I couldn’t come out and say it in Part I of this series, but now I can report that Mobile Content Networks (MCN), a U.S. mobile search platform provider, has acquired Caboodle Networks Inc., a company best known for its multimedia, multi-service semantic recommendation platform. Translated: Caboodle has the know-how and the patents to deliver recommendations based on the user’s context and the intelligence to tackle age-old semantic search problems such as Jaguar (the car) and jaguar (the cat). It doesn’t have a silver-bullet solution, per se, but it does go a long way toward disambiguating similar search terms.

More importantly, Caboodle’s open source approach to recommendations transcends content types and silos to deliver recommendations from across all available content types. (In this aspect, the recommendation capability reminds me of Xiam, a server-based recommendation solution that may not have solved the semantic Web problem, but has developed a capabilities mix to offer a user who types in “Beyoncé”, for example, a wide choice of news, gossip, tracks and ringtones – the works. More about that when the service formally launches soon.)

Marc Bookman, MCN CEO, who graciously agreed to be grilled for over an hour in a pre-briefing, puts it this way: With Caboodle under its belt MCN can create the associations between content and successfully cross-sell and up-sell the users to similar content. Caboodle doesn’t index content; it relies on taxonomies and contextual information to categorize it and play matchmaker between users and the content likely to matter most. It’s an interesting hybrid approach that builds on collaborative filtering, but goes a step further. “It’s going to be very much a content-driven approach to this [recommendations] in terms of trying to get the correct associations between the content and the user’s queries.” (Further down in the story, I deep-dive and provide a few excerpts from an internal white paper, so read on. Thanks again to MCN marketing execs Stephen Burke and Michael Crane for passing on excellent background material!)

Listen to Part II of the podcast. [27:36]

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

Just the facts: First, let’s be clear on Caboodle. Vendor pitch aside, here’s what the company tried and tested. A study was conducted to compare the existing text-based, keyword search with Caboodle Networks’ Semantic Web-based search solution. A music catalog of over 5,000 entries from a wireless operator was chosen to provide an adequate number of ringtones to search. 4,654 queries were performed, first using the text-based keyword search, then Caboodle Networks’ Semantic Web-based search. The results: “Caboodle had a 93 percent match vs. a 70 percent on a text-based keyword search. An even higher percentage match is achievable by augmenting the search with spellchecking.” No details on the methodology, but the stats speak in favor of an approach that can understand the data and the original query and not just match keywords in a relational database. (I’ll watch the rollout and circle back with an informed opinion once I’ve seen it in action.)

Follow me: Marc doesn’t see tough competition on the horizon yet, but that could always change. In his view, there is some tug-of-war at some points on the value chain, but, overall, it’s a non-issue for MCN. “We got in there early with this, it’s complete, we’re expanding it, we’re adding more advanced features, and I think it’s going to be difficult to catch up with MCN now for this particular application.” He adds: “It’s a very competitive environment, but, in terms of a company that’s built a search management platform that sits in the middle of the value chain and doesn’t compete with content and with distribution, I do believe we’re the only ones that position ourselves this way and actually are executing on a business plan to do this.”

Where it counts: Asia is a vibrant market for mobile search, with IBM pegging usage in Japan at 20 percent of users. Why does usage in the U.S. and Europe lag behind? “We need 3G. Although there are people who say 2.5 is good enough, good enough isn’t going to create a mass market.” On top of that, the markets in these regions need flat data plans to bring down the prices and pique interest in surfing the mobile Web. “That’s when advertising kicks in. The carrier business model [also] has to support content to be [sold] both off-deck and on-deck.”

The good news: Mobile search in markets that satisfy these requirements has exploded. “Some of the markets [where] we are have about a 10-15 percent data adoption rate. We can see traffic grow 100x because we get 10 times more users and 10 times more page views for 3G versus 2.5 G.” The bad news: It takes about two years to get to this inflection point. “Europe is just starting to travel along that curve and so Europe is very interesting to watch right now.” In line with this observation, MCN has beefed up its presence and efforts in Europe, installing a team in Europe to “talk to many players in Europe.” In the U.S. MCN is “quietly” working on versions [no additional details for the moment] with partners. “I think the U.S. market is going to need time to evolve.”

This entry was posted on Wednesday, September 5th, 2007 at 6:10 pm and is filed under Mobile Advertising, Mobile Music, Mobile Search, Multimodal Search, Podcasts. 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.