Optimized product distribution with the help of analytics
Machine Learning models to optimize product distribution to
stores by analyzing factors like past sales, weather and local trends.
Process
The client is a leading global retailer; the focus of the work done was for the Indian operations of the retailer. Client has approximately 2,000 stores spread across the country. Different store locations have different customers, different weather, and different capacity to stock and display products which resulted in very different needs. The initial product distribution was one size fit all for specific geographies, which resulted in under-inventory in some locations and over-inventory in others.
Solution
Artificial Intelligence (AI) is ideal for optimizing product distribution for retailers. We created an AI model for the client, which looked at a variety of factors like past sales, weather forecasts, local trends, and store display space availability. The model optimized the product distribution by sending more inventory to stores where products were being sold more. The AI model also tracked sales in real-time and was able to predict when rerouting of inventory was required from one store to another.
Challenges Addressed
Out of stocks
Store inventory was planned more on geo perspective, so caused some of the store to sell out on products, hence loosing on potential sales.
Discounts impacted profitability
Products left unsold in stores were sold off on discounts, which impacted hard-earned profits for the retailer.
Low customer satisfaction
Due to unavailability of hot selling items in some stores, customer used to leave dissatisfied or had to come back to collect the item when it became available. This adversely effected customer experience.
Manual process
Stock routing was planned manually which was prone to human errors.
Higher retail space requirement
Some locations used to be overstocked with products and hence storage space was kept occupied with inventory.
Outcome
- Increased profitability by 18%.
- Decreased unavailable product requests by 38% across the board.