Observing the Terrain

The English landscape gardener Lancelot Brown earned the nickname "Capability" because he had a way of looking at a client's estate and seeing its latent potential. He didn't impose a garden on the land. He read what was already there — the ridgelines, the water sources, the natural clearings — and built to reveal it.

Growth engineering works the same way. Before you run a single experiment, you need to understand the terrain. Where does traffic already want to go? What funnels are almost working but held back by one friction point? Which channels have early signal that nobody's cultivated yet?

This is the work that comes before the spreadsheet. And most teams skip it.

What the survey looks like

A terrain survey is not an audit. An audit tells you what's broken. A survey tells you what's possible. The difference is the frame of reference: broken relative to what? Possible relative to what the land can support.

The survey questions I return to:

• Where does the existing traffic concentrate naturally, before any optimization?
• Which acquisition channels have the highest intent signal per dollar?
• What does the activation funnel look like by cohort, not by average?

You won't find the high-leverage interventions in a dashboard. You find them by reading the grain of the terrain — the kind of work covered in the shed.

Capability, not just competence

Brown's insight wasn't just aesthetic. It was economic. Working with the land's natural contours required less labor, less material, and less maintenance than imposing a formal French garden on unsuitable terrain.

The same logic applies here. A channel that already has intent signal scales cheaply. A channel you're forcing against its nature will always be expensive and brittle.

Find the natural contours first. Then build. (See also: this article for a live wikilink test.)