JapanRetail and e-commerce2021 – ongoingCainz
Demand forecasting and automated model improvement
Cainz cut long-term forecast error by 19%, Google Cloud reported
Problem
Ordering depended on individual experience and spreadsheets. Manual model tuning later reached a limit.
Approach
Cainz began its forecasting project in November 2021 and started production use in June 2022. It expanded Vertex AI Forecasting across all stores by October 2025. The team later used AlphaEvolve to search for better algorithms and introduced the resulting long-term model into production.
Results
Takeaway
The team automated model improvement after repeated manual tuning had stalled. Engineers still designed the evaluation logic and production integration.
Categories
- Size
- Large enterprise
- Industry
- Retail and e-commerce
- Business function
- Logistics and supply chain
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and MLOptimization
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Google Cloud described Cainz's forecasting rollout and subsequent algorithm search with AlphaEvolve. It reported a 19% reduction in long-term forecast error against the previously tuned model.
Full details are in the original sources.
