The shelf has not always looked the way it looks today. It has been rebuilt, from scratch, more than once.
Until 1916, the shelf was a counter. You told the clerk what you wanted; the clerk fetched it. You could only buy what you already knew you needed.
Then, in 1916, a grocer in Memphis, Tennessee named Clarence Saunders ripped the counters out of his store and let customers walk in, pick things up, and choose for themselves. The store was called Piggly Wiggly. His competitors thought he had lost his mind. The first year, nine stores. Within five years, 615. Within seven, 1,267 — a growth curve retail has barely matched since.
Why did it work? Because behind a counter, customers buy what they need. In front of a shelf, they buy what they want. You came in for vinegar and walked out with a new chili sauce you had never planned to buy. The shelf had let you discover a want you didn't know you had.
From Piggly Wiggly to today's AI chat interfaces, the underlying logic of the shelf has not moved an inch: be present at the moment of decision, and be chosen. The casing has changed six times.
Six shelves
Before 1916 — the counter. Narrow. It served demand: you named the need, the clerk fulfilled it.
1916 — the self-service shelf. See it, take it. The shelf's role flipped from responding to demand to generating it.
Mid-1900s — television, the first virtual shelf. You saw a cola on screen and bought it at the store. Any medium that shows you a product is a shelf.
Early 2000s — the search shelf. You type into a box; hundreds of results line up. The wall became a search field.
2010s — the algorithmic shelf. Short-video platforms stopped letting you search. They decided what you saw. Whether your product was visible depended not on its position on a wall but on whether the algorithm granted it traffic.
Mid-2020s — the AI-native shelf, now emerging. The customer no longer searches and no longer scrolls. They delegate. "Pick me a red for a family dinner tonight, budget around 500 yuan." From that sentence on, the competition is on the AI-native shelf — and in it, decision authority passes from the person to the model. Your symbol must now be readable twice: by eyes, and by machines. The S40 in the Old Vine Series can no longer be only a visual mark; it has to travel as a clean data packet — forty-year-old vines, earth-tone packaging, formal-dinner occasions. On the AI shelf, the contest shifts from grabbing eyeballs to owning the words the model recommends.
What counts as a shelf
By today's definition: the wine cooler in a supermarket is a shelf. The phone screen is a shelf. A social feed is a shelf. A livestream is a shelf. A chat with an AI assistant is a shelf. A shelf is any touchpoint where a purchase decision can happen.
The qualifier matters. Brand exposure is one thing; a shelf is only the interface where the customer can buy now or change their mind now. A billboard shows you a brand — but you cannot order from it, so it is not a shelf. A short-video feed is a shelf — you can watch and buy in the same gesture. Get this wrong and you will believe you are running shelf strategy while actually doing brand exposure. Shelf strategy governs one thing only: being chosen at the moment of decision.
And the modern shelf is more brutal than 1916's. Attention has not increased; shelves are now everywhere. What you designed on a 27-inch monitor will shrink a hundredfold on the shelf — not because it got smaller, but because a hundred other products are fighting for the same eyes.
Subtraction at the wine shelf
When the Old Vine Series went to shelf, it faced a wall of colorful labels and dense foreign text. We chose subtraction: the entire design rests on two things — the words "old vine" and the age number. Earth tones throughout. Among shouting bottles, it is the quiet one.
Every other bottle was shouting. This one stayed quiet. And quiet made people walk over.
The order of communication at a wine shelf: color at three meters, symbol at one meter, hand out at arm's length — because the neighbor doesn't have "forty years." Drop any one of the three, and the customer is gone.
A costly lesson. In an earlier project — a cultural tea — we commissioned a well-known ink painter for the packaging. The original artwork satisfied everyone. On the shelf it was unreadable: a landscape painting reads as gray fog at three meters. Customers give a shelf three seconds; your package has half a second to finish communicating. We re-launched with a minimal symbol design — same tea, new skin — and sales nearly doubled. Product unchanged. Symbol changed. Result completely different.
