Notes from the body data layer.

Writing on fit, geometry, and what commerce needs from a body once machines start doing the buying.

Shoes are designed for a foot that a third of people don't have.

Every last describes the same smooth, symmetrical foot. Published prevalence data says close to a third of European adults have a structural feature it does not anticipate — and the consequence is not only discomfort. It is shoes that fail early, and get replaced.

Read →

The return shipment is the small story.

Our emissions modelling put a footwear return at 0.33–0.47 kg CO₂e — with click and collect the worst, not the best. But a pair of shoes carries twenty to thirty times that. Which means the whole industry, us included, has been optimising the smaller lever.

Read →

Fit is an inverse problem.

What you feel is the gap between your foot and the inside of the shoe — and in e-commerce neither surface is known. On the four kinds of indirect evidence, and why a fit answer without an error bar is not an answer.

Read →

The agent knows what the return costs. It doesn't know if the shoe will fit.

A short guide to the agentic commerce stack — MCP, UCP, ACP, AP2 — what each protocol actually does, and the one layer nobody has built.

Read →

You're a 39 in one brand and a 40 in another. Your foot never changed.

Every shoe is built around a physical form called a last. It decides the fit — not the number on the box. On grading, on why "runs small" tells you almost nothing, and on the question the industry has been asking wrong for a century.

Read →