80–120 compatible marketing combinations in one cycle

Finding a working customer-acquisition system for small businesses

In 2018, I created a product to help small businesses find a working acquisition system instead of betting everything on one advertising hypothesis. At each level — audience, channel, strategy, positioning, and conversion method — we selected several promising options, combined them, and removed combinations that did not fit together. One cycle produced 80–120 complete combinations to test.

The savings came from how we produced the materials. Ad campaigns, landing pages, quizzes, and other assets were developed as one coordinated package and built on each other, rather than created from scratch for every combination. This brought the cost of testing one combination down to $12–24. I explained the difference with two games: test one combination at a time, or reuse components to explore many compatible options.

According to the data from that period, finding a working combination cost about 50 times less than developing separate hypotheses one by one, while leads from successful combinations cost 3–5 times less. The product was sold to and implemented by dozens of small-business owners.

I see new potential for this method today: AI agents could make it cheaper to prepare and maintain many ad and landing-page variants — the work that limited the breadth of experimentation in 2018. Advertising spend and validation with real demand would still be necessary, but the search system itself could become more accessible. It could form the basis of a new AI tool, and I want to open the method for the community to build on.