Changing shelves entirely. Among recent Chinese consumer brands, Saturnbird Coffee deserves study from both directions. Its coffee comes in mini cups, and its returning-recycling program turned the packaging itself into a symbol on the social shelf. Customers are not "buying coffee"; they are collecting little cups. Saturnbird's move: it took a fast-moving product and put it on the collectible's shelf. Competitors were fighting over brightness and traffic on the coffee shelf; Saturnbird switched shelves. That is not subtraction. That is jumping to a different battlefield.
Same category, different denominators. The Old Vine Series runs on time: "I drink forty-year-old vines." Guizhou Xijiu's Jiaocang 1988 Mid-Autumn edition runs on ritual: the bottle sculpted with surging-wave shoulders, glass graduating from deep-sea indigo to moon-silver, the neck ring engraved with Zhang Jiuling's line — the bright moon rises over the sea — the box opening to trigger an LED while a moon rises across the ceiling. The liquid is 1988 base spirit finished in coastal locust-wood barrels, salt-tide and osmanthus on the nose. Sold only during the full-moon window: uncork it and it is Mid-Autumn; raise the glass and it is reunion. Same category. Completely different denominator. The Old Vine does subtraction on time — standing quietly. The Xijiu Mid-Autumn does addition on ritual — the unboxing is the ceremony. Both reached number one on their shelves.
Shelf strategy by category
The subtraction strategy is not universal. Different categories face different problems at the physical shelf.
FMCG (food, beverage, daily goods): the problem is density. Hundreds of SKUs, less than a second of visual attention each. The answer is jump — in color (fluorescent orange, bright yellow) or in form (shaped bottles, sculpted packs). Old Vine chose not to jump but to sit still, because its price band buys the customer an extra second. At two yuan a bottle, nobody gives you that second.
Durables (appliances, furniture, electronics): the problem is decision length. Nobody impulse-buys a refrigerator; they compare. What the shelf needs is not color but information architecture — model, spec, price, legible down the vertical. The classic durables mistake: a package so dense that after three seconds the customer still cannot say what makes this unit different from the one beside it.
Luxury (high-end spirits, jewelry, couture): the problem is distance. Luxury does not need to be seen. It needs to be recognized. The strategy is to stand near competitors and stay restrained. Old Vine's composure works partly because of the subtraction — and partly because the neighbors are so loud. Silence is suicide on a quiet shelf. On a noisy shelf, silence is strategy.
The three e-commerce shelves
E-commerce is not one shelf. It is three, with different physics.
The search shelf (Taobao, JD). The customer types; results line up. Visibility here depends on keyword match, main-image click rate, and sales-review volume. The symbol works through the main image. Old Vine's earth tones survive it: at thumbnail size, the S40 number is the only thing still readable. This is what "screenshot-able" means in practice.
The recommendation shelf (Douyin, Xiaohongshu). The customer is not searching; the algorithm is feeding. The question is no longer "will they pick you" but "will they stop scrolling." You get three seconds. First second: curiosity — what is this? Second: recognition — I get it. Third: intent — I want to know more. Menglan's sticker works in feeds because "the cheeky prince" completes curiosity and recognition in the first second.
The livestream shelf. The strangest shelf — the symbol is not displayed, it is spoken. The host holds the product, turns it to the camera, narrates it. On this shelf the symbol migrates from visual to verbal. A host selling Old Vine does not say "notice our packaging." He says: "This vine has been growing for forty years." If you invest only in graphic design and not in language design, you have built no symbol for the livestream shelf at all.
Three laws unique to it:
- Language is the first symbol. In a livestream the customer cannot inspect the embossing or the foil. They hear words. "Forty-year-old vines" is typeset metal on the shelf and a story told with feeling on the stream.
- Trust outweighs price. On a search shelf the customer checks ratings, reviews, price. On a livestream shelf they watch the host — does this person actually know wine, or are they reading a script? Did they drink it, or were they paid? Trust transfers faster here than on any other shelf — and collapses faster. One meltdown costs more than a hundred careful designs.
- Emotion amplifies. A physical shelf shows a quiet price tag. A livestream shows a contested object: "Only three thousand bottles exist. I fought the brand for three months to get them." The same bottle stops being merchandise and becomes an opportunity someone won for you.
The algorithmic shelf's logic
Every shelf before this one let you choose your customer. The algorithmic shelf assigns customers to you.
