Somebody is standing in a supermarket holding a carton of oat milk, and she has been reading it for eleven seconds, which is roughly ten seconds longer than the shelf was designed to hold her. Shoppers give any single item about a second before their eye moves on, and almost every rule of packaging design exists to win that second: the name set large enough to register peripherally, one benefit stated once, a hierarchy arranged so the decision is most of the way made before the shopper is consciously aware of deciding anything. The carton in her hand breaks all of it deliberately. The product name has been demoted to a supporting element, there is no calcium callout and no photograph of anything being poured, and the primary real estate on the front of the pack has been given over to a paragraph of copy.

Thirteen years of agencies have looked at that carton and copied the wrong half of it.

Stand in the alternative dairy aisle in 2012 and take it in as a single object rather than a set of products. What you get is a wall of soft green leaves, pastoral vectors of farms nobody has ever visited, and calcium percentages arranged in neat circles, with an almond frozen mid-splash into a pool of white, shot from below, on roughly every third carton. Silk did it, Almond Breeze did it, and the store brands copied whichever of them was selling that year.

The design logic underneath was reasonable enough for the market as it existed. Plant milk was still largely medical, bought by people with an intolerance or a diagnosis, so the packaging borrowed its visual grammar from pharmacy: cool palettes, clinical sans-serifs, nutrition presented as reassurance, everything arranged to make an anxious customer feel handled. It worked well enough on early adopters, and it produced an aisle in which every carton visually cancelled out the one beside it. Then plant milk went mainstream, every large corporation in dairy arrived carrying the same sterile health codes, and the shelf went dead. At three feet, the only way to tell two brands apart was to read the logo.

Oatly had been sitting in that aisle since the early nineties, invented by researchers at Lund University who worked out that oats could be turned into something you pour on cereal. By 2012 it was a fifty-person company behaving like a multinational, running an internal report that put brand recognition among Swedes at 70% when the real figure was under 5%. The product sat on shelves in a category growing quickly, and nobody could see it.

The board told its recruiters not to bring them a food-industry CEO and ended up with Toni Petersson, who had come from restaurants, a nightclub and property, and who called a creative director he had worked with for fifteen years to ask whether he fancied a go at oat milk.

Lars Elfman at Forsman & Bodenfors hand-drew the type and the images. John Schoolcraft wrote the copy. The working arrangement was that Schoolcraft would write and the agency would sign it off, which is the client-agency relationship running backwards. The logotype is the part everyone remembers, and the logotype is a container. What went into it was a copywriter handed the most valuable surface the business owned and told to use all of it.

A Tetra Brik gives you a fixed format, a set number of panels and no possibility of structural differentiation, so the only variable available to any brand in the aisle is what gets printed on the same box everyone else is using. Filling that box with dense copy should produce something that reads like a terms-and-conditions panel. It doesn't, because the text sits inside a strict structural grid with the hierarchy inverted: message promoted to hero, product name demoted to a supporting role, everything aligned tightly enough that hand-drawn type reads as attitude rather than as an error at the printers. Flat single-colour print on uncoated board did the rest. There was no media budget, so the carton had to carry the advertising, and in an aisle of glossy stock and pastel gradients, matte board with no photography on it reads as a hole in the shelf. Attention came from subtraction, in a category that had spent a decade adding.

The copy did the work the subtraction made room for. WOW NO COW on the front. YES WE ARE VEGAN SO? on the side. IT'S LIKE MILK BUT MADE FOR HUMANS across the middle. Down the back panel, where convention reserves the space for compliance information, rambling monochrome paragraphs insulting the dairy industry or describing whatever the CEO had been thinking about that month, and at one point Schoolcraft's own phone number, which people rang, mostly to ask why a Swedish product was speaking English.

Every element a shopper uses to decide in a second had been given away, and to a grocery buyer the result looked like a mistake. What Oatly bought in exchange was twenty minutes at a kitchen table, a trade that only works if you are prepared to be invisible to everyone who won't stop walking. Oatly was already invisible, so it cost nothing to find out.

By 2012 the pastoral reassurance had stopped reading as reassurance and started reading as a company working very hard to be approved of, which is what polished wellness design tends to signal once a category matures. Need is not an attractive quality in a person, and it doesn't improve on cardboard.

Somewhere in the same period, buying plant milk had quietly stopped being a medical decision and become an identity one. A corporate carton on the counter makes you look like someone managing a digestive issue, while the matte, text-heavy, faintly annoying one makes you look like you are in on something, and the pack behaves like an outsider so that the person holding it gets to be one too, for the price of a carton.

The design carries that signal but cannot generate it. A hand-drawn logotype is incapable of insulting a lobby group or making a joke about its own CEO's inner life. Only sentences do that, which is also why only sentences can get you sued.

In 2015 the Swedish dairy lobby LRF Mjölk sued over IT'S LIKE MILK BUT MADE FOR HUMANS and won. Oatly was banned from the slogan, ordered to stop calling its product milk, and fined around $126,000. It published the full text of the lawsuit as content, and sales rose by more than 45%.

Four years later Oatly applied to register the same slogan as an EU trademark and was refused for lacking distinctiveness. The General Court overturned the refusal, holding that the line challenged the accepted idea of milk as a staple of the human diet, and that the controversy it had caused in Sweden, the Netherlands and the UK showed the phrase was memorable enough to identify one company.

