The New Must Have Accessory for Online Retailers? Recommendation Engines
These days, recommendation engines are so common on larger online retail sites that you might not give them a second thought. Amazon and eBay have long been using personalization technology to create a shopping experience thatâ€™s highly customized.
The whole area of collaborative filtering, the behind-the-scenes technology used to drive customized recommendations, has advanced to the point that now, it has become easier than ever to implement without the huge investment in development. It doesnâ€™t even have to degrade site performance like it did in the old days.
Susan Aldrich of the research firm Patricia Seybold Group believes recommendation engines will become commonplace next year. A survey Seybold and the Institute Telecom Paris conducted with 100 European and U.S. companies indicated that every one of the responding companies was planning to add a recommendation engine if it didnâ€™t already have one.
Aldrich thinks one main reason for the new popularity of recommendation systems is advances in collaborative filtering. She says it uses modern web analytics instead of longs lists of business rules. The new systems are easier to update and more efficient.
Companies like Webtrends and MyBuys are on the forefront of collaborative filtering. Webtrends just announced full availability of its â€œOptimize Profile Enhanced Targetingâ€ solution. It provides marketers with the ability to deliver a personalized web experience based on past behavior of visitors on their websites. Casey Carey, vice president at Webtrends, says profile-based onsite targeting â€œpresents a great opportunity to maximize relevance, increase conversion rates, and ultimately see a significant return on [usersâ€™] marketing investments.â€
MyBuys, which provides personalization for multi-channel retailers, has introduced â€œKinetic Advertising.â€ According to CEO Bob Cell, this â€œis a format that enables our display ads to incorporate movement, that let users interact with them and see a dynamic set of products that fit their profile.â€ With a Kinetic ad unit, says Cell, you can â€œpopulate a custom set of products and offers based off a shopperâ€™s interactions with a brand that are specifically tailored to that individual.â€
Aldrich draws an intriguing analogy for the recommendation engine â€“ she likens it to a global positioning system. â€œIf a site visitor doesnâ€™t take the recommended path,â€ she says, â€œit will come back and recommend another path.â€ A key strategy of the recommendation engine is to use a site visitorâ€™s known shopping history to put â€œthem on a shopping path that other shoppers with similar interests have taken before completing a purchase.â€
Not surprisingly, says Aldrich, companies making use of or planning to implement recommendation engines see value in cross-selling and upselling. Obviously , if an online retailer can get a customer to make an additional purchase at the same time as he or she is buying a product, the incremental cost for the sale is negligible. Recommendations, says Aldrich, have â€œhigh impact on conversions and order size.â€
Expect to see more recommendation engines popping up on even smaller online retailersâ€™ sites as plug-and-play collaborative filtering technology becomes more widely available.