When Oshi-Katsu Collabs Go Wrong: Case Studies in Random Goods, Sell-Outs and Sloppy Operations

Team up with a popular series, prepare limited-edition goods, announce it on social media. That alone doesn’t guarantee fans will be happy. The more passionate the fans around a project, the more strongly they feel any discomfort with the sales method or with how well the work was understood.

Write a failure off as “the fans were too harsh” and the same problem will come back. Drawing on the cases in our script for Episode 48, “Heaven and Hell in Oshi-Katsu Collabs,” this article looks at which designs break trust, and where they can be repaired.

How to read the cases: We give priority to what official announcements from the companies and organizers confirm, and use press coverage and our own reporting as support. “Failure” here means the design of sales, operations and explanations on the planning side. It is not a judgment on the work or the performers.

The Conclusion: Failures Happen at the Seams, Not in the Product

In an oshi-katsu collab, planning the product carefully isn’t enough.

What fans see isn’t only whether the artwork is cute. It’s one experience that runs from reservation through pickup, sharing and later inquiries.

The collabJudged as a single experience

• Work and productSettings, colors, symbols
• Product and salesRandom variants, purchase limits
• Announcement and stockDemand, sell-outs, restocks
• Sales and the venueLines, delays, guidance
• Problem and responseRefund, exchange, restock

A polished product can still fail at any seam.

Case 1: A Random Design Where You Can’t Choose Your Oshi

Rethinking a Random Sale Across 66 Variants

For a limited-time shop held in August 2025, SHIBUYA TSUTAYA, the bookstore and media retailer in Tokyo’s Shibuya district, announced in July of that year that some items would be sold as a “66-variant random.” The store said it had received many opinions and criticisms, apologized, and changed the sales unit to five groups of 12 to 14 variants each (SHIBUYA TSUTAYA’s official X, in Japanese / official event page, in Japanese). The VTuber Inugami Korone also voiced concern about the sales method on her X sub-account (her post, in Japanese). We aren’t assessing the company or the performer here. We’re looking at a sales design in which the more firmly you’ve chosen your oshi, the more attempts it takes to get one.

66

Variants in the original random sale

5

Groups after the change

12-14

Variants in each group

Shrinking the pool made an oshi easier to reach.

Source: SHIBUYA TSUTAYA event page (in Japanese)

The problem isn’t only the random format.

• Even when your oshi is already decided, you’re drawing from a pool of 66
• Several items are each random
• Trading takes a lot of time and shipping costs
• The person being supported also voiced concern about the sales method

Buying had stopped being a fun test of luck and become a burden you carry to reach your oshi.

Fix: Leave One Path Where They Can Choose

Even when you offer random products, you can combine the following options.

• Narrow the range to a group or series
• State clearly that buying a box gets you every variant
• Offer made-to-order or named-pick purchases in a separate lane
• Provide an official trading space and safe trading rules
• Check the maximum total cost of completing the set from the unit price and number of variants

Random goods are also a culture in which people enjoy chance encounters. That is exactly why it matters to separate a situation where random is your only option from one where you take part by choice.

Case 2: Treating a Sell-Out as Success

McDonald’s “Happy Set Pokémon”

McDonald’s Japan announced that it would sell the Pokémon Happy Set, its children’s meal with a toy, from August 8, 2025, and give purchasers on August 9 to 11 a limited quantity of a two-card Pokémon card set. The official notice stated that purchases were limited to five sets per person, that customers couldn’t choose the card type, and that even buying five sets didn’t guarantee collecting them all (McDonald’s Japan official announcement, in Japanese).

At a press conference by the head of Japan’s Consumer Affairs Agency, in Japanese, the commissioner said that, given reports of bulk buying for resale and of food being left behind or thrown away, the agency had asked McDonald’s Japan to improve its sales method and prevent a recurrence. What this case asks us to consider isn’t only the drawing power of a popular IP, but the sales design when a limited-edition bonus becomes the purpose instead of the product itself.

A sell-out may mean the sales goal was met, but it isn’t a successful experience if any of the following happen.

• It doesn’t reach the children and fans it was really meant for
• Lines and inquiries pile up at stores and in their neighborhoods
• Resale prices become the way in
• Food or packaging is thrown away
• Guidance differs by store, creating a sense of unfairness

Fix: Forecast the Chaos, Not Just the Demand

Based on past similar projects, the IP’s fan base, resale prices of the bonus and how fast social media reacts, build an overheating scenario separate from the regular sales forecast.

1. Consider measures against repeat purchases across stores and ordering channels
2. Consider lotteries, pre-orders and later delivery, not just first-come sales in store
3. Consult flea-market operators in advance about measures against listings
4. Design bonuses and sales that don’t encourage people to buy more food than they can finish
5. Decide the restock policy for sell-outs before the announcement

Simply increasing stock also increases the risk of storage and disposal. The goal isn’t a sell-out but a project in which the people who want the goods can take part without strain, and the goods are received with care.

Case 3: Supervision by the Source Alone Doesn’t Protect On-Site Quality

In a collab, a single error in the colors, shapes, names or relationships that fans have followed for years can read as a lack of understanding of the work. At the same time, even when dishes and products have passed supervision, that doesn’t automatically guarantee quality on site, including serving and guidance.

For the 2025 “Pretty Guardian Sailor Moon CAFE,” built around the long-running magical-girl series, the organizers acknowledged delays in serving food and drinks, flaws in entry guidance, content that differed from the advance notice and some items being sold out. They announced refunds for the affected menu items and improvement measures (Sailor Moon 30th Anniversary Project official site, in Japanese). Assessments of the taste and criticism of individual designs, meanwhile, are evaluations by visitors and the press, and need to be kept separate from facts the organizers themselves acknowledged.