The sentence that lost the 2015 case became the evidence that won the trademark. A typeface cannot do that.

Legacy alternative milk brands belong to conglomerates organised around mass-market appeal and legal compliance, where packaging passes through dozens of focus groups and legal reviews before it reaches a shelf, and where a legal team is unlikely to approve a design that spends its primary panel on a three-hundred-word stream of consciousness about nothing in particular.

They came anyway. Nestlé launched Wunda in 2021, built expressly to take on Alpro and Oatly, and Danone moved at the level of the line rather than the layout. Alpro sells an oat drink under SHHH...THIS IS NOT M*LK, with a white droplet standing in for the i. Zurich's regulator ruled the labelling unlawful in 2022, and in March 2026 the Swiss Federal Supreme Court rejected Danone's appeal four votes to one, one judge remarking that if you have to call a spade a spade, you have to call milk milk.

The same kind of courtroom produced the same kind of loss a decade apart. Oatly printed its lawsuit on the internet and sold 45% more product. Danone Switzerland acknowledged the ruling and said that various options were currently being evaluated.

The moat sits in the gap between those two responses, and it is not the uncoated stock, which Danone bought, or the asterisk, which Danone also bought. What could not be bought was a person inside the building permitted to answer a court with a joke rather than a holding statement. The asterisk gives it away in a single character: a brand doing the bit while hedging the bit, which shoppers read as a corporate suit in a leather jacket long before they could explain why. The inability to be genuinely unhinged is the ultimate corporate moat.

Walk the aisle now and the almond has largely gone, replaced by uncoated stock, muted single-colour print, wordmarks drawn by hand at speed and a small joke somewhere on the front. By 2021 the observation was already common enough to be a cliché among strategists, one of whom noted that every new oat milk company looks like Oatly, which is the ordinary fate of any distinction successful enough to be worth stealing. An agency that worked with Oatly for years puts it more flatly: the house style and tone have since been widely copied inside and outside the category.

Elfman's type was distinctive in 2013 because it was one carton behaving oddly among two hundred behaving identically. Distinctiveness in packaging has never been a property of the object so much as a property of its relationship to whatever sits either side of it. In 2026 the same drawing is the house style of an entire discipline, so a shopper facing eight handmade-looking cartons is back where she started, spending her second on nothing, and the aesthetic that once bought twenty minutes now buys nobody anything.

One brand took the other half. Surreal launched into cereal in 2022 with a writer who had come from Innocent, and became the fastest-growing cereal brand in the UK inside two years against opponents with a century of shelf equity. The test its own admirers apply is the useful one: cover the visual identity entirely and you can still name the brand from a line of its copy. Nobody can do that with a hand-drawn wordmark.

The branding argument this year has moved from what a brand looks like to what it sounds like. Generative tools optimise for what has already worked, so a fast identity comes back as a competent recombination of the last decade's winners, and the same tools produce the same flattening effect on copy. The brands still being recognised are the ones whose writing could not plausibly have come out of a machine.

None of which saved the business. Oatly is worth around $296m today against the $10bn it listed at in 2021, having scaled its factories faster than it could run them, and no carton has ever fixed a manufacturing problem. The company now behaves like an ordinary FMCG operation, with a first seasonal limited edition in Hot Cocoa, a new Vanilla, and Popcorn-Flavored Barista Edition in test overseas, competing in an aisle that has become crowded and price-sensitive, where oat is the only segment still growing and growing slowly.

1. Photograph five competitor packs and list only what they have in common. Not what separates them, which is the audit every agency runs and which produces nothing. The shared list is the category convention, and it is usually four or five items long: a colour, a photographic treatment, a claim structure, a typeface class. Nobody in the category has tested any of it in a decade.

2. Pick the one item on that list the client is most frightened to lose, and lose it. Oatly gave away the second on the shelf, which was the most expensive thing it owned. The space that opens is the only place attitude can go, and it will be roughly a third of the front panel.

3. Hold the result inside a rigid grid. The difference between Oatly and a bad zine is entirely structural. Chaotic elements read as a decision when the alignment underneath them is strict, and as a printing error when it isn't.

4. Then give the front of the pack to a writer, and pay them properly. This is the step the industry has skipped since 2013. Copy is not the material that fills the space left over once the layout resolves. On a Tetra Brik, where every brand has the identical box and no structural differentiation is available to anyone, the sentences are the only asset a competitor cannot buy off the shelf.

The method works where a category has gone visually dead and every brand is fighting over the same functional attributes. It fails where precision is the product and a visual mistake signals real danger.

The almond splash was a convention nobody had ever tested. It was there because it was there, and the aisle had agreed not to ask. In 2026 the convention has moved and now arrives disguised as non-conformity, in the form of the hand-drawn mark, the lowercase wordmark and the little joke on the pack, all of it producible in an afternoon and defensible in any meeting, which is what makes it the almond.

Most creative professionals treat brand guidelines as a set of rules that keep them safe, where this treats category convention as a map of what to avoid. You trade broad demographic comfort for fanatical relevance with a smaller group, which means accepting in advance that some people will dislike your packaging. Brands that try to please everyone across every touchpoint end up being remembered by nobody.

The woman in the aisle is still reading, and thirteen years on, the only part of that carton anyone has managed to copy is the part she isn't looking at.