CroatiaRetail and e-commerce2021 – ongoingFortenova Group
Machine learning demand forecasts for fruit and vegetable orders in Konzum stores
Fortenova's Konzum pilot stores cut fruit and vegetable waste by 7% with a machine learning order forecast
Problem
Fortenova Group had large amounts of sales and loyalty-card data, but its on-premises systems made advanced analysis slow and hard to manage. Fruit and vegetables have a short shelf life, so wrong store orders led either to empty shelves or to unsold stock being thrown away.
Approach
In 2021 the group started a central data platform and set up a DataLab with the consultancy Artefact on Google Cloud, using BigQuery ML and Vertex AI. The DataLab built a model that reads historical sales, promotions and inventory data and produces a daily forecast for store managers placing orders. It ran first as a pilot in 65 Konzum stores in Croatia, and from mid-June 500 Konzum stores used its recommendations.
Results
Takeaway
Fortenova started with one perishable category and proved the forecast in 65 stores before it moved to 500. The story gives the waste and profit results for the pilot only.
Categories
- Size
- Large enterprise
- Industry
- Retail and e-commerce
- Business function
- Logistics and supply chain
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Partner co-development
- Evidence
- Vendor case study
Sources
Source summaries
Google Cloud's customer story describes how Fortenova Group built a DataLab with Artefact and used machine learning to forecast daily fruit and vegetable demand for Konzum stores. It reports a 7% cut in food waste and an 8% rise in profits in a 65-store pilot, and says 500 stores later used the recommendations. The page carries no date; the Internet Archive first saved it on March 8, 2024.
Full details are in the original sources.