Fix: Don’t Make Supervision the Last Check

If a supervisor sees only the finished product, there is no time for corrections. Check at each of these stages:

• Planning: Which appeal of the work will become the experience?
• Design: Colors, names, symbolic motifs, prohibitions
• Prototype: Check the food, goods and space in physical form
• Operations: Check wait times, sell-outs and the wording for wrong deliveries
• After launch: Send fans’ comments back to the supervisor

1. 01PlanningWhich appeal becomes the experience
2. 02DesignColors, names, motifs, prohibitions
3. 03PrototypeFood, goods and space in physical form
4. 04OperationsWait times, sell-outs, wrong deliveries
5. 05After launchFans' comments go back to supervisors

Each stage catches a different kind of mistake.

You can’t use fans as a substitute for supervision. But a way to test the experience through the eyes of core fans before it goes public can surface the discomfort the production side missed.

Issue 4: Treating On-Site Operations as Outside the Plan

At collab cafes and pop-ups, it isn’t only the work and the products. Entry guidance, serving time, how staff explain things and how stockouts are handled are all part of the IP experience.

For example: no estimate of the wait even though reservations are required, sell-out information that never reaches the floor, or different staff giving different answers to the same inquiry. These look like backstage problems, but fans can read them as “my oshi isn’t being treated with care.”

Fix: Rehearse the Busiest Day

• Reserved guests arrive earlier than scheduled
• Dishes or products run out
• Payment or systems go down
• A wrong delivery or defective bonus is discovered
• Misinformation spreads on social media
• Individual accommodation is needed, such as for wheelchair users or families with small children

Go beyond the everyday service manual. Imagine the most chaotic day, and decide who makes decisions and where announcements go out. When you plan a collab cafe, operations design needs to carry the same weight as menu development.

Issue 5: Vague Explanations After a Problem Occurs

An apology like “we’re sorry to have made you uncomfortable” doesn’t tell anyone what will change. Even a quick apology brings more inquiries if the people affected and the actions are unclear.

Six Items to Show in the First Response

1. Confirmed facts — What happened, when and where
2. Range of impact — The affected products, dates, stores and purchasers
3. Current measures — Sales suspended, announcement corrected, under investigation and so on
4. Options for users — Refunds, exchanges, restocks and how to inquire
5. Time of the next update — When you will issue a follow-up
6. Prevention of recurrence — What you will change in supervision, stock, purchase limits and operations

Rather than staying silent until everything is settled, show it in stages: “this much is confirmed” and “we will update again at this time.” If there are errors, keep a record of the corrections.

1. 01FactsWhat happened, when, where
2. 02ImpactProducts, dates, stores, buyers
3. 03MeasuresSales stopped, corrected, investigating
4. 04OptionsRefund, exchange, restock, contact
5. 05Next updateWhen the follow-up comes
6. 06PreventionWhat changes in the future

Say what is known, and when you will say more.

What Separates Success From Failure: Turning Respect for Fans Into Actions

In the original script for the episode, we sorted what divides good outcomes from bad ones in oshi-katsu collabs into three points.

1. Is the understanding of the work visible in the details?
2. Is it designed for participation rather than exploitation?
3. After a problem occurs, can you make concrete fixes quickly?

A philosophy of “we value our fans” can’t be judged on its own. It comes across only once it becomes action: ways of buying that let people choose, participation from outside the big cities, guidance on wait times, restocks, refunds and inquiry handling.

This ties in with the design of worldview, participation, safety and belonging introduced in oshi-katsu marketing. An IP’s name recognition creates the entrance, but it doesn’t guarantee the experience through to the exit.

Pre-Planning Checklist

The Work and the Person

• Did you confirm official settings, colors, names and relationships with the supervisor?
• Is there a reason behind combining the work’s context with the product and venue?
• Are the conditions ones the person being supported can recommend to their fans?

Sales

• Are the number of variants, unit price and purchase limit a combination people can accept?
• Is there a way to get the item for sure, such as named-pick purchase, boxes or made-to-order?
• Did you plan countermeasures on the premise that resale, hoarding and waste can occur?
• Did you decide the restock and refund policy for sell-outs?

On Site

• Did you estimate wait times and staffing at maximum crowding?
• Did you standardize the guidance for stockouts, delays and wrong deliveries?
• Did you assign an owner for inquiries and set answer deadlines?

After a Problem Occurs

• Is it settled who decides to stop sales?
• Can you update the official site and social media at the same time?
• Can you show concrete remedies such as refunds and exchanges to those affected?
• After a fix, can you publish what you changed?

Notes From The Oshi Lab: The Stronger the Passion, the Larger Unfairness Looks

Oshi-katsu marketing failures don’t happen because fan enthusiasm was weak. It’s because enthusiasm is strong that the unfairness of being unable to get an item, misunderstandings of the work and pressure to pay in feel so large.

At the lab, we don’t treat avoiding online backlash (enjou) as the whole goal. When a concern is raised, do you process fans’ voices as “unexpected complaints,” or take them as data on where the design can improve? That attitude separates companies that end with a one-off collab from companies people look forward to seeing again.

Summary

What the failure cases teach isn’t how to pick a popular IP. It’s the operations that support the project.

• For many-variant, high-priced random items, leave a path where people can choose
• Look beyond the sell-out to rates of getting the item, resale, waste and burden on site
• Bring in the work’s supervisors from the planning stage, not just before it’s finished
• Rehearse the floor, including crowding, stockouts and system failures
• When a problem occurs, show concrete relief along with the apology

Don’t borrow fans’ love as a tool to draw crowds. Return it as an experience they can join with peace of mind. That is the foundation of oshi-katsu marketing that doesn’t repeat its failures.