Note: This interview was conducted on February 26, 2009, prior to the annoucement of Amazon's acquisition of Zappos.com.

eMarketer: What are Zappos.com’s plans for personalized product recommendations?

Brian Kalma: We do some personalization on the site now, and we’re going to be redesigning to create the recommendations in a way that are more tailored toward someone’s previous experience or interaction on the site.

eMarketer: Where are the product recommendations on the site now?

Mr. Kalma: They’re on some product pages, not all.

Right now, the personalization is driven more by visual merchandising needs as opposed to more collaborative browsing. [Visual merchandising is suggesting products based on what complements the other items the consumer is purchasing. Collaborative browsing is suggesting products that are likely to be of interest based on the consumer’s past shopping behavior.]

eMarketer: What is your vision of personalized recommendation at Zappos.com?

Mr. Kalma: First off, personalized recommendations are what they’re called in general, but I think a better way to look at it as is more relevant search results. So if you consider the concept of getting search results based on the aggregate worth of users’ social behaviors, hopefully one day our search engine will become so effective that it replaces what we are currently calling a recommendation engine.

And I don’t mean as a little module on the side saying, “Here are the best sellers,” but integrate that into your actual search itself.

eMarketer: What about personalized recommendations in e-mails—promotional e-mails?

Mr. Kalma: We’re definitely considering it, but we’re very sensitive to being intrusive, so it’s not somewhere we want to go into just yet.

eMarketer: What do your personalized recommendations do best?

Mr. Kalma: They deliver the right product at the right time during the browsing experience and shopping experience. Specifically, people are more pleased with their purchases because we deliver to them—easily—what they are looking for.

eMarketer: All right, and on the flip side, where are they weak?

Mr. Kalma: The weakness is primarily in how they are executed.

Quite often sites rely too heavily on creating recommendations as a module as opposed to integrating the feature into the shopping experience. That’s not necessarily the recommendation engine itself falling short—I just haven’t seen recommendation engines deployed in a way that’s terribly impressive yet.

eMarketer: What about the issue of relevancy? Do you feel comfortable about personalized product recommendation engines delivering relevant recommendations?

Mr. Kalma: We’ve been pleased with what I’ve seen out there. It gets tricky when you have folks who are shopping for other people. If you’re coming in and shopping for your wife, and that’s a one-time occurrence, but if you come back next time and you’re getting a lot of women’s products geared toward you and your recommendations. That’s potentially a challenge.

eMarketer: Will people who receive a personalized recommendation on your site have an opportunity to give you feedback—let you know whether that recommendation is on target or off target to help your system learn or adapt to that individual?

Mr. Kalma: Well, we want to get there. But being the type of company that we are—really loving to talk to our customers and encouraging them to pick up the phone or enter live chat—I’m almost positive we’ll be getting feedback that way, which is more qualitative, of course, but very useful.

The full version of this interview is available here, to eMarketer Total Access subscribers only. Every day they have access to new interviews with digital marketing leaders and trendsetting entrepreneurs.

Click here for information on how to subscribe to Total Access, or call 1-800-405-0844.  

Get more articles like this one delivered every day. Click here for the eMarketer Daily newsletter.