With so much information, content and so many services now living online, there's a lot of choice -- even for something as simple as where to go to buy a new pair of socks. Oh, and there's a lot of...
With so much information, content and so many services now living online, there’s a lot of choice — even for something as simple as where to go to buy a new pair of socks. Oh, and there’s a lot of data. As it’s evolved and gotten better at making sense of its new Big Data, the Web has become an extraordinary engine for discovering new stuff: News, cat videos, porn, you name it. Naturally, scores of sites are becoming (or are building) recommendation engines to help users wade through the noise, and, dining on Big Data, they get smarter every day.
However, as it stands today, the discovery process is pretty fragmented, as recommendation engines tend to be domain-specific. Want to find a good movie? Try Netflix. Want to find a good book? Go to GoodReads, etc. And this fragmentation makes for a crappy user experience.
So, frustrated with the fact that there’s no one-stop shop for great recommendations on, well, everything, a couple of seniors at MIT have developed, and quietly launched, Tipflare to be that general solution. While the site’s creators, Hayden Metsky and Thiago Vieira, have bigger ambitions, today Tipflare focuses in on giving users one place to find quality recommendations on books, movies, songs, and restaurants — in a jiffy.
If you’ve ever struggled to find a recommendation for a good book that’s actually based on what you like to read and not just your browsing or purchase history (Amazon, for example) and then repeat that for a good movie, Tipflare’s value is obvious. If not, you might just say, hey, why not go to GoodReads or Amazon or Netflix? And it’s true that, using Netflix as an example, those that specialize in one domain will probably be better at recommending (say, movies) than a generalist.
But the Tipflare creators aren’t worried about that. If you want to watch something on Netflix, you’ll discover on Netflix. Instead, Tipflare really aims to differentiate itself from other services by being broad in focus while trying to keep the design and UX as clutter-free and as easy to use as possible.
Users simply enter what movies or books they like — or import their “likes” from Facebook, and the site instantly serves recommendations on what they’ll go bananas for based on what they already enjoy, factoring in Facebook friends’ interests, your location (for restaurants, specifically), the day of the week, time of day, etc. so that it can offer “New for You” recommendations when you come back.
Like Triposo, the travel and destination recommendation service we recently covered, Tipflare leans heavily on its algorithms to serve quality recommendations. But it also incorporates social cues as well, though Metsky says that Tipflare’s social integration is really a tool to back up and support its algorithm. It’s the supporting cast, not the star.
While Tipflare has a lot to recommend it (see what I did there?), it’s not a panacea. In other words, it’s not yet that next-gen, hyper-intelligent personal assistant; it still requires work on the part of the user. One day in the not-so-distant future, the Web, search, etc. will transition from a demand-based system (where you input what you want and a service tells you where to go), to one that’s more actively directive in telling you what to do and where to go based on your history, behavior, browsing, and more.
The site is early in its gestation and still has a long way to go, but the founders are asking the right questions. When it comes to algorithmic recommendations and discovery, there’s a trust issue. Google and so many others are addressing this with friendsourced and social graph-authenticated recommendations and results. The core reason for this being, hey, you’re probably going to trust recommendations from real people, especially those who know you and your tastes, more than a machine — even if those algorithms are wizardry.