With summer here and the fall TV season still months away, how do you decide what shows you might want to stream on your HDTV or tablet to fill the void? Ask your friends? Check out your Netflix recommendations? Consult tea leaves? Wrong, wrong, and oh so wrong. At least according to Televisor, a new site/recommendation engine launching Wednesday that aims to take what you like and spit out other stuff you'll like too. I got to play around with a preview version of the site--which has the tagline of "Discover Great TV"--for a few days, and it definitely shows some promise.
Visit Televisor and you're greeted with a big search box with the words Find me a show to watch like... with a Recommend button beneath it. The button is somewhat pointless, because as you start typing, a drop-down menu appears with suggested completions. (The only time you'll then see the Recommend button is if you enter something not in Televisor database.) You can scroll up and down within the list, and when you select a show from the menu you're taken to a page of recommendations based on that show.
A search for U.K. workplace comedy The IT Crowd, for example, turned up an expectedly large batch of British shows like That Mitchell and Webb Look, Coupling, and Rowan Atkinson's wonderful Black Adder, but also short-lived (and underrated) American workplace comedy Better Off Ted, Scrubs, Community, and Eureka. All solid choices. And on the recommendations page for The Big Bang Theory, I'd seen every episode of six of the first eight shows (even if the results were tainted a bit by Two and a Half Men appearing in the group).
At the top of the results page are several buttons, one for each genre of the show you started with (you can select more than one to focus results); Recent and Classic buttons that let you narrow by the age of shows; and Hulu, Amazon, and two Netflix buttons (one with a maple leaf, which I assume is for Canadian Netflix subscribers) that let you see only those shows available on the services to which you subscribe.
Each show on the page displays when it ran, its genres, a brief description, and either an IMDb rating or Metacritic score (which link to the show's page on either site). And you can give each suggestion a thumbs up (to see more shows like this), a thumbs down (you don't like this), or click an X to indicate not to display that show ever again. As you click one of those icons, the site runs the search again to provide more-focused results based on your feedback. Which, oddly, also removes a particular show you liked from the list of results.
When you see a show that interests you, click its name or poster art to go to its unique page. There you'll find reviews, links to the streaming sites that offer the show--currently Hulu, Amazon, and Netflix, with others promised--trailers, several other recommended shows, and a button to launch the larger recommendation engine based on that show.
So how does it work? The company told me that Televisor uses several layers of data: Collecting more than a million TV show ratings from sites like Amazon, IMDb, and Metacritic on about 3500 shows at launch; analyzing how TV show preferences correlate across that large group of people; integrating genre data; and finally personalizing based on individual input (the thumbs-up and -down buttons mentioned earlier).
To be fair, this isn't a new concept. Every Amazon page has a Customers Who Bought This Item Also Bought section. Netflix makes suggestions "because you watched" something. But what makes Televisor interesting is that it's a free service, dedicated solely to streaming TV, and makes suggestions based on the shows you feed it instead of your viewing/purchase history from a single service.