Trapit Aims To Be Your Search “Personal Assistant”
Somewhere between the “precision of search and the serendipity of social media” lies Trapit. Trapit is a “personal discovery engine” based on the same underlying technology that forms the core of Apple’s Siri virtual assistant. From the outside, however, it looks something like a cross between StumbleUpon and Flipboard or Zite. Co-founders Gary Griffiths and Hank […]
Somewhere between the “precision of search and the serendipity of social media” lies Trapit. Trapit is a “personal discovery engine” based on the same underlying technology that forms the core of Apple’s Siri virtual assistant. From the outside, however, it looks something like a cross between StumbleUpon and Flipboard or Zite.
Co-founders Gary Griffiths and Hank Nothhaft are quick to rebut each of those comparisons by pointing to Trapit’s underlying technology as the differentiator.
Users input keywords or entire URLs and create “traps,” which are content modules drawing upon approximately 100,000 “high quality” content sites. Thus the corpus being tapped is a tiny sliver of the web; however Trapit’s co-founders believe that the sites they’ve identified offer a much better experience than a large index of open web content.
For example, Trapit won’t use content from “aggregator” sites (e.g., Yahoo News) or content farms. The company uses a combination of machines and humans to identify high-quality sources. Trapit then uses “semantic extraction” to identify pieces relevant to your query or URL. Over time Trapit becomes more personalized, based on explicit user ratings (thumbs up or down, like Pandora) and other signals.
In earlier interviews, Trapit’s co-founders were explicitly invoking Pandora to describe the nature of the service. With me on the phone earlier today they used a number of analogies and metaphors that indicate they’re struggling a bit to convey what’s useful and what’s different about Trapit. They said they were seeking to walk a kind of tightrope between search and social discovery and get that balance right.
Trapit has been around for about a year in private beta with 10,000 users. Data provided by Trapit showed impressive engagement metrics among these early private beta users:
- 100,000 traps (content modules) created
- 31% active users
- 24 minute average stay (vs. Facebook’s 22:53 or New York Times’ 7:13)
- 17 page views per visit (vs. 21 for Facebook)
Trapit will release iOS and Android apps, including for tablets in Q1 2012. That’s when the Flipboad and Zite comparisons will really kick in. Again the co-founders assert that their technology makes Trapit more personalized and sophisticated than more basic “social news readers” or aggregators on tablets.
Yet without the benefit of having spoken to Trapit’s co-founders as I did, you wouldn’t necessarily put it in a separate category from the crop of tablet-oriented news readers to come out in the past year: Flipboard, Pulse, Taptu, Zite, AOL Editions and Yahoo Livestand.
Another way of describing what Trapit is doing is “persistent search.” But you wouldn’t use Trapit instead of a search engine. You’d use it as a news reader. Trapit may also expand into other areas beyond news, such as shopping or deals for example. There’s nothing about the underlying capabilities that are wedded to news content.
Set up of personalized “traps” — Trapit will also recommend traps — is not as intuitive as it might be. On balance, however, the user experience was promising.
One reason Siri is engaging is because it puts a voice interface on top of its sophisticated technology. That creates greater usability than in the absence of voice. And while Trapit may have the same intelligence under the hood, the exterior isn’t quite as sexy.