How I approach an analysis
A repeatable method for turning a messy question into a clear answer.
When I sit down to analyze something — a market, a dataset, a decision — I try to follow the same loop. It keeps me from fooling myself.
1. State the question precisely
A vague question produces a vague answer. "Is this a good market?" becomes "How many people have this problem badly enough to pay, and how is that number changing?"
2. Decide what would change my mind
Before looking at any data, I write down what evidence would push me toward yes and what would push me toward no. This is the single most important step — it's the antidote to confirmation bias.
| Signal | Bull case | Bear case |
|---|---|---|
| Demand | Growing search volume | Flat or declining |
| Willingness to pay | Existing paid alternatives | Everything is free |
| Competition | Fragmented, sleepy incumbents | One dominant, fast player |
3. Gather, then weigh
Collect the evidence, but don't treat every data point equally. A single conversation with a real user often outweighs a glossy industry report.
4. Write the conclusion as a bet
I end every analysis with an explicit, falsifiable claim and a confidence level. "I think X, at roughly 70% confidence, and here's what would change that."
That's it. The structure is boring on purpose — boring is repeatable.