Search across social networks can be maddening. Combing through the archives of Facebook or Twitter for a post that you know is real, but seems to be a figment of your imagination, is a feeling that I’ve heard described too many times. For this very reason, there’s little doubt in Silicon Valley that social search will be the next big thing…eventually. But there’s a reason it hasn’t been done well yet. It’s hard as hell.
Today, YourTrove is launching its platform to unlock the secrets hiding within status updates, pictures, videos, and shared links from across users’ private networks. The company offers not just text-based search, but true contextual search, making relevant connections across a given social network.
“The problem is that social content is effectively like an iceberg,” says co-founder Seth Blank. “You miss 99 percent of the content each day while you’re away. Facebook even chooses what to show you while you are there.”
Facebook CEO Mark Zuckerberg told the audience at last week’s TechCrunch Disrupt that his company was working on its own search solution. Twitter has been dabbling in the same area recently and will no doubt crack the nut sooner rather than later. Blank predicts that both will be “awesome.” What they will lack, however, and where YourTrove can provide users with massive value, is cross-platform functionality.
At launch, the company is only able to search Facebook, but it aims to extend to other networks in the near future, including most likely Twitter, FourSquare, Tumblr, and Instagram, among others. The company has its roadmap to be sure, but until the platform partnerships are signed in
blood ink, they are little more than good intentions.
The use cases for YourTrove are myriad. Imagine that the next time you travel to a new city, you’re able to search across Facebook, FourSquare, Yelp, and maybe even OpenTable for local restaurants that your friends have tried, easily aggregating their recommendations. Want to find that one picture of a cat riding a dog dressed up like a horse? Good luck, unless you can remember which network it was posted on and when. And even then, it’s a hit or miss proposition, weighed by getting the exact right search terms.
“It shouldn’t matter where your content lives, or who owns it,” says Blank. “What should matter is that it’s there when you want it, regardless.”
The thing is, it does matter who owns this content (at least today), and the gatekeepers have gone to great lengths to protect their competitive advantages. Contextual awareness across this massive sea of interrelated data has the potential to be unbelievably powerful. How willing the platforms will be to provide this level of access, however, is yet to be seen. Previously, Google was best-positioned to deliver such a solution, but with the launch of its Google+ social network, it’s highly unlikely that any competing social network will willingly share its data. The conclusion, then, is that any solution will have to come from an independent third party.
Therein lies the difficulty for YourTrove and others. Accessing all the necessary data to make this work is no sure thing. According to the CEO, preliminary conversations with each platform appear promising, but nothing is done until it’s done. Even with access granted, rate limits on the number of searches the company can perform could very well cripple the service.
Since cross-functionality is the primary value proposition of YourTrove, securing these deals is critical. Given the precarious negotiating position that this puts the small startup in, the company would do well to position itself as an ally to the large social networks, rather than a competitor.
YourTrove is coming out of a two week private beta during which it said it searched more than 100 million unique pieces of content. The company expects to pass 1 billion at some point today once the service is opened to the world. The numbers sound outrageous, but really they just speak to the scope of the problem being solved.
“It shocked us how huge the data we were playing with was,” says Blank. “We expected search to be difficult to do, not painful to do. It was actually quite painful. We expected to launch this back in April, but it took us four extra months to do it right.”
Blank and his team built to search technology on top of an open source Elasticsearch backend. The real magic comes in how to get the raw data into the platform. The usual big data toolset of data normalization and contextual mapping play a big role. According to the founder, the company has solved a significant technical problem and, while it expects competition, Blank feels well positioned to defend whatever headstart it can build.
When YourTrove first released its beta product, searches were organized according to relevance by default – you know, the way every other search engine on the planet does it. Overwhelming user feedback led the company to change this to chronological sorting. The next phase following today’s launch will be offering more flexible search filters such as by friend or by location.
Along the way, the company has been incredibly mindful of privacy. When poking around in the world of users’ personal and private data, few companies get the chance to ask for forgiveness before skittish users bolt for the doors. YourTrove silos each user’s search index so that no one can see his results. All searches are encrypted, and each user can delete his personal index at any time.
YourTrove has been largely bootstrapped by its founders thus far, raising only a nominal angel round to date. The team of three San Francisco-based co-founders has a lot of startup experience, including the successful turnaround and sale of a hosting company and consulting roles with Justin.TV, WordPress/Automatic, and gdgt. But none of this makes the task at hand anything less than a frighteningly difficult challenge.
“The search space was littered with the ashes of dead search startups,” Blank remembers about the early days of the project. “It’s a scary sector.”
The co-founder points to the fact that companies need to solve equally challenging problems of user acquisition, platform business development, and technical development as reasons why no one has succeeded previously. This said, data demands to be liberated. The greater the obstacles on a path, the greater the rewards waiting at its end. Social search is no different, and the winner’s spoils promise to be glorious.