First, imagine that a majority of households have a digital video recorder of one kind or another (such an event is predicted to occur by the year 2009, according to Forrestor). Further, imagine that this DVR has a “search history” of everything you’ve watched and are planning to watch (this is already done by most DVRs). Further still, imagine that this history is – with your tacit approval – blended with an edited profile of your online searching habits, forging a marketing precise of your likes and dislikes, your wants and needs (doing this is a matter of a marketing deal between DVR providers and search engines). Perhaps you use Google Desktop Search, or A9, or Ask, or Yahoo – it matters little, all of them create a search history already.
In his book, The Search, he goes on to paint a picture in which (a) television’s business model is reinvented as niche, long-tail advertisers who could never afford national television ad buys come out of the woodwork—just as they did online—when they’re able to pay more modest sums to target subgroups of likely customers; and (b) TV no longer broadcasts, but becomes a massive series of database objects—your favorite shows are all automatically transcribed and searchable, giving you the ability to bring up not only episodes, but every line of dialogue or appearance by a favorite actor or character, every mention of a particular topic, and so on and so forth.
Well, that future is arriving quickly—or at least parts of it are. Connected televisions and set top boxes are proliferating, of course, and so is video search. In 2006, just two years after Battelle made some of these original predictions, scholars at UC Santa Cruz created MetaVid, a project which ingests C-SPAN footage and mines the closed-captioning data to produce a searchable video archive of every utterance made on the floor of Congress. Search a phrase like “healthcare,” and you’ll bring up a page of results showing every time the word has been mentioned, paired with transcripts and the footage of the Senator or Representative doing the talking. You can also categorize results by member of Congress. And not only that, once you’ve found the transcript and video you’re looking for, you can create a custom embed of just the part of the clip you want, and add that to your Website or blog. Way to keep your Representative accountable, no?
Last year, MSNBC.com followed suit with its new video player (introduced above by Lester Holt), which not only pairs every video with the best available transcript, making its video archive text-searchable, but it also allows the same sort of embeddable subclip creation that MetaVid does. Even cooler, the technology is good enough that you can click or highlight a portion of the transcript and have the player (a) jump to that part of the video, and (b) produce an embed code for that portion of the video clip.
And then, of course, there’s Hulu, which introduced its caption search feature in 2009, allowing users to do many of the same things with all of its own content.
Moreover, last year after Congress mandated that all professionally produced online video contain closed captioning, they not only improved its accessibility, but ensured that pretty much all premium video content on the Web will have the kind of metadata that makes these technologies possible.
So the stage is set for the fusion of search and video—and we’re beginning to see it arrive in some prominent places around the Web. What’s not ready yet? That business model Battelle talked about in 2004. Certainly the potential for it is there. I’ve written about this before and recently made a similar comment on NewTeeVee explaining my thoughts:
Why not charge advertisers more to provide [shows] free on Hulu and other online distributors? Friends and I have hypothesized that people retain ads from Hulu and its ilk better than either TV or TiVo. TiVo’ers skip ads altogether. Live television ad breaks are long enough that people tend to talk through them, run to the bathroom, or make themselves snacks. But on Hulu and the network video players the 30 second ad breaks are (a) impossible to skip, and (b) too short to run away from. Thus, the ad recall online is likely to be superior to that from other types of media, to say nothing of the medium’s ability to target customers. All these are reasons the networks should be able to market ad space online for great rates.
But despite all these hypotheticals, we’re finding out that online video viewing is far outstripping ad spending, and that online audiences are currently valued at about a tenth of television ones. We’ll see how fast that changes. Then again, perhaps I shouldn’t be so eager or quick to speak—Joseph Turow over at Annenberg is in the process of writing what promises to be a fairly compelling book about some of the perils of this sort of surveillance marketing and advertising. For those of you with library access, you can see some of the seeds of his argument in this 2005 paper.
Author’s Note: I do dissertation research on this topic. You can see my “Editorial Standards” page for any questions about my relationship as a researcher to MSNBC (the short version is that none of the facts or analysis here rely on any inside knowledge of MSNBC or its Web operations, nor of any of the other Websites mentioned).