Jennifer Zaino
SemanticWeb.com Contributor

Matchmine is at the center of efforts to help users manage their own IDs in the open Web 3.0 world, says J. Trent Adams, founder and chief innovator of the startup and one of the speakers scheduled to discuss the business risks around this Internet evolution at the Jupitermedia Web 3.0 conference in Santa Clara, Calif., this week.

As an officer of the Data Portability Project, Adams is an advocate on the topic of abolishing the re-entry of data. The stated purpose of the Data Portability Project is to put existing technologies, techniques, policies and initiatives in context in order to facilitate translation, education, advocacy and ultimately implementation of data portability. Portability is defined as both physically moving data or simply porting the context in which the data is used.” At the session, panelists are expected to explain how easily transportable data helps brings companies closer to their users, and that in fact the risk companies take is in not working to this end, as it may limit their ability to reap competitive advantage.

“Matchmine is a poster child in its ability to leverage this,” Adams says. Matchmine is a media discovery network that identifies a person’s tastes and interests in specific media and then applies that to provide other media items they also probably would like to consume. What sets a discovery network apart from search or a recommendation engine, Adams says, is that it is the most passive of the three. “With discovery you go somewhere to do something and are enabled to also encounter discovered elements that may relate to what you are doing there,” he says. The service differentiates itself in two respects: one is its focus on specific media groups (film, music, blogs, podcasts, short form-factor Internet video) and the other is that it is a portable user-centric solution. “We give the consumer power over their experience and the ability to interact within our network; any partners within the network can be part of the dialogue and the consumer has control over that dialogue with partners.”

The ‘matchkey’ is the portable representation of a user’s taste and interests. When you walk into a system — say, the music site Fuzz.com — with a matchkey, you are providing your tastes and interests in a wide-ranging representation that leverages the system’s semantic technology. While you can create a matchkey at the matchmine.com site (either trained via your input of interests or just a baseline key), the portability angle comes from the fact that matchkeys can be automatically built with tastes and interests based on the accounts users may already hold with one of its partners (such as Fuzz.com or Blogged.com) or users may “feed” the key with information they may have with an off-network, non-partner service they subscribe to.

Here’s how the service works, as something of a nuanced version of a recommendation engine. You enter a new site cold with an abstract view of your interests and tastes. For example, you may not have any data in your matchkey related to music–maybe you just came from blogs.com and your tastes and interests are informed by the manner in which you have consumed blogs. But if you walk into Fuzz, Matchmine can leverage its understanding of your taste and interests according to media groups and inform your discovery experience at the new site.

“So when you bounce out of a silo and interact in a portable fashion in other systems, discovery becomes more prevalent, obvious and distinguishable,” says Adams. “So our system provides the ability to walk into a [user] data-poor environment and still provide a discovery experience before you reach critical mass. That helps both startup partners [in targeting to visitors] and it helps the user showing up at the site for the first time.” It’s always the user’s decision, though, about whether they want to use their matchkey in a portable cross-site way.

The semantics and natural language processing technology behind the service are basically this: As users move through a site, Matchmine collects telemetry events that tells it what kinds of interactions they are having on the site. It uses those events in its reasoning engine to apply to a map of understanding what those behavior patterns mean. Matchmine has 300 attributes of users’ tastes and interests across media; those interests may have overlapping attributes. So, to go back to the example of moving from blogs.com to Fuzz.com, it applies the differences of the patterns it’s identified on blogs.com and bubbles up from Fuzz’s catalogue attributes that map the blog experience to the music experience.

For example, your blog patterns may indicate you have been reading about politics, and it may combine that information with demographic data; then, as it detects further patterns, it may become obvious whether your political focus leaned to red or blue states, urban or rural, “and as you interact with these types of parameters that floats nicely into the music space,” Adams says. “Because types of music do vary in a measurable, detectable signal way. The urban to rural, north to south, these patterns do exist in generalized interests in music. So the technology narrows the universe of discourse for you.

If you walked into Fuzz without the blog experience, there are 360 degrees of options and you are like anyone else walking off the street,” Adams says. “But if you first bounce through blogs.com and then show up at Fuzz (matchkey at the ready), we essentially in a probabilistic way have narrowed the degrees of freedom from 360 oftentimes to 60 or 70 degrees.” Thus, it can better discover the information you may be most interested in.

If there’s no attribute overlap between the blogs and music sites, though, then the service won’t discover targeted and relevant data for you. “But if in your experience in blogs we do detect a pattern in media within the music space, then we will suss that pattern out and flavor your experience,” Adams says.

The company’s science is patent-pending, says CEO Mike Troiano.

“Helping partners present more relevant content enables us to deliver more targeted advertising, which raises CPM and creates new, off-site revenue streams – a share of which is paid back to participating network partners. It’s a behavioral targeting approach and business model, but with some unique data portability and privacy features for users (think TACODA + Pandora + OpenID),” Troiano says.

SemTechBiz is Less Than 2 Weeks Away

The Semantic Tech & Business Conference (SemTechBiz) is coming to San Francisco on June 3-7! Join us for case studies, innovative panels, tutorials, and keynotes that will provide you with practical advice, hands-on guidance, and breakthrough approaches to solving business problems with semantic technology. Passes go up $200 at the door. Sign up now and save !