How we choose the styles we sell

A style in our catalogue is not a mood board. It is one set of decisions — what you are wearing, what is behind you, where the light comes from, and how tightly the frame closes — applied identically to everybody who buys it. We wrote eighteen of those sets, and this is an account of how, including the two we got wrong.

Four things constrained every one of them, and not one of the four is a question of what looks nice. They are worth stating in order, because most of the decisions further down are really about which of the four we agreed to give up:

Written down after the fact, the working rules are shorter than the catalogue suggests. They are not original — any portrait photographer would tell you the same things — but having to apply them without ever seeing the subject made us more literal about them than a photographer in a room would need to be.

It is the only one of the five whose reason is physics rather than judgement, and the one most often stated badly: avoid patterns is not the rule. That argument is long enough to have its own page — what actually happens to a pattern in a photograph — and it shaped these wardrobe lines more than anything else did. For clothing decisions with no camera involved, the dressing guide is the broader treatment.

Navy suit, white shirt, mid-grey wall, soft light from one side. It is the highest-contrast and lowest-risk combination in the catalogue, and it was the first one written, because a catalogue needs a floor before it needs a ceiling.

The arithmetic is why it works, not the taste. Navy is dark enough to separate from almost any skin tone and light enough to separate from mid-grey. The white shirt puts a bright wedge under the chin that survives being shrunk. The grey wall neither swallows dark hair nor blows out beside a pale face. Nothing else we ship fails for fewer people.

The cost is that it is unmemorable, and we chose that knowingly. A photograph nobody could object to is also one nobody remembers, and if every style were built this way the catalogue would be a single picture repeated. The Corporate style page shows the result. Everything below it gives something up on purpose.

We tell people to avoid black, and then we sell a look built on a dark fitted top. That is not an inconsistency we failed to notice. It is the clearest evidence we have that these rules are about conditions rather than about garments.

Black fails in an ordinary photograph because the room is usually darker than the person, so the shoulders sink into it and what is left is a face with nothing underneath. Every part of that failure is environmental. Fix the environment and the garment stops being the problem: put a mid-tone paper backdrop behind the shoulders so the dark top has something to sit against, then light hard from one side so a bright edge runs down the shoulder line.

Neither condition exists in a phone photograph taken in a living room, which is exactly why avoiding black is right for most people most of the time and wrong here. The colour post is the general form of that argument. The specific form is that we only put somebody in black where we also control the wall and the light.

The first draft of this style had a jacket in it. That was wrong, and it was wrong for the same reason a suit is wrong on the person the style is meant for: formality is not a scale where more of it is safer.

A photograph is read against the expectations of the field it will be seen in. In technology, product and agency work, a suit in a head-and-shoulders frame does not read as smart. It reads as borrowed — taken for an occasion rather than for a job, which is the impression this style exists to avoid.

Landing on a crew-neck knit took longer than choosing the background did. A t-shirt is underdressed. An open shirt collar is close, but claims a slightly older register. A knit in oat or charcoal turned out to be the one garment that is unmistakably not formalwear and equally unmistakably deliberate — the difference between casual and careless, carried almost entirely by the weight of the cloth.

The obvious way to make a photograph approachable is to ask for a bigger smile. It does not work — or rather it works for a fraction of a second, and then reads as effort. The parts of a picture that actually carry warmth are the ones nobody consciously notices.

Two of them do most of the work. Soft frontal light with almost no shadow removes the modelling that makes a face look serious: shadow under the brow and beside the nose reads as gravity, and taking it out reads as openness. A light layer over a plain tee then sits a register below the same jacket over a shirt, lowering the formality without lowering the care.

The bright off-white background is doing a third job that has nothing to do with warmth. It is the highest-contrast backdrop we ship for anyone with dark hair, which is why this is the style we would choose if the picture were only ever going to be seen small. Approachability and legibility turned out to want the same photograph, which does not happen often.

The professional styles all answer a single question — what should a stranger who is evaluating you see. The Social pack came out of noticing that most people's faces are in several places at once, and that only one of those places contains a stranger.

Written down, the differences are not about quality. A work chat puts your picture beside every message you send, in front of people who already know you, so an over-polished portrait reads as trying too hard beside colleagues who used a phone camera. A messenger shows it smaller than anything else on the device, so the crop tightens and the background collapses to one saturated colour. A press page sets it next to a headline, where blending in is the failure mode, so the light hardens and the camera drops below eye level.

The honest caveat is that this is both a real distinction and a convenient one for us, since it is a second purchase. So the test we hold it to is whether any two of the six are interchangeable at the size they are actually seen. If they were, it would be padding. The pack page lists all six, and the argument stands only while they stay distinguishable.

The first was arithmetic. Our catalogue contains styles that no professional pack has ever included, and for a while a pricing surface counted the whole catalogue when describing what a pack contained. The most expensive pack was advertised as twelve of eighteen looks, which made the top of our range sound like a partial version of something else. It was a page counting the wrong list, and the count now comes from the professional catalogue alone.

The second was more substantial, because it was about what we were charging for. For a long time the larger packs gave you more shots of each style rather than more styles, and we called that more choice. It was not. Each extra shot repeated the same wardrobe, the same wall and the same light: another roll of the same dice. Somebody buying three of one look was buying redundancy and being told it was variety.

What changed is the axis we sell on. The largest pack now buys more looks rather than more copies of each, and the work in progress is giving every style a few genuinely different setups, so a second shot of a look differs in something a buyer would have chosen between. Until that lands, the honest description of shots per style is insurance against a bad frame, not a wider selection.

None of the reasoning above is worth much if the pictures do not hold up, which is why the useful version of this page is not the page. Every style at full size is the evidence, and each one carries the four lines it was built from — wardrobe, background, lighting, framing — so a decision can be checked against the photograph rather than our account of it.

Where the two disagree, believe the photograph. Several of these changed after we looked at what came back, and the four lines get updated when that happens rather than left standing as the original intention. A style whose description and whose sample tell different stories is a bug, and one we would rather hear about than defend.