our series on recommendation engines, we look at ChoiceStream - a recommendations vendor which counts Overstock, Borders and AT&T among its high profile clients. ChoiceStream has recently turned its attention to using recommendations in online advertising, and in this post we look at how the company is doing this. The ChoiceStream advertising product aims to generate personalized banner ads for each consumer, using data on shopping and buying patterns that it collects from the advertiser's website. The company claims that this technology improves click-thru rates, conversion rates and average order size.In this latest installment in
There would obviously be an affinity with Overstock already for the user, if they had visited Overstock (or similar sites also using ChoiceStream's platform) before and then saw an ad for Overstock on AT&T. Nevertheless, despite this bias it makes sense that conversions would increase if the Overstock ad that a user sees on AT&T's website is personalized based on that user's past shopping or browsing behavior. The company claimed that for this particular campaign, user click-thru rates doubled, conversion "about doubled" and Overstock orders increased by an average of 20-30%. If accurate, then those are compelling figures for Overstock.
It remains to be seen whether this technology can be used generally. Right now it appears to have a relatively limited niche in the retail market and its success is very dependent on users having visited and used participating sites (such as Overstock) before.