- Looks at Meta, Contextual & Social Network Data
Anyone who has used the online version of Netflix or LoveFilm knows how hard it can be to find what they want to watch once the favorites have been viewed. With tens of thousands of selections available, most of which users don’t want to watch, finding the right one for the moment is difficult. And the choice depends on their location, home, office, train, and the time of day. We think that recommendation engines like the one The Filter has developed will be an essential part of future online video services.
In the UK, TV-viewing time increased on average by two hours per week as a result of companies providing recommendations for what to watch from social media friends, according to The Filter CEO David Maher-Roberts, citing numbers from the UK’s Broadcasters Audience Research Board (BARB). Speaking at SXSW he said it’s no surprise that some TV and film producers are looking at including social media streams with their broadcasts.
Since 2009, The Filter has provided its recommendation engine to Web services such as NBC, DailyMotion, Warner Bros, Nokia and Sony Music. Roberts says the recommendations, which are based on taste, location, time of day and prior consumption, are more accurate than Amazon’s recommendations are with books and DVDs. Roberts says the company currently serves over 100 million recommendations per month.
“We can look at content in context of each individual, based on mood and taste, for example,” he said. “Connected devices produce mountains of relevant data in real time, and this can be used to understand specific demands of individual users.”
The recommendation engine, he said, looks at three different data types:
- Metadata offers content descriptions, performers, titles, images and trailers that can be pulled through via search.
- Contextual data provides the location, times of day, day of week, device and subscription package for every view. As an example he cited children watching TV after school with more child-relevant content recommended between 4 PM and 6 PM.
- Social data based on shared interests and reinforced through social networks is more likely to lead to program recommendations within specific subject areas. It’s like looking to see what a person’s Facebook friends are watching. It has an increasing influence on program discovery.
He said smart TVs with a combination of these data types are already possible. It is also possible, he said, to measure the ongoing effectiveness of each type within certain viewing groups. He said real time analysis is essential to the future of TV and film services and to the future of advertisers that want the most efficient route to consumers.
The Filter has recently received a third round of venture capital, which is believed to be about £800,000 ($1.3 million).
Peter Gabriel, once the lead singer of the rock band Genesis, is a major backer. Gabriel also started Spotify competitor We7, a music streaming service that allows users to download music to their handsets for offline listening.
There are also online recommendation services like moki.TV that are not integrated in an online site. Consumers can use moki.TV to select a show and be shown a list where it can be viewed now as part of a subscription service or purchased. Moki.tv is not a widget on any smart TVs, smart TV adapters or Blu-ray players that we know of. That makes it a two-step process to watch a selected show on a TV set: first find it on the moki.tv Web site and then find it on online service on the TV set if the set or adapter supports that particular service. For example, LG TVs and Blu-ray players do not support Amazon.com and Apple TV only provides access to Netflix.
To see 4 free editions of The Online Reporter, the weekly source for competitive intelligence about digital content, online entertainment services, mobile media and wireless networks, visit www.onlinereporter.com/trial_copies.php
