JapanConsumer goods2021 – ongoingCoca-Cola Bottlers Japan
Machine-learning model for choosing and relocating vending machine sites
Coca-Cola Bottlers Japan used a model trained on data from about 700,000 vending machines to pick and move machine sites
Problem
Sales staff chose vending machine sites by experience, and foot traffic shifted during the COVID-19 pandemic. Machines in areas with falling demand needed to be moved.
Approach
The company trained a model on data collected from its vending machines and divided its sales area into hundreds of millions of grid cells. The model picked promising spots, including ones staff often missed, such as a back-alley community facility. Sales staff were reluctant at first, then adopted it. During the pandemic the company moved machines from declining areas to promising ones and adjusted refill frequency and placement details.
Results
Takeaway
Staff adoption took persuasion, and the article reports full use among business developers. It gives no sales figures.
Categories
- Size
- Large enterprise
- Industry
- Consumer goods
- Business function
- Marketing and salesProduction and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Press reporting
Sources
Source summaries
ITmedia reports how the bottler uses vending machine data and a model that scores grid cells to choose sites. AI use began in April 2021. The article says the tool is used by all business developers and that machines were moved during the pandemic. It gives no sales-uplift figure.
Full details are in the original sources.
