Thinking

The biggest black box in any business is why its customers do what they do. This is how I open it, and the ideas that hold the method together.

Opening the black box

Nine principles behind my work, and the books whose ideas run through several of them.

  1. Understanding is the product

    What a business gets from my work is a written understanding of its customers: what they are trying to resolve before they buy, what they doubt and what they ignore. Every experiment adds to it, the losing ones included.

  2. Behaviour before opinion

    People are poor judges of what they will want and how they will feel, and they misjudge in consistent ways. So I start from what visitors do: where they come from, what they read, where they stop. Asking them comes later, when the behaviour stops yielding answers.

    On the shelf: Stumbling on Happiness

  3. Trust the numbers before the results

    A dashboard can be precise and wrong. Before any experiment counts, the measurement has to be proven end to end, revenue included, and a test in which both versions are identical has to show how far the numbers move by chance alone. That noise is the yardstick for every result after it.

    On the shelf: Fooled by Randomness

  4. Explain before you persuade

    Most businesses fail first at explaining themselves: what the product is, which problem it solves and what it costs. Plain, complete information earns trust, and polish cannot replace it.

  5. A few variables decide

    A business can list a hundred things to improve. Usually a handful decide whether people buy: which problem feels urgent, which promise matters, what creates doubt and what settles it. Price, guarantees and refunds come later in the same sequence.

    On the shelf: Predictably Irrational

  6. One change, one lesson

    Each experiment changes one meaningful thing at the biggest bottleneck and tests it against the current version, so the result answers one clear question.

    On the shelf: Poor Economics

  7. Losses are evidence

    A losing test is evidence about what the audience ignores, distrusts or misunderstands. I record it beside the winners and revise the belief that made it look right.

    On the shelf: Antifragile

  8. Count what reaches revenue

    A test is a win only once it holds downstream, in revenue or in the quality of the customers it brings. What it taught about the audience carries over to other businesses that share that audience, more than to others in the same industry.

  9. The tool is not the strategy

    I choose tools for the measurement and the experiments they make possible, and the stack a business already has can be kept and improved. The strategy is the order of the questions: what to measure, what to believe, what to test first.