Sushiro Case Study: Plate-Level Data That Cut Waste by 75%
Sushiro case study: the conveyor-belt sushi chain uses plate-level data to forecast what sushi to put on the belt 1 minute and 15 minutes ahead. Waste fell by about 75%, to roughly a quarter of the previous level. Sushiro installed the system in all its stores in 2002, before "big data" or "AI" were common terms.
Want your team to use frontline data for decisions like this? See our AI Workshop for organizations.
Key numbers
About 75% less sushi thrown away after the system (Nikkei, Jan 27, 2014)
1 minute and 15 minutes ahead: the forecast windows (Nikkei)
350 meters on the belt before tuna is removed automatically (Nikkei xTECH, Sep 8, 2014)
About 1 in 100 plates wasted, according to Sushiro's president in 2017 (Business Insider Japan, Sep 28, 2017)
About 50% food cost ratio, so every wasted plate is lost profit (Business Insider Japan)
The old problem: every plate starts as a guess
Conveyor sushi is made before anyone orders. Make too little and customers have nothing to pick. Make too much and plates circle until they go stale and are thrown away.
Decisions relied on the store manager's eye and experience. Sushiro wants customers to enjoy picking what they want from the belt, so it has to read demand and make the right amount in time (Nikkei xTECH).
How Sushiro uses data and forecasting
1. An IC chip under every plate
The store management system installed in all stores in 2002 tags plates with IC chips, so sales are tracked item by item in real time. Nikkei xTECH describes it as the world's first single-item tracking system of its kind.
2. A freshness rule based on distance
The system tracks how far each plate travels. Past a set distance for each item, the plate is treated as no longer fresh and removed automatically. For tuna, the limit is 350 meters.
3. Forecast 1 and 15 minutes ahead
The system uses the time of day and how long customers have been seated to predict how much of each item will sell, then tells the kitchen (Business Insider Japan).
4. Managers keep their judgment
The system adds IT to the manager's intuition and experience. It does not replace them (Nikkei).
Results
Fewer stale plates, and waste fell to about a quarter of the previous level (Nikkei).
In 2017, Sushiro's president said waste was about 1 plate in 100, and controlling waste is what lets Sushiro run at a food cost ratio of about 50% (Business Insider Japan).
What executives can take from this
If you have no feedback data, start by collecting it. Sushiro started with chips under plates, not with a model.
Let the system handle simple rules, such as freshness limits, that need no human judgment.
The goal is to give every branch the thinking of your best manager.
More cases in this series: How FamilyMart uses AI for fresh food ordering and How IKEA uses AI and moved people into sales.
Sources
This article is based on public information. IM Impower was not involved in this project.


