Product Recommendation

  HomeFusion Informatics   : :   Retail - Case Study   : :   Product Recommendation

Project Info

About

About Client

A leading retail brand based in Iran. The client offers a wide range of products over 500 SKUs.

The objective of client is to reduce the sales representative’s time while they visit the store & place the order and would like to accelerate the sales so that they can improve the bottom line by optimizing the supply chain.

DATABASE

PLATFORM

TOOLS

About

Technologies & Tools

Using ADF, Data Bricks and Spark we develop recommendation engine that make use of algorithms and SAP data to recommend the most relevant items to a sales representee for a particular store.

 

Strategic Approach

Customer-facing applications are mobile-based, whereas data is stored in an SAP database, therefore, we setup the data lake solution that ingests the source data from SAP to be furtheranalyzed holistically.

The data ranges from product data, retailer data, purchase order data, inventory data, deals & offers data. We categorized and segmented the top 50 SKU for each retail store and mapped deals & offers with each of them.This helped to easily access the top 50 SKU by the Sales Rep while placing the order.

We further cross-linked the recommended products for upselling and cross-selling the items. Apart from this based on the past purchase history, we applied predictive analytics algorithms to its data and used machine learning to optimize its insights overtime for the sales forecast and inventory planning.

Challenges

  • High order processing time.
  • Poor product discovery from 500 SKU to place the order
  • Inefficient ordering system which is taking a
  • High time to visit retail shops
  • Stagnant Sales
  • Fierce competition
  • Inefficient inventory management and production planning
  • The client had data of over 7 years in SAP that was not analyzed due to lack of AI, ML technology, and infrastructure

Basic Impact

  • Sales rep order processing time was reduced by 30% hence, due to which overall sales witnessed a spike of 6% growth.
  • The recommendation system helped retailers to recommend relevant products which enabled company to improve production planning and delivery.
  • Improved net promoter score (NPS) by 2%
  • The bottom line of the company improved significantly due increased visibility in inventory and better forecasting needs