Affordances, or how objects tell you what to do

Shape as instruction, and where it succeeds without any label.

An object tells a hand what to do through the shape it presents rather than through any word printed on it, a quality called an affordance, and a device built for one specific person can lean on this far harder than a device built for a crowd, since it only ever has to communicate correctly with one particular hand rather than guess at thousands of unfamiliar ones.

The short version

A well-shaped control does not need a person to think about how to hold it, the correct grip simply presents itself the moment a hand meets the object, and the entire purpose of designing through shape rather than through instruction is removing that moment of thought altogether. For a device meant for one named user, this stops being a nice-to-have refinement and becomes something closer to the whole design brief, since a control that has to be relearned every single time defeats the point of building something custom in the first place, when a generic, labelled alternative would have done almost as well.

The scissor-handle comparison

A pair of scissors has two loops of noticeably different sizes, a larger one for the fingers and a smaller one for the thumb, and that asymmetry is not decorative, it is the entire mechanism by which scissors tell a hand which way round to hold them without a single word of instruction. Grab a pair of scissors the wrong way round and the mismatch is felt immediately, the thumb too cramped in the larger loop, the fingers too loose in the smaller one, long before any cutting is attempted, and the correction happens automatically because the shape itself refuses to feel right until the hand has found the one orientation it was actually built for. A device built for one named user can use exactly this trick deliberately, shaping a grip so it only settles comfortably into that person's actual hand in the one orientation the mechanism needs, rather than leaving the correct orientation to a diagram or a printed arrow.

Why a custom affordance can be more exact than a general one

A device meant for a broad, unknown market has to shape its affordances conservatively, using only the cues that will read correctly across a huge range of different hands, ages and levels of familiarity with similar tools, which rules out anything too specific to one person's particular grip. A device built for one named person carries no such constraint, since its shape only ever has to work for the one hand it was measured against, which means a designer can build in a much sharper, more specific cue, a ridge exactly where that person's index finger naturally rests, a contour that matches the curl of that particular hand at rest, than any general-purpose product could responsibly commit to.

This is also where a custom affordance can afford to be far less conventional than a mass-market one dares to be. A generic product usually has to borrow familiar shapes, a trigger that looks like other triggers, a dial that looks like other dials, because an unfamiliar shape asks an unknown user to learn something new before the product can be trusted. A device built for one person who will use it daily has no such obligation, since that person only has to learn the shape once, and an unusual but genuinely better-fitting form can earn its keep over months of use in a way a mass-market design, judged mostly on a stranger's very first encounter with it, never gets the chance to.

One figure worth keeping in mind

Testing a control's affordance usually comes down to a simple, almost embarrassingly direct measure, whether a first-time user picks it up and uses it correctly without being told anything at all, and a custom-shaped control built around one specific, measured hand can reach a first-attempt success rate close to certain, since there was never a range of different hands it had to accommodate, only the one it was actually built to fit. A generic control aimed at a broad population settles for a lower, though still respectable, success rate by comparison, simply because it is doing a harder job, communicating clearly to hands it has never met and never will.

What this changes in practice

Designing a custom affordance starts with measuring the actual hand, not a published range, watching where the fingers naturally fall at rest and building the grip around that resting position rather than around a generic ergonomic template borrowed from a product meant for everyone. It also means testing the shape against that one specific person early, since a shape that feels obvious to the designer holding it is not proof it will feel obvious to the person it was actually built for, and the only real test that matters is whether that particular hand finds the correct grip without being shown. A designer's own hand is, in fact, the least reliable judge available, since it already knows which way the part is meant to be held before it ever makes contact, and that foreknowledge is exactly what the intended user will not have on their first attempt.

What this does not explain

A strong affordance still cannot substitute for basic mechanical soundness, since a beautifully shaped grip on a mechanism that jams, binds or fails under load has only succeeded at communicating an instruction the device then fails to carry out. Shape as instruction solves the problem of a user not knowing what to do, which is a different problem entirely from the mechanism not doing what it was asked, and a device for one named person needs both solved, not merely the one that happens to be more visible on first contact. A custom affordance can also drift out of date in a way a general one rarely does, since a hand's grip strength, reach or dexterity can change over months through injury, recovery or simple ageing, and a shape measured once against a hand that has since changed slightly is no longer the exact fit it was built to be, a maintenance cost a device built for one person carries that a mass-market product, aimed at whichever hand happens to pick it up, never has to think about at all.