The World Wide Web is becoming the World Wide Watch -- a creepy place where people talk about behavioral targeting and learning your intent. The goal is less nefarious, because implied in that is the ability to offer consumers more meaningful content. Startup Atigeo promises to make this all palatable by giving users access to their social profiles, including how that profile is acted upon.
The World Wide Web is becoming the World Wide Watch -- a creepy place where people talk about behavioral targeting and learning your intent. The goal is less nefarious, because implied in that is the ability to offer consumers more meaningful content. Startup Atigeo promises to make this all palatable by giving users access to their social profiles, including how that profile is acted upon.You can watch a ReviewCam of this technology below.
I'll start with a simple explanation: You create a profile, and that profile can be accessed with your permission on any site that has implemented the Atigeo platform. That profile learns as you visit sites, and as the platform learns on its own; you can also teach it what you like manually.
The Atigeo platform is called XPatterns. Part of this platform is a semantic learning engine (my description, not necessarily theirs) that uses things like Wikipedia to build relationships between topics. Atigeo points its own domain experts at certain representations of knowledge in particular topic areas, and the engine crawls, watches and learns (there's a video representation of this in the video above). These domain experts can be pointed at web sites, databases, blogs, and updates on sites like Facebook and Twitter. It is constantly learning and re-shaping its knowledge.
Some of those entities (some sites, and certainly Facebook and Twitter) have their own profiles -- ones the user creates, but that the sites manage. Atigeo professes to give the profile (the persona) back to the customer, who defines and shapes it. The data belongs, in other words, to you. You manage it. You teach it. You decide how it gets shared, and, thus, how you are targeted. You can go into your profile and see what the system has learned, and you can alter it or add to it or categorize things.
Sites or applications that take advantage of XPatterns pop up what's called a LifePass bubble on the site, and you can choose to share your profile with it. For example, you might want to share your profile with the New York Times and have it deliver content that is relevant to your tastes and habits (and intent, and even location). The persona you build even knows when you're in work mode or social mode. And, yes, those third party sites and applications can also target you -- but hopefully with offers that are relevant; if they aren't, you can tweak your profile further. These third parties can't see what's in your persona, they just get to take action on it through the Atigeo platform.
To demonstrate all of this, Atigeo created TweetPatterns, an application that can pull out the relevant tweets of the people you're following. In the demonstration in the video above, you can see that it isn't just matching text, it's extracting a level of meaning.
Atigeo isn't out just yet -- it's in trials with what it calls major brands. The company also says there will be a mobile agent for the product. Although the idea seems compelling, Atigeo has its work cut out for it on both ends of the transaction -- getting consumers to build profiles and getting merchants and web sites to implement the technology.
Fritz Nelson is an Executive Editor at InformationWeek and the Executive Producer of TechWebTV. Fritz writes about startups and established companies alike, but likes to exploit multiple forms of media into his writing.
Follow Fritz Nelson and InformationWeek on Twitter, Facebook, YouTube and LinkedIn:
Server Market SplitsvilleJust because the server market's in the doldrums doesn't mean innovation has ceased. Far from it -- server technology is enjoying the biggest renaissance since the dawn of x86 systems. But the primary driver is now service providers, not enterprises.