters the game, the competition is pressured into acquiring it - or risk death. Try to name something more persuasive.

This pattern has happened continually across history. Time to equilibrium differs but it’s almost always <10 years and doesn’t correlate with the size of the EBITDA margin gains.

The obvious next question then is “What makes it stickier?”.

The best conditions seem to be:

1. High copying friction. Not just patents or trade secrets, but real operational complexity. Amazon’s fulfilment network isn’t valuable because it’s secret - it’s valuable because it’s hard to build. The steam engine wasn’t enough - you needed the whole railway network.
2. Compoundable. The really good innovations get better with scale. TSMC doesn’t particularly focus on making chips; they focus on getting better at making chips faster than anyone else. Each generation of process technology builds on the last and forms a compounding moat.
3. Valuable ecosystems. Innovations that create systems where staying becomes increasingly valuable, like how each AWS or Google Cloud service makes the next one more useful.
4. Regulatory moats. The surprising thing about railroads wasn’t that they made money - it was how regulation eventually protected their margins by increasing barriers to entry. The best technological advantages often end up wrapped in regulatory frameworks that institutionalise them.

The 747 case is a great anti-example of sticky innovation. Airlines couldn’t keep the margin expansion because:

• The innovation was purchasable (just buy the plane)
• There was no compounding gain with the rest of their fleet
• Customers faced no cost switching between airlines
• There was no regulatory moat

-

It’s clear that some innovations are stickier than others, but are some industries more ripe for EBITDA margin expansion?

The most interesting thing about these opportunities is that great ones do tend to appear alongside specific structural inefficiencies:

1. High transaction costs. Industries where a lot of value is lost in the process of doing business. Commercial real estate leasing is a classic example, ~6% of the transaction goes to brokerage fees, legal reviews, and document processing - none of these add value to the ‘matching’ of tenant to real estate space.
2. Artificial capacity constraints. Industries where capacity isn’t really limited by physics or economics, but by organisational structure. Before container ports were standardised different loading systems and incompatible container sizes significantly reduced efficiency.
3. Misaligned middle-men. Industries with powerful middle-men who aren’t adding value proportional to their take. Travel agents and the internet are a classic example. Agents weren’t bad actors, they just weren’t necessary after the internet hit the mainstream.
4. Hidden cross-subsidies. When one customer group is unknowingly subsidising another, there’s usually room for EBITDA expansion by segmenting properly. Enterprise software lived off this for years - the same product would cost $1M or $10K depending on the customer’s size. Find the segment subsidising another, solely target them, remove the subsidy, and lower your prices.
5. Manual processes at scale. Any industry where humans are doing repetitive tasks at massive scale is basically asking to be disrupted. The interesting part isn’t replacing the humans - it’s capturing some of the savings as margin instead of passing it all to customers.

The best opportunities usually have several of these characteristics layered together. Retail stock trading pre-Robinhood had high transaction costs, unnecessary middle-men, cross-subsidies, and manual processes.

What’s counterintuitive is that these inefficiencies often persist in plain sight for years. Everyone in the industry sees them, but existing players are too optimised around the current structure to be incentivised to fix them.

This might explain why outsiders often drive the biggest EBITDA expansions. They’re not trying to optimise the existing system - they’re building a new one that makes the inefficiencies obsolete.

The really valuable question is: which industries today have these characteristics but don’t look tech-enabled yet? Those are probably where the next big margin expansion opportunities are hiding.