A large electronics retailer wanted a solution to find out which of their products are compatible with which other products. They had a process of carrying this out manually but besides being cumbersome, this process of manually mapping compatible was also very slow and could not scale to their catalog size.
They wanted us to develop a solution with the following features:
- Ingest multiple sources of information like product descriptions, product specifications, customer reviews and Q&A.
- Ability to draw out compatibility and tag the right products automatically on periodic basis
- Determine compatibility across newer models of devices and the old ones.
Rare Mile Solution
We designed and developed an automated solution using our text analytics engine and IP. The delivered solution had the following features:
- Extracted information from a variety of sources like product reviews, descriptions, specifications and Q&A.
- Extrapolated compatibility across devices to find out new compatibilities based on existing compatibilities to cover scenarios where Camera A is compatibile with Battery B and Camera C is compatible with Battery D. If B is a replacement of D then extend B's compatibility to D and C's compatibility to A.
- During the first pass, the solution tripled the existing compatibility mappings based on new analytics.
- Since the solution was based on our iGimlet text analytics and rules engine, it was extended to a multiple product and accessories categories.
- Created an ability to increase the compatibility mapping thereby offering more cross sell and up sell opportunities for the customer
- Developed algorithms to learn from existing compatibility relationships.
- The solution was run as a managed service for the customer where it ran in the cloud and provided enriched catalog to the customer's online applications
Development of a learning system to detect compatibility between devices
- Client: US Based Electronics Retailer
- Execution Time: 3 months
- Category: Text Mining & Analytics