Reverse Auction for e-Commerce Portal

As a merchant, how do you create a way to use consumer requests to effectively suggest products?

Customer & Background

Our customer is a leading consulting services provider who was looking for a product engineering and solutions partner who could facilitate building analytics solutions in the areas of consumer and retail industries. One such customer scenario was related to e-commerce reverse auction domain.


Innominds developed a solution utilizing our big data and analytics practice competencies and our context aware analytics solution platform. The solution includes the following features:
E-commerce portal to provide reverse bids based on consumer’s requests.
Consumers can request any kind of product against their pre-paid amounts by inputting their own description.
Multiple merchants can fulfill consumer requests by offering products in a competitive manner.
Developed a merchant web portal and iOS consumer app.
The analytics platform directs a consumer’s request to the correct merchant by analyzing merchant profile, inventory, transactional history, and so forth.
Analytics platform uses consumer behavior patterns, market trends and other factors to help merchants effectively target the appropriate consumers.

Algorithms Developed

Innominds developed three algorithms for the reverse auction platform:

  1. A relevancy algorithm to send a consumer request to merchants best suited to fulfill the request. This is done through an unsupervised learning differential model that computes the distance between the consumer’s position along product attributes (such as price and shipping) and the merchant’s position along the same product attributes, using the scale invariant Mahalanobis distance.
  2. A ranking algorithm that orders the offers made by merchants for relevancy to the consumer. A logistic regression was fitted with offer attributes as predictors.
  3. An NLP algorithm that identifies products and feature requirements from consumer requests in free text through unsupervised learning using topic modeling.


Data from multiple sources and forms such as from web logs, transaction processing systems, social media and enterprise systems running potentially into several Terabytes of data.


Practices Involved

Case Studies

Retail Live Connect

How do you leverage big data to improve consumer experience and effectively target shoppers in real-time?