1. Conclusion: if an ordinary-looking plate sells a million meals, that is powerful business
Sometimes a collaboration menu looks suspiciously familiar. No character face. No elaborate recreation of a scene. The food looks like a normal family-restaurant dish, while the name sounds like a finishing move.
As a customer, that can feel underwhelming. From an operator’s perspective, it can be excellent. If a chain can keep most of its normal kitchen, ingredients and procedures intact while using a popular character to generate extra orders, the economics can be very attractive.
Joyfull officially reported that its 2024 Demon Slayer collaboration surpassed one million collaboration-menu meals sold. In 2026 it returned with 13 menu items nationwide, and the third and fourth campaign phases kept six popular earlier dishes while adding seven new ones.[1][2]
So “low character density means it will not sell” is clearly false. In fact, selling a million meals without manufacturing character-shaped konjac may be the more elegant business model.
2. There are two broad models: the Joyfull model and the PARCO model
A rough way to classify collaboration food is to split it into two directions.
The Joyfull model starts with familiar chain-restaurant dishes, then adds character names, attack names, colors, toppings and limited bonuses. Hide the menu title and the plate may simply look like normal comfort food.
The PARCO model pushes the food itself toward an exhibit. Faces, colors, props, story scenes and references become part of the plating. Even with the title hidden, you can often tell it is collaboration food. Then you look at the price and your wallet abruptly returns to reality.
PARCO’s 2026 Chiikawa movie collaboration café offered 16 items inspired by meals and scenes from the story. At the Nagoya venue, ramen, curry and shaved ice were each ¥1,870, while a tropical drink was ¥1,296.[5][6]
In short:
- Joyfull model: the food stays on normal duty; the name uses the special move
- PARCO model: the food becomes the character; paying becomes part of the event
2.1 “Water Surface Slash” — where is the water?
Joyfull’s “Tanjiro Flame Dance Hamburg Steak & Giyu Water Surface Slash Grilled Beef” makes the contrast unusually clear. The official description says Tanjiro’s slashing flames are represented with cheddar cheese, garlic sauce and shredded chili. Giyu’s side, however, is described as thin-sliced grilled beef meant to evoke Water Surface Slash. The official text does not describe a water-themed sauce, blue coloring, ripples or a similar water motif.[12]
So the Tanjiro side has a concrete mapping — flame to visible ingredients — while the Giyu side leans much harder on association: thin slices, therefore something slash-like, therefore Water Surface Slash.
“Which part is Water Surface Slash? Where is the water?”
That joke surviving intact is exactly the point. The restaurant does not need to redesign the beef into something watery if the technique name and character name already carry the meaning. The flame is implemented on the plate; the water is implemented in the name.
3. In the Joyfull model, “few changes” is itself a competitive advantage
As of September 8, 2026, Joyfull’s regular chicken nanban set was ¥955 including tax. The collaboration version themed around Tengen Uzui was ¥1,098.[3][4]
The dishes are not identical, but they sit very close in operational category. Other collaboration items also use familiar chain-restaurant territories such as hamburg steak, grilled beef, pork rice bowls, Napolitan pasta and pancakes.[2][4]
That matters because the dangerous part of a new menu is not the creative name. It is everything the name can force the operation to do:
- new dedicated ingredients
- extra refrigerator and warehouse space
- additional prep
- extra plating steps
- more staff training
- leftover special inventory if demand is wrong
At national scale, one “cute little detail” on a plate can become a logistics and training project multiplied across hundreds of stores.
Joyfull instead can push much of the collaboration value into naming, visuals and bonuses. The 2026 campaign prepared 330,000 clear files and 370,000 acrylic key holders for customers ordering collaboration items.[1]
Rather than heavily modifying the kitchen, put the fandom value into paper and acrylic. The kitchen stays calm. Fans go home with something limited. Efficient.
4. Add one character-shaped konjac and suddenly you are doing manufacturing
Imagine deciding that ordinary konjac is not distinctive enough, so you want a small piece shaped like a character’s face.
To the customer, it is one cute bite. To the supply chain, the conversation changes.
