Xconomy is reaching a milestone of sorts: some of the startups we profiled last summer when we were just getting started have now had time to evolve through at least one major generation of their technology, giving us the opportunity to come back and see where things stand. That’s definitely the case with Cambridge, MA-based EveryZing, the BBN spinoff we featured last July when the firm was just launching its speech-to-text system for indexing audio and video content and making it discoverable by major search engines.
Last summer, as EveryZing CEO Tom Wilde told me this week, the company was offering “a capability rather than a solution.” The company’s automatic transcription technology produced text files that could be published alongside audio or video files, opening up those files’ contents to traditional text-based Web searches and therefore, in theory, making it easier for the owners of multimedia-rich websites to attract traffic. But it wasn’t a full-service system that companies could simply buy and implement. Today, though, EveryZing is launching the full commercial versions of two products, called ezSEARCH and ezSEO, that establish the company’s technology as a serious platform for publishers who want to boost consumption of their online multimedia content (and, more to the point, of the ads published alongside that content).
Here’s how it works: a client, typically a media company with a large Web property, signs up with EveryZing, whose Web-based service then runs alongside the client’s existing content management system, extracting full-text transcripts from audio and video files and creating an index of these transcripts for search purposes. The ezSEARCH product places a single search box on the publisher’s website that allows visitors to browse all of the site’s multimedia content (or, optionally, all of the site’s content, period, including text). When the search results come up, users can either view clips from the beginning by clicking on them, or, using an innovative time-stamped bar under each file, jump to the moment in the audio or video when their keywords appeared.
If an EveryZing customer also signs up for ezSEO—the letters stand for Search Engine Optimization—a lot more happens behind the scenes. EzSEO is, essentially, an automatic publishing engine. It uses the indexed, time-stamped multimedia files as the core content for hundreds or thousands of new “landing pages” organized around keywords drawn from the text transcripts. These new pages aren’t necessarily reachable via links from inside a publisher’s website; while they have plenty of interesting, human-viewable content, they’re designed primarily to attract the attention of—and get highly ranked by—the general search engines.
Wilde argues that getting a site’s multimedia content noticed by Google or Yahoo is a much more sure-fire way of drawing traffic than just waiting for users to stumble across the content while browsing, or publishing it via special video portals such as YouTube. “There are all these vertical multimedia search engines like Google Video, Yahoo Audio Search, and Yahoo Image Search, but honestly, users don’t want to search that way; they really want to use the single, Google-style search box,” Wilde says. “And when the industry created Web search technology, it was centered around HTML and text, and it still is, and it will be that way for a while. What we do is to optimize multimedia content for the big Web search engines—because that’s where 80 or 90 percent of the search activity is happening.”
For an example of the EveryZing system in action, check out Boston.com, one of EveryZing’s largest customers during the ezSEARCH and ezSEO beta-testing phase. Actually, start at Google, by typing in “Tom Brady audio video.” The first result will likely be
