AI in retail: how technology boosts margins and transforms business processes

Client: a major FMCG retailer having more than 1,000 stores.

Previously, the company’s category managers and procurement team relied on manual price management using fixed rules in Excel files. This approach was inefficient and did not enable them to quickly and flexibly adapt to market changes. It lacked the capability to personalize prices across stores factoring in local demand, the product’s role in the assortment matrix and the market share.

The primary objective of the project was to increase both margin and retail sales by implementing an automated pricing system.

BaOne solution

A machine learning (ML)-powered pricing system was developed and implemented, fully tailored to the client’s specific needs. The project involved the following tasks:

1. Automated pricing factoring in:

• Product’s role in a specific store
• Product perception by local customers
• Market share and competitive environment

2. Integrating the price management interface for the category management department

3. Phase-by-phase implementation – the project was delivered in three eight-month phases:

• Developing a methodology with optimization to drive margin growth and testing on a limited number of categories (hypothesis validation through A/B testing with a confidence level of 95%)
• Developing a product and putting it into operation
• Adding optimization for retail sales and running local promotions

Project specifics

• Price differentiation for different stores depending on local factors
• Using ML for demand forecasting and dynamic pricing
• Integration with electronic shelf labels for prompt data updates
• Pilot testing on a small sample with subsequent scaling

Project deliverables

• Increase in front margin (without a decrease in sales):

- By 0.9 p.p. for regular pricing
- By 0.5 p.p. for local promotions

• Automation of 80% of pricing processes
• Prices sent to stores in real time as a result of integration with the ERP system

As a result of the project, the client shifted from manual price management to data-driven approaches, thus improving key metrics (margin, retail sales) and becoming more competitive in the market. The solution is scalable and tailored to the specifics of a large retail network.

Client: a major retailer specializing in children’s products, operating more than 1,000 stores across 442 cities.

The primary objective of the project was to automate and optimize the sales management process for clearance and repricing, discontinued and expiring items, as well as to boost margin while maintaining a controlled pace of sales.

BaOne solution

• The pilot was implemented across 30 chain stores and tested on selected product categories (45,000 SKUs)
• A cloud-based AI-powered B1 Pricing system was deployed to optimize sales management
• During the pilot, over 400 million aggregated sales records from several years were processed

Project deliverables

• A 4.6 p.p. increase in margin compared to the control group, while maintaining the same sales velocity
• The solution scaled across the entire chain of 1,123 stores

The sales management process was automated and optimized with the AI-powered B1 Pricing cloud model. Algorithms analyze data on shelf life, inventory levels and demand to generate personalized promotion campaigns.

Client: a major food retailer operating over 900 supermarkets, with an active customer base of 13+ million customers and 100 million transactions.

With its huge customer base, the retailer faced challenges in providing personalized product recommendations, assessing performance of personalized promotions and managing the assortment matrix through targeted product offers to specific customers.

The project’s primary objective was to increase margin by 1-2 p.p. and the average number of purchases by at least 20%, as well as to expand the active product matrix by 10%.

Solution
We developed and deployed in the client’s mobile app a hybrid recommendation model based on neural network embeddings and collaborative filtering to:

• Leverage personal purchase histories and customer preferences, even with very large customer bases
• Ensure continuous retraining on various user signals, such as purchases, searches, comparisons, cart activity, etc.
• Predict future purchases based on purchase frequency

Deliverables

• A 2% increase in front margin through strategic category promotion (PL – private labels, GE – global exporters)
• A 2.5 increase in add-to-cart rates as a result of considering customers’ behavioral and historical context
• A 10% increase in activity across product matrix categories

The project enabled management of personal offers through customized channels, such as email, mobile apps, push, SMS, etc. A functionality for flexible segment management and user activation through RFM-based offers was developed.