Amazon and Netflix were pioneers in building recommendation engines, and they used the systems they developed to become leaders in their respective categories.



Until recently only large retailers had the money, expertise and time to develop the complex algorithms that drive personalized product recommendation systems. But the market has advanced and a number of vendors have proven they can deliver real customer value. Now the technology is affordable to even small retailers.

Yet, according to Vovici, only slightly more than one-fourth (27.6%) of online retailers offer any type of personalized product recommendations.



But that proportion could be changing.

"A growing number of success stories is spurring online retailers to jump on board," says Jeffrey Grau, eMarketer senior analyst. "Zappos.com, for example, currently uses a low-tech solution to offer personalized recommendations on only some product pages, but its recent expansion into product categories beyond shoes may be just the impetus needed to push the retailer over the hump."

Brian Kalma, director of user experience at Zappos.com, said the retailer’s goal is to roll out recommendations on a massive scale with a powerful engine at the center: "Hopefully one day our search engine will become so effective that it could replace what we are currently calling a recommendation engine."

"Even though automated recommendation systems are rapidly improving, there are situations where peer recommendations provide more meaningful results," says Mr. Grau. "Take for example a woman who has just moved to Manhattan and needs help choosing and purchasing appropriate business attire. An automated recommendation system would be challenged to find suitable suggestions based on her historical shopping data. A better source of apparel advice would be other professionals working in New York."

Peer recommendations are most commonly shared on shopping-oriented social networks such as Kaboodle, ThisNext and StyleHop. While these social shopping sites are very much a work in progress, their popularity is growing.

"Retailers like personalized product recommendations because they create cross-sell and upsell opportunities," says Mr. Grau. "They also cite long-term benefits such as a higher degree of customer engagement and loyalty."

Despite low adoption rates, 40% of online merchants surveyed by Internet Retailer said they planned to add personalized product recommendations to their Websites by the end of 2008 - second in priority to video functionality.



"Despite being around for 10 years, the personalized product recommendation market is still in its nascency," says Mr. Grau.