Last Friday I had a phone conversation with Rich Pearson and Matt Robinson of Attributor, a Redwood City, CA, startup. The nominal reason for the call was that Attributor announced this week that it has signed up Reuters as its second big customer using the company's content-tracking platform. In actuality, we had a wide-ranging conversation on the company's products and philosophy.
People who've been reading me for a while will know that I feel we desperately need new business models - the old dinosaurs aren't going to survive. Attributor wants to make the case that they provide a technology platform on which new models can be built, based on the ability to find out quickly where content is appearing on the Web.
Attributor's own business model is that a content creator signs up with them and identifies content to be tracked - so far they have announced deals with news syndication giants AP and Reuters and are in "closed beta" with 15-20 other content creators, including bloggers. The Attributor platform monitors Web pages and identifies instances of use of the tracked content.
The "secret sauce" (as they put it) has two parts: one is they claim to be able to track billions of Web pages in near-real time; the other is that their idetification algorithms are claimed to be able to identify the amount of content reuse, down to about the paragraph level within a text document.
The first is interesting in that it's a claim about the value of the long tail. Does AP care that its story appears on someone's blog or MySpace page? Does it care as much as it cares when that story appears on your local NBC affiliate's Web site? Maybe, maybe not. One of the arguments Pearson and Robinson made is that their customers care about the "reach" of certain messages. It's one thing to see your stuff show up at all on Digg or Slashdot; it's another thing to know that within a week of that appearance the story was found on 300 other blogs. From just a simple marketing perspective, it would be useful to know which types of content propagate in which ways. Attributor claims that people care about the long tail into time as well - not just the the first 72 hours.
The second point is truly key to discussion around fair use. If we believe that anyone in the Cartel still respects the principles of fair use doctrine, then it makes a huge difference whether I've copied a whole AP story or am just quoting a paragraph in a Copyfight posting. And of course once a "hit" has been found it's simple to analyze the page for the presence of paying ads, giving prima facie evidence of whether someone using a piece of content is making money off that use.
So if the technology works - they claim 99th percentile accuracy in both precision and recall - what can you do with it? Well, one of the services Attributor offers is automated DMCA takedown notices. It's up to the customer to decide to send them, but they're there. This led me immediately to wonder how they plan to avoid fiascos like SFWA created with its spurious takedown notices. To wit, what prevents me from using Attributor to generate improper DMCA notices?
The answer is "not a whole lot." Attributor have some identity-checking capability, as do most online Web companies. And they provide linkbacks so that takedown notice targets can see what content specifically is claimed to be infringing as well as who generated the notice. But the DMCA remains a blunt, clumsy instrument and Attributor isn't going to fix that. Pearson and Robinson want to focus more on the notion of "multiple remedies". If I know who's using my content in what quantity how often for what purpose I have much more information to go into a business negotiation. They believe their customers are more forward-thinking and interested in new revenue opportunities more than locking down use.
Speaking of lockdowns, what about DRM? Attributor paints itself as "the opposite of DRM" and on this point I agree. The fundamental notion of DRM is control over action by people with content - it's necessarily a preemptive strategy since you can't attach DRM to content post facto. The fundamental notion of Attributor is "know what's going on then decide what to do". By providing visibility into how, where, and when content gets used, they believe they can open up opportunities for new product development, such as use-based licensing.
Will this really happen? Magic 8-ball says "ask again later". It's quite clear we need new tools and new ways of thinking about managing digital content, but I don't see any guaranteed winners this early in the game.