# What sells isn't what's fun to generate > Everyone knows specific beats generic and almost nobody acts on it. The reason is that the people who enjoy making images and the people who buy printed clothes are two different crowds. August 17, 2026 · https://justog.club/stories/what-designs-actually-sell Of the two shirts above, one is unquestionably the better picture. Mountains, a lake, a well-handled palette, hours of craft implied. The other is a cement mixer drawn in about nine lines with a joke under it. The cement mixer sells. Not because it's better — because it belongs to somebody. ## What kinds of designs actually sell? Specific ones, and this part isn't in dispute. Every serious account of the market says the same thing and the examples repeat: *Golden Retriever Mom* outsells *Dog Mom*, a trade-specific joke outsells a generic funny slogan, a profession with an in-group outsells a mood. Buyers are reacting to communities and identities rather than to pictures. So the advice is correct, universally available, and almost nobody follows it. That's the part worth explaining, because a rule everyone agrees with and nobody applies usually means the explanation is wrong somewhere. ## Then why does everyone keep making the beautiful generic one? Because the person making it and the person buying it are two different crowds, and this is the finding that reframes everything else. The population that enjoys generating images and the population that buys printed clothing overlap far less than you'd assume. Different demographics, different motives, different taste. One is there for the act of making something appear — sci-fi, fantasy, concept art, the pleasure of the render. The other is buying a way to say who they are in a room: streetwear, subculture in-jokes, gym language, identity worn where people can see it. The uncomfortable consequence: **your enjoyment of a design is not evidence about its market.** You're in the first group. The buyer is in the second. The mountain landscape is a genuinely satisfying thing to have generated, and that satisfaction is telling you nothing. ## What is the buyer actually buying? Membership, most of the time. The image is how it gets carried. A shirt is worn in front of people, which makes it the rare purchase whose value is mostly external. The cement mixer works because a specific group reads it instantly and everyone else doesn't — and being read by the right people is the entire product. The landscape is legible to everybody and says nothing about anybody. This is why a worse drawing beats a better one so reliably here, and why "make it look more professional" is such a common and expensive wrong turn. ## How narrow is too narrow? Narrower than you think, and the old limit was economic rather than aesthetic. Niche advice has been around forever, but it used to run into a wall: you couldn't make a shirt for eleven people, because the minimum order was seventy-two and somebody had to eat the rest. So "niche" in practice meant *a niche large enough to absorb a print run*, and a great deal of received wisdom about how wide a design must be is really a memory of that constraint. Made to order removes the floor. A shirt for eleven people is a shirt. Which means the honest answer to "is this too specific?" is now almost always no — [the whole production stack that enforced breadth is gone](/stories/start-a-shirt-brand-with-ai). ## Isn't the design the product? Less than it used to be, and this is worth being clear-eyed about. Generation made competent design abundant. Anyone can produce a good-looking graphic now, which means a good-looking graphic has stopped being scarce and therefore stopped being defensible. What stayed scarce is people who already know who they're talking to. That's the actual asset — not the ability to write prompts, but an audience. A design with fifty people who'd recognise it beats a better design with none, because the first one has a route to the people it's for and the second one is waiting to be discovered by a search engine. ## So beautiful designs are pointless? No, and this is the wrong lesson to take. The landscape's problem isn't that it's beautiful. It's that it's **anonymous** — it could have come from anywhere, be for anyone, and hang on any wall. Beauty and specificity aren't opposed; the best things are both, and a gorgeously rendered in-joke is better than a crude one. Craft is worth having. It just can't do the job specificity does, and pouring more of it into a generic idea is [the most common way to polish something into nothing](/stories/why-a-wolf-gets-the-average-wolf). ## Is there a test before making it? One question: who is the person who'd see this and know it was aimed at them? If a name comes to mind — your group chat, the people at your gym, everyone who has done this particular job — the design has a reason to exist. If the honest answer is "anyone", what you've got is decoration, and decoration is competing against every other pretty thing on the internet. The test is uncomfortable on purpose, because the generic idea always feels safer. It isn't safer. It's just aimed at nobody. ## How do I get one made? Describe the idea at [JustOG](/) — including who it's for, which is the half that does the work — and the directions come back as a set. Pick one, drag the crop frame, see it composited on the real garment, and it's made to order and shipped. Designs other people have published are in the [shop](/shop). The mountain took longer and looks better. Somebody's brother-in-law is going to buy the cement mixer.