DESIGN INTELLIGENCEJuly 22, 2026·Mary · DEPIX Design Intelligence

You Don't Draw an Emoji, You Design a Meaning

There is no more widely used visual language on earth than emoji — billions are sent every day, across every culture and platform — and in September 2025 it gained 163 new words at once. Unicode 17.0 added a treasure chest, a distorted face, a fight cloud, ballet dancers and a bigfoot-inspired hairy creature, among others, and they are landing on iPhone and Android through 2026. But here is the thing almost everyone gets wrong about how an emoji is made: the people who "design" it never draw it.

Adding an emoji is a two-stage process, and the two stages are done by different people. First, the Unicode Consortium approves a concept — a code point, a name and a short description ("treasure chest"). That is all. No picture. Then each vendor — Apple, Google, Samsung — independently draws its own version of that concept. Apple's treasure chest and Google's treasure chest look nothing alike, and both are correct. The same emoji is a different picture on every device in your pocket.

Which means the design decision that actually matters is not the drawing. It is the concept. The question a new emoji has to answer isn't "how should this look?" — it is "is this idea clear and universal enough that a dozen strangers, drawing it independently and never speaking to each other, will all produce something people read the same way?" That is a far harder question, and it is settled years before any illustration exists.

This is the contrarian heart of it: the hard part of emoji design is not the illustration — it is designing a meaning robust enough to survive being redrawn by many hands and still communicate. When a concept is ambiguous, the system breaks in exactly the way you would predict. Emoji have meant different things on different platforms — a grin that read as friendly on one phone and pained on another — and in the most famous case, vendors unilaterally redrew the "pistol" as a bright green water gun, changing the meaning of a message after it was sent. Each platform's picture was fine. The concept wasn't pinned down, so the meaning drifted.

That is why Unicode's approval process looks less like an art studio and more like a concept-phase review board. Proposals are debated for years against a strict brief: is the idea frequently needed, is it distinct from what already exists, is it not so specific it will date, and — crucially — will it still read at sixteen pixels? Only concepts that clear that bar become code points. The nine candidates now waiting on Unicode's 2026 vote, part of a broader 2026–2027 draft, are going through exactly that gauntlet: not "can we draw it?" but "should this meaning exist, and is it clear enough to be drawn by everyone?" Because once it is a code point it is effectively permanent and rendered by every vendor on earth.

The lesson travels far beyond your keyboard. Any time a design will be executed by many independent hands — a design system rolled out across a hundred product teams, a brand interpreted by a dozen agencies, a franchise built by different studios — the concept is the only thing you truly control. You will not be in the room when most of the work is drawn. If the shared meaning is crisp, the independent executions cohere without anyone coordinating; if it is fuzzy, every hand drifts a little, and the drift compounds into incoherence. The coordination you can't do later has to be designed in, upstream, as clarity of concept.

Emoji are the purest proof of a principle that is easy to forget when you are staring at a beautiful render: the design is the idea, not the picture. The 2026 rollout will bring a dozen different-looking treasure chests to a dozen different phones, and it will not matter one bit that they don't match — because the real design was finished before any of them was drawn. The consortium didn't design a picture. It designed a meaning, and trusted the meaning to hold.

Sources: