Personalizing eCommerce: The Evolution of Filters

The Numbers: Online Shopping is The Future
Online retail sales grew 12.6 percent in 2010, hitting $176.2 billion. By 2015, Forrester predicts the market share for the online retail industry will be worth $279 billion in the US alone -- more than $900 per American annually. This growth isn't expected to slow anytime soon; Forrester expects that online shopping will eventually overtake in-store shopping, a phenomenon it dubbed the "cannibalization of store sales."

The Challenge: Personalization
Despite the shift, web-based retailers have not been successful in personalizing the online shopping experience in the way that traditional retail has. Companies like Amazon and eBay have added traditional filtering features to help consumers narrow their options, but conventional filters can sometimes be limiting.

Online retailers have access to both implicit and explicit data, which they can leverage in their recommendation algorithms. Implicit data is information gained from a user's everyday actions; explicit data, on the other hand, is information garnered from surveys that users can opt to take -- data that will greatly improve personalized recommendations. But very few online retailers have developed a simplified way to quickly survey a customer and, using that data, provide active and targeted recommendations.

Implicit Data: Passive Personalized Recommendations
Netflix and Hunch have taken a passive predictive approach -- creating personalized recommendations that are being heralded as 'revolutionizing e-commerce.' This type of personalization requires a lot of time and data accumulation in order to build accurate, passive recommendations.

Explicit Data: Active Personalized Recommendations
Sites like eHarmony are active in asking consumers exactly what they are looking for and returning results accordingly. FindTheBest is taking the next logical step by introducing a tool that makes active, personalized recommendations not only based on what features are important to a particular consumer, but also how important those features are -- similar to an online personal assistant.

FindTheBest founder and CEO Kevin O'Connor summed up the newly launched feature called AssistMe. "Many have attempted recommendation engines with big promises and dismal results," O'Connor said. "We take a far more structured approach -- tell us which product or service you are interested in and through a couple of simple, but important questions, we can show you which products best fit your requirements."

More on AssistMe
Engineering Manager Ivan Bercovich and Senior Engineer Abhishek Rajendra, who both worked on the new recommendation feature, explained how AssistMe has become the first online personal assistant that gives active, personalized recommendations across a broad range of topics.

"Most comparison sites today offer, at best, a passive approach at guessing a customer's preferences. Missing from these online sites, however, is the active dialogue and personal interaction that in-store sales representatives offer," Rajendra said, adding, "We created AssistMe to bridge that gap as one of the first online, personal assistants."

Explaining the new feature, Bercovich said, "AssistMe is a new way of finding what is best for you. Rather than eliminating potential matching products that don't fit exactly within the specified filters, AssistMe will find the products or services that fit closest with the user's needs. We do this by first asking the user which aspects of a product or service are important to them, and second, asking how important those aspects or features are -- on a scale of unimportant to very important. Then, the magic happens. FindTheBest's AssistMe algorithm calculates the options and returns a set of results that most closely match the user's personal needs. I think you'll find that people are willing to put in a little more effort in order to get a better recommendation--especially when they're making big, considered purchases."