
A business that adopts the standard model competes where the established companies are already strongest. When we built the marketing agency in the United States, our model was almost identical to everyone else’s on paper. We changed how the work was done, and we built that around the one thing we were actually good at.
The same model as everyone else
The model was simple: create and launch marketing campaigns alongside our partners. Nothing in that was new. We had no experience creating marketing campaigns, and we were doing it in the US, an enormous and brutal market.
So we asked ourselves a specific question. What are our actual strengths, and where is our leverage? We did not have much. Our competitors had some of the best copywriters I knew of, and we were not going to out-write them. What we were good at was the technical side. We understood systems, and we could find patterns.
Why did we reverse-engineer campaigns instead of copying them?
The plan followed from that. We would reverse-engineer every campaign those companies had made and build something at least as good. If it did not work the first time, we would iterate until it did.
It was a strange plan, because nobody else was doing it. Everyone in the space copied everyone else, but at the surface: they copied finished campaigns. We sat down and analyzed the logical structure of the campaigns so that we could rebuild them technically, from the inside out.
The difference is between an output and the mechanism that produced it. Copying an output leaves a company one step behind for good, always chasing the last thing someone else made. Understanding the mechanism puts it level, and then a little ahead, because a mechanism can be run faster and in directions its owners have not tried. From the outside the two activities look alike.
The analysis also did something we had not planned. Taking the campaigns apart and understanding the patterns behind them trained us. We became better copywriters and marketers than we had any right to be. What began as cold technical analysis turned into a learning loop, and that loop eventually produced campaigns better than the ones we had started from.
We could not win on copywriting, so we refused to let the fight happen there.
We moved the contest to technical pattern recognition and speed, where our advantage was largest and theirs was smallest. On that field the best copywriters in the industry had no answer. Why we chose speed, and what it takes to keep it, is in Speed is a structure.
Differences nobody could see from outside
From the outside our agency looked like every other one: the same service, the same category, the same customers. The differences were internal: how we contacted partners, how we built campaigns, how fast we iterated and who did what. None of that was visible to an observer, and all of it showed in the results.
That is why I pay little attention to whether a business model looks original. A model is a description. The game is the specific way a company sets its strengths against the gaps in a market, and two companies with the same description can end up in opposite places. It is also why I distrust off-the-shelf models. They are not wrong. They were designed for someone else’s hand, and adopting one whole means competing where the incumbents have years of accumulated advantage.
The game is not fixed either. The version that wins today is tuned to today’s market, today’s competitors and today’s tools, and all of those move. The first winning version is only a version, and the rules have to be rewritten as the field shifts.
A weak hand was the reason
We had no money, no experienced writers and no reputation. That forced us to build a faster, more iterative machine than the companies that had all three. Had we been comfortable, we would have copied, and copying would have kept us a step behind.
Not knowing the standard system helped as well, which I understood only in hindsight. Had we arrived trained and fluent in how the industry did things, we would have done them that way, behind people who had been doing them that way for years. I am not romanticizing inexperience. We could not see the standard path, so we built from what we had.
It was also lonely for a while. When we decided to reverse-engineer campaigns technically, there was no one to tell us it would work, because no one else was doing it. Copying at least has the comfort of company.


