One who increases sales. One who makes the buyer happy. One who has the right stock and a wide range of shelf products. One who adopts a competitive price but differs from competitors.
Retail is a highly competitive market characterized by low margins and strong variability. Customer expectations constantly change. Technological, social and economic variables determine the changes.
Today retail requires fast processes, such as artificial intelligence applied to warehouses and stores, an “agile” supply chain, making better forecasts, minimizing risks, optimizing the processes and oblique decision support.
OUR SOLUTIONS FOR YOUR INDUSTRY
OPERATION & TRANSPORTATION
A logistics partner of a grocery retailer complained about a significant increase in transportation costs due to the policy requirements established by the retailer. However, the retailer needed to increase the service level without impacting their costs.
Explore the possible levers to improve the process governance, avoiding extra-costs, aiming to increase the service level.
The solution was applied in two phases:
- off-line analysis and simulation of an optimal distribution. This simulation included the optimization of the fleet scheduling and optimization of the delivery-calendars;
- roll-out of the TMS for daily management. Optimization of the delivery calendars on a seasonal basis.
- Costs reduction: 10-15%
- Reduction of the variability in terms of volume delivered per day: 30-40%
As a consequence, the proposed solution has improved fleet utilization, in particular for smaller vehicles (often a bottleneck when delivering in city centers).
DEMAND FORECASTING AND PROMO MANAGEMENT
Italian retailer with hundreds of POS and 3 distribution centers.
Support the process of promotional purchase, at distribution center level.
Improve the forecast accuracy and the number of managed product categories, with respect to the current tool.
Ublique’s promotional purchase module was installed.
The module employs Machine Learning models to cluster products by several features (e.g., category, promotion type, product attributes, …) and produce a quantity forecast, which is then input to an optimizer allowing to suggest the quantity to buy, aiming to balance overstock and stock-out probabilities.
- Forecasts and proposed purchases for all the categories
- Risk management approach
- Forecast error lowered from above 40% to under 25-30%
- Buyers’ working time almost halved