The mechanism: push content to a few hundred people; measure completion, likes, comments, shares. High engagement — expand to thousands. Higher — to tens of thousands. Your symbol's fate on this shelf is decided by whether the first few hundred interact.
Which changes the design brief. On a physical shelf, a symbol needs to be seen. On an algorithmic shelf, a symbol needs to be interacted with — there must be an impulse to like, an itch to comment, a reason to share.
Three traits trigger it:
Discussable. "The Third Prince of Xizhimen" is a built-in argument. Someone asks why "prince." Someone explains the Xizhimen location. Someone starts ranking pandas by fame. Ask of your own symbol: can it start a small fight?
Remixable. Users caption Menglan stickers — "me at work," "Monday, 9 a.m." — and every send is an impression. The vine, too, has vintages.
Accumulative. The algorithm rewards continuity. First exposure: not remembered. Third: noticed. Fifth: interacted. If the symbol changes mid-sequence, the accumulated recognition burns.
Control and cede
The digital shelf poses a problem the physical shelf never did: hand users the right to remix, and you hand over control of how your symbol is used.
Ninety-nine percent of Menglan remixes are good-faith and expand the character's reach. There will always be the one percent you would rather not see — the aggressive thread, the crude caption, the off-brand joke. You cannot stop them. You should not try.
This is the paradox: the tighter you grip the symbol, the slower it spreads. The looser you hold it, the faster it spreads — and the larger the risk.
Three principles hold the balance:
- Guard the core persona; release the periphery. Menglan's core is "cheeky prince." As long as that survives — as long as he never turns vicious or sleazy — captions and edits pass. The round face and the smirk are the root; the root is not negotiable.
- Be a curator, not a judge. Chasing bad remixes for deletion drains the energy that amplifying good ones would earn. Officially liking and sharing the best user work beats deleting a hundred bad ones — and the algorithm rewards the positive interaction.
- Keep an official anchor. Users will play their game; you keep shipping yours. As long as a stable, high-quality official version exists, the remixes orbit a fixed reference point.
Tolerance is the floor. The ceiling is designing for it — treating user-generated content not as a hazard to be ceded but as material to be co-designed. Menglan became a cultural symbol quickly not because we finished an IP and begged people to share it, but because we released only a core persona — the symbol seed — and let users complete the expressions, the scenes, and the lines.
We call this the Symbol Co-creation Protocol:
- Lock the tone. The brand fixes the core persona: Old Vine's "settled by time," Menglan's "cheeky," Shihuji's "immortal herb." Users cannot rewrite this.
- Leave blanks. Deliberate empty space in packaging, merchandise, and assets — blank speech bubbles, swappable backgrounds, spliceable numbers. Where the user fills in, the symbol stops being theirs and becomes mine.
- Lower the barrier. Provide templates, sticker packs, brand filters — one-tap tools that let an ordinary user produce quality remixes. Lower barrier, more participants.
- Recycle into the system. Absorb the best user language and imagery back into official assets. The strongest cultural symbols are always finished by users, not by designers.
One symbol, many shelves
The customer meets you in more than one place: a Menglan sticker on Douyin, a Menglan ticket at the zoo gate, Menglan merchandise in a friend's feed. In every place, it must be the same Menglan. Not visually identical — the same persona, the same tone, the same values.
The difficulty of omnichannel consistency is not what the symbol looks like. It is what the symbol says on each shelf.
On Douyin, Menglan says cheeky. On the ticket, welcome. On the merchandise, cute. Three shelves, three messages — one persona underneath. Cheeky is the attitude; welcome is the scene; cute is the base note.
The Old Vine Series does the same. On the wine shelf it says forty years. On the e-commerce home page, the taste of time. In a livestream, what this vine has been through. Different information, one core value: quality settled by time.
Occupy the shelf territory
The strong brands do not sell one at a time. They land as a series and fill the territory. Old Vine ships S10, S15, and S40 side by side — three numbers telling one story. Menglan appears on tickets, merchandise, installations, stickers, feeds — occupying every touchpoint it can reach.
The shelf does not care how good your product is. It cares about two things:
That eyes find you first. And that minds understand you in a second.