Patterns of Business

The more businesses I examine, the more I find that companies which appear entirely different often share similar structures at a higher level of abstraction.

Perhaps this is because human beings do not change very quickly. People have always wanted to communicate, acquire information, trade, and do less work.

Communication → letters → telephone → messaging → social media
Information   → libraries → search engines → AI
Transactions  → cash → cards → Stripe
Less work     → tools → software → agents

What changes is not the demand itself, but the form, cost, speed, and frequency with which it can be satisfied.

New technologies usually cause three changes.

1. They satisfy an existing demand far more cheaply.
2. They release latent demand that had been suppressed by cost and friction.
3. They create new supply structures and new bottlenecks on top of the resulting activity.

AI has not invented an entirely new desire either. It is giving a new form—calling an agent—to the old desire to accomplish more with less effort. When the cost of production collapses, output grows, and scarcity moves from execution toward judgment and trust.

The most important question to ask about a technology therefore comes before its features.

Which enduring human demand does this technology satisfy in a new way, and what becomes explosively more abundant as a result?

Business models follow a similar logic. Technology may change the interface, but the means of making money are often recombinations of established patterns such as subscriptions, transaction fees, advertising, and bundling. The important task is not inventing a novel label. It is applying the right old pattern to the bottleneck that a new technology has relocated.

Finding patterns in successful companies is not enough, however. In retrospect, every choice made by a successful company appears wise, while every choice made by a failed company appears mistaken. Unless we also study companies with similar traits that disappeared, we are likely to mistake correlation for cause.

My conclusions so far are these.

• Fundamental human demand lasts longer than technology.
• Technology changes the cost and frequency of behavior more often than it creates demand.
• The largest opportunities appear in new bottlenecks after an activity explodes.
• A repeatable business model and a durable moat are separate questions.
• We should study recurring paths to failure alongside the common traits of success.

Books to Read

1. The Business Model Navigator — Drawing on 350 companies, it argues that most business model innovation recombines existing concepts and organizes them into 55 patterns. It is the most direct framework for this intuition.
2. The Halo Effect — It warns against retrospective stories that blur cause and effect when we study successful companies. It is necessary for finding patterns without being fooled by them.
3. 7 Powers — It explains how a good business becomes an enduring advantage. It should clarify the difference between a repeatable business model and a moat that is difficult to copy.
4. Why Startups Fail — It examines recurring paths to startup failure rather than reducing failure to an incapable founder. It balances the bias created by studying success alone.
5. Good Strategy/Bad Strategy — It returns strategy to diagnosing the central challenge and forming a focused response. It should help turn an identified bottleneck into coherent action.

I plan to read them in that order: The Business Model Navigator, The Halo Effect, 7 Powers, Why Startups Fail, and Good Strategy/Bad Strategy. First learn the patterns, then guard against analytical error, understand moats and failure, and finally compress the result into strategy.

Understanding a new technology may not require imagining a new kind of human. It may be closer to observing what unchanged human beings begin to do far more often once they acquire a new tool.