Humans need better sizing.

A great deal of sizing technology today relies on trying to predict the measurements of a human body. I argue that this is, in fact, a misidentification of the underlying problem.

First, attempting to accurately infer body measurements from a handful of inputs such as height, weight, age, or photos is inherently difficult and often inaccurate.

Even if there were a perfect black box capable of inferring every measurement of a human body from some combination of input data, it still would not solve the fit problem in the long term. In fact, such a black box arguably already exists: a human tailor can measure a body, and 3D computer vision models, a field I have personally spent time working in, are becoming increasingly capable of doing the same.

The problem is that human bodies are not static.

The average human body undergoes many compositional changes over relatively short periods of time. Most simply, people frequently gain and lose weight; an athlete, for example, spends different periods aggressively bulking or cutting. A developing child, a patient taking medication to gain or lose weight, and patients recovering from an accident all experience significant changes in body composition. The list goes on.

Returns are always a net loss for consumers and retailers. American consumers pay an average of $8 to $12 in restocking costs per returned item and return 25% of online apparel purchases. Retailers face a harsher punishment: roughly 15.8% of annual retail sales are returned, and size-influenced returns reduce the probability of repeat purchases (and thus LTV). Again, the list goes on.

Sizing extends itself into many other human problems: consumer problems such as ordering glasses too thin for your face, rings too tight for your finger, ordering and shipping shoes from overseas that after 4 weeks of waiting, turn out to be inches too short to fit, or traveling to Japan as an American and not being able to interpret size labels. Again, the list goes on.

I believe that fit-based returns are inherently an expectation-mismatch problem. Snapshots, whether by prediction or by measurement, are not sufficient in the long term.

OpenTailor is taking a novel approach to this problem. Humans don't know their measurements. They know how things they've worn felt. OpenTailor builds a persistent memory of those experiences and reasons relative to them, and early data shows meaningfully that buyers using OpenTailor can better visualize and be more confident in what they buy and experience a near-elimination in returns and dissatisfaction on the basis of fit. Now, OpenTailor exists so you never have to check your size again, and you never have to go through the hassle of returning an item after. Over time, OpenTailor will become the infrastructure layer for the human-wearable industry.

OpenTailor is onboarding its founding team. If you want to be at the forefront of this change, reach out to info@opentailor.com or send a DM. We want to speak with retailers, manufacturers, tailors, engineers, and anyone building for the human-wearable supply chain.

Humans, and agents acting on their behalf, need sizing that evolves with them.

Glossary

Sizing is the process of determining which variant of a wearable product a person should purchase in order to achieve their intended fit.