RichRelevance.com - Dynamic Product Recommendations

Ever since online stores like Amazon hit the big time, programmers have been under a lot of pressure to come up with systems for recommending products to customers, in a way as personalized and accurate as possible.

This startup aims to make a stand by presenting a system that is termed “Adaptive AI”, and which has been adopted by retailers such as Sears and KMart. The main strength of this new recommendation engine is that it tailors every customer individually. It is not a matter of “those who shop at ABC are different from those who shop at DEF”. Rather, it seems to be a matter of “Those who shop at ABC are different one from each other; let’s create an engine for suiting each individual as minutely as possible”. This is accomplished by bringing personalization to the fore, but also by employing collective intelligence to the full.

In practice, this means that item-based recommendations are employed intermingled with a personalized approach. Furthermore, the company dispenses with isolated algorithms like the ones used for collaborative learning, and focuses its approach on an ensemble learning model that runs several algorithms at the very same time.

If the abovementioned sounds good to you, nothing will prevent you from trying this system out at any of the supported retail locations like Sears, The Vitamin Shoppe and Burton. Just visit any of these stores and see what is put your way by the system. Of course, it will work better once you have a couple of transactions behind you, but it is also designed to work out from square one.

“We deliver richrelevance through a highly consultative partnership with our clients based on our team’s experience in e-commerce, merchandising, personalization and network architecture for industry leaders including Amazon, Overstock, Akamai, PayPal, eBay, Microsoft, Hotmail and Qualcomm. “Ensemble Learning” constantly evaluates multiple and interrelated shopping behaviors, creating a picture of customer desire in real time. The technology gets “smarter” as customers advance through a site by collecting feedback via a loop as customers interact—or don’t interact—with recommendations. Based on this active feedback, the technology optimizes in real time choosing the most effective recommendation types (out of 20+ options) at each stage of each shopper’s experience.”

It is a viable alternative to other recommendations systems geared at customers.

What is the minimum number of transactions that must be carried out by a customer for RichRelevance to yield the best results