ChoiceStream has released its Personalization Survey, an excellent read that never fails to shed important light on industry trends. In view of my interest in the figures, ChoiceStream has provided MSG with the results and an exclusive pre-briefing with Toffer Winslow, ChoiceStream EVP for all-things personalized, that covers all the bases including an update on ChoiceStream’s recent tie-up with white-label mobile search and advertising company JumpTap. More on this a little further down in the post.

Vendor spin aside, I have been a fan of this survey since I dissected the results for my first mobile search and content discovery report back in 2003. However, this time the results are much more science-fact than fiction. Users are not only interested in personalization; they see a real value in it. In fact, 37 percent of the 811 U.S. users surveyed reported that they would have bought more music, bought more movies (42 percent) and bought more ringtones (57 percent) if they could have found more content that matched their preferences. The content industry is clearly leaving money on the table…

The survey also asks users’ attitudes towards more personalized and targeted advertising. The results are encouraging: Overall, 41 percent of all respondents said they would be “more willing to pay attention to advertising if it was personalized - based on personal tastes and interests.” When broken down by advertising medium, the results show users interested in personalized advertising on their mobile devices is on the rise, up almost 60 percent from last year to 38 percent.

But users can also be brutal if they receive a recommendation that fails to make a perfect match between them and the content they’re likely to appreciate. Overall, 39 percent of users who receive a poor quality recommendation are less willing to return to the site again. Granted, the survey looks primarily at online consumer habits, but the results speak volumes when I think of the operators and content providers lining up to offer recommendations as a means to cross-sell and upsell their try content and differentiate their mobile search services. The takeaway: Quality matters!

I asked Toffer for his take on the results. In his view, personalization is “table stakes for companies serious about selling content.”

Q: In view of the increasing importance of recommendations and the low tolerance for poor matches, how should providers, content providers and operators tweak how they offer recommendations?

A: They need to move beyond the basic approaches that have categorized recommendations as simple pattern matching: If you bought item A, then let’s take a look at all of the other people who bought item A and use that to identify the next most popular item B and make that recommendation of item B to you. But that’s not really personalization; that’s not taking into account the fact that you’re a unique individual and you have unique tastes and you’re trying to accomplish unique things. That’s just basically treating you as part of an overall pattern.

I think that approach has resulted in poor recommendations and almost half of the folks we talked to say that they’re getting just plain bad recommendations; suggestions that didn’t match their preferences, the item that was recommended was inappropriate or it was something they already owned. Content providers and operators with some form of personalization or recommendations on their sites need to be increasingly critical of themselves and about how they’re providing recommendations. In a word, they have to be more sophisticated and accept that users are also more sophisticated.

Operators have to think about the user experience they’re delivering and to keep users on-deck, if that is their aim. To accomplish this they have to have rich content and be able to extend a comparable Internet browsing experience. This means helping the users parse through content effectively by understanding their interests, their near-term intent at any given session, and serving up the right results. Not all content is created equal and helping users get to content they most want and are willing to pay for has to be central to an operator’s mobile strategy.

Q: What is ChoiceStream doing to improve the quality of recommendations? What does it bring to the table?

A: We offer a platform for personalization that can enable a variety of approaches depending on the available data. Just like your site, we see ourselves at the intersection of content and context. To make a great recommendation that’s on-target and unique to the individual, you’ve got to understand a great deal about the individual, about the content, and about the context into which you’re making that recommendation. But you’re not always going to have that deep knowledge, so you will need a recommendation platform that can accommodate different amounts of input and still make quality recommendations.

The survey shows user expectations around recommendations and personalization are changing; the majority of approaches today are not satisfactory. And essentially, we believe we’ve built a better ‘mouse-trap’ and we want companies to understand that they’ve got to try and up their own game when delivering recommendations to their users.

Q: The other side of the coin is search. You have a partnership with JumpTap. Please fill me in on your go-to-market strategy.

A: We’ve been working with JumpTap because we’re hearing in the marketplace that personalization, recommendations and search go hand-in-hand. JumpTap has some level of personalization capabilities; we bring them – through the interoperability work that we’re doing – a deeper level of personalization and recommendation.

There are a couple of use cases. One, we can help deliver personalized search results. Based on what we know about an individual user and their preferences and their overall profile, we can help sort search results in a different way. If you and I were to search for the same term, we’d get different results because the underlying ChoiceStream engine knows more about you as the user and can help re-sort the search results in such a way that the most relevant ones for you appear on top.

Second, if you’re doing searches for content items that we know a lot about – such as entertainment content, ringtones, wallpaper, music, video- we can make recommendations about similar items you’d be interested in as a user based on the terms that you’ve already searched for.

Third, there’s a huge opportunity further down the road around personalizing the ads that are delivered to users.

The go-to-market strategy is two-tier. There’s a certain class of early adopter customers who are willing to sort of build their own and pick best-of-breed components and do some of the integration work themselves. As the market matures, customers will be looking for these solutions pre-integrated and delivered from a single source. We expect that partnerships with search vendors will see us doing the work to either embed ourselves inside their solutions or do much of the interoperability work so that – out-of-the-box – you’re able to use the search infrastructure you get from one party, and the personalization and recommendation capabilities from us. The partnership approach with other vendors represents a go-to-market strategy for ChoiceStream, designed to access the more mainstream market.

Again, thanks to Toffer for the pre-briefing. He’s promised to circle back with some exclusive stats on mobile video personalization, so check back regularly.

This entry was posted on Tuesday, December 4th, 2007 at 11:01 pm and is filed under Content Discovery, Mobile Advertising, Mobile Content, Mobile Search, Personalization, Recommendation. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.