There are real examples of food-processing molds designed specifically to shape konjac. Konjac contract manufacturers also offer development and production across multiple special forms such as blocks, balls, granules and noodles.[10][11]
Depending on the specification, a custom shape may create work around:
- shape design
- food-safe molds or processing methods
- prototyping
- mass-production consistency
- packaging, transport and storage
- compatibility with restaurant cooking
- leftover dedicated inventory after the campaign
This does not mean every custom food requires a new factory. Existing equipment, dedicated molds or outside contract production may be enough. But the customized piece creates tasks that did not exist when the chain used the normal ingredient.
Customer: “This character konjac is adorable!”
Product development: “Thank you.”
Procurement, logistics, quality assurance and production engineering: “Who brought this project in?”
One tiny konjac can reveal an entire org chart.
5. PARCO can charge more because it sells “time inside the work,” not just food
The PARCO model accepts much more of that complexity.
The 2026 Chiikawa movie café recreated story meals and scenes, while an adjacent shop sold original goods. A spoon was ¥1,540, a salt-and-pepper container ¥3,080, a tray ¥3,300 and a T-shirt ¥3,850.[5][7]
So this is not merely a business selling ¥1,870 ramen.
The customer is buying a bundle:
- see food modeled after the characters
- take photos
- receive limited bonuses
- talk about the story
- browse goods
- take home the memory of having been there
Historically, PARCO collaboration cafés have proven they can draw serious traffic: the 2014 Kiki & Lala café attracted 70,000 visitors at one location, and the 2017 Minions cafés drew more than 130,000 across four locations.[8]
So the model is not “too expensive, therefore nobody comes.” It is a higher-priced experience model that justifies the premium by pushing the intellectual property deeper into the physical experience.
6. Which model makes more profit?
Public information does not let us settle that question. Neither side discloses campaign-level food cost, licensing cost, labor allocation and operating profit in enough detail.
But the structures can be compared.
| Factor | Joyfull model | PARCO model |
|---|---|---|
| Dedicated food design | Easier to keep low | Higher |
| Reuse of existing ingredients/processes | Easier | Sometimes harder |
| Scale of locations | National chain potential | Mostly limited venues |
| Spend per guest | Relatively lower | Easier to raise |
| Goods sales | Supporting role | Can be a major pillar |
| Time per seat | Closer to normal dining | Often longer |
| Character fidelity | Low to medium | High |
| Operational change | Easier to minimize | Often larger |
If the question is revenue per visitor, PARCO has a strong setup because food, drinks and goods can stack.
If the question is how many meals can be sold per unit of extra operational pain, Joyfull looks formidable. If a chain can avoid major equipment changes, deploy nationally and reach million-meal scale, the return on development complexity can be very strong.
We cannot say which has the higher absolute margin. But it is useful to see Joyfull’s lower character fidelity not simply as a weakness, but as a design choice that protects scalability.
7. “Mostly a name” does not mean licensing work is trivial
There is one important correction. You cannot simply put a famous character’s name on an ordinary dish and call the project finished.
Typical intellectual-property licensing can involve proposal review, permission and contracts, supervision of designs and promotions, production approval and final checks. KADOKAWA’s public licensing guide also describes a general merchandise royalty model based on list price, royalty rate and production quantity.[9]
That does not reveal the terms of Joyfull’s Demon Slayer agreement. It only shows that small physical changes in the food do not imply casual rights management.
A better description is not:
“food that casually borrowed a character name,”
but:
“properly licensed food designed to minimize kitchen-side change.”
“Name-license food” is a fair customer joke. Behind the joke, contracts and supervision still exist. Even the flashy name has an owner.
8. Final conclusion: what looks thin to the customer may be deliberate product design
As a fan experience, the PARCO model is easier to understand. You look at the plate and immediately think, “That is the character.” It is expensive, but the reason for the premium is visible.
The Joyfull model invites the opposite reaction: “Isn’t this pretty close to the normal menu?” Yet that similarity may be exactly what a national chain needs.
Do not break the existing kitchen. Do not break logistics. Do not drown staff in new procedures. Do not overstock dedicated ingredients. Still use names and limited goods to increase orders.
Customer:
“Did they basically just borrow the name?”
Management:
“If that sells a million meals, what exactly is the problem?”
Then someone in product development says, “To make it more authentic, what if we manufacture custom character-face konjac—”
And a voice rises from the back of the kitchen:
“Stop. That konjac is bringing capital expenditure with it.”
Collaboration food is where character love and cost accounting share the same plate.
