Personalize Video's Pull

Start talking about the concept of personalized video with industry stakeholders and you'll quickly find yourself debating interactive applications such as a mosaic view that lets users watch six shows simultaneously or the ability to use Twitter or Facebook while watching TV.

In interactive TV jargon, these all are “lean forward” applications that require the user to make selections. But traditional television has been anything but “lean forward,” and some people within the video development community believe some consumers will want to retain the laid-back approach of traditional TV while also making it more personalized. The concept is to give people a personalized video channel that knows what each person would like to watch and automatically serves it up with little action required.

The idea already has taken root on the Internet, where sites such as ffwd and MeeVideo stream music videos or other video content to the user based on preferences or usage patterns. Some developers are working on bringing this capability to the TV screen, where content could be drawn not only from the Internet, but also from the service provider's content library.

“The promise of TV is that you flop on the couch and it entertains you,” said Simon Steward, CEO of Vizimo, a company that develops personalization and recommendation software to mobile, set-top box and middleware manufacturers. “That's why on-demand video has kind of failed.”

When a service provider has thousands of titles in its on-demand library, finding which one to watch can be challenging. That's why service providers are working on refining their search and recommendation engines. At press time in May, Verizon was set to launch a software widget for the FiOS STB that would simplify the process of searching for video content in its own library and on the Internet. The company also has opened its widget platform to about 300 of what Joe Ambeault, Verizon's director of product development for consumer TV, called “Ring Two FiOS television partners.” And based on research from The Diffusion Group about what consumers are most interested in, those developers are likely to pursue recommendation apps such as one that suggests TV shows based on a user's interests.

Personalized channels would take the concept of the recommendation engine a step further by automatically presenting content to consumers. “You could do a virtual channel that blends linear and non-linear programming,” Steward said. “It would play from a schedule and fill in gaps with on-demand content because you might not have back-to-back linear content you like.”

Instead of asking, “What's on now?” consumers increasingly would think in terms of “What's on next?”

Steward predicted, however, that people will want to jump in and out of their personalized channels, and that developers of personalized channels will need to give people the capability to override the pre-selected content.

Vizimo has been working with service providers to develop personalized channel capability. “They're starting to ask for this,” Steward said.

At least two developers of advanced video apps claim they already can deliver personalized video channels.

“Our playlists are personalized channels,” said Buno Pati, CEO of Sezmi, developer of a wholesale offering that uses a combination of broadcast and DSL or 4G wireless to enable service providers to deliver video services without investing in headend equipment. Sezmi's service uses a combination of previous routines and a recommendation engine to determine content that would be most appropriate for any individual family member.

Individual family members identify themselves by pushing their own button on the remote control. There also is a family button. “When you go in, shows that you have watched or set to record are presented, and there is no distinction between whether it's recorded, live, Internet or something you have purchased on demand,” Pati said.

The goal, Pati said, is to “minimize the time between when you sit down until you are watching something you care about.”

By selecting a playlist function, users can view a collection of shows strung together to form a channel that makes sense for them by pulling material from cable, broadcast, the Internet and recorded material.

Typically when people sit down in front of the TV, they are not looking for a new show they have never seen before, Pati said. “The most powerful thing in television is the routine,” he said. Accordingly, Sezmi's personalized channels draw first upon material the user traditionally likes to watch, filling in the gaps with similar content.

Sezmi already has service providers signed up to use its video offering, which is expected to launch this year.

TellyTopia, a company that offers what CEO Kshitij Kumar calls “the best of Internet video” through a broadcast or on-demand method, is currently doing a beta test with an undisclosed customer that includes personalized video channels.

“We provide the intelligence: what the person has been watching and what they're sharing with others,” Kumar said. “We use an off-the-shelf partner to stream that content.”

One of its first customers is SureWest Communications, which is offering a subset of TellyTopia's full suite of capabilities. For example, SureWest is letting customers upload videos they have created to TellyTopia, which makes them available on a dedicated local television channel.

The TellyTopia software is capable of taking that a step further by letting users share their own videos only with other users whom they have designated. That content also can become part of a personalized video channel. One of TellyTopia's hooks is that it screens content prior to making it available through its system and can screen out or tag mature content.

Kumar is a bit less enthusiastic about the concept of personalized video channels than Steward and Pati. “We'll make it available,” Kumar said. “If they watch it, great. If not, we'll let them create a playlist on line.”

The success of personalized video channels will depend, in large part, on how relevant end users find the content. “The crux of the problem is that no single technology will give you the answer all of the time,” Steward said.

Vizimo uses a wide range of factors behind its recommendation engine, taking into account everything from the top shows people have set to record to newspaper reviews to what friends are watching

Another critical variable is metadata — the content descriptions that accompany both traditional television programming and Internet video. Steward said Vizimo looks “deep inside” metadata to determine what he called the “aboutness” of the program.

Personalized video channels could work on a household basis, but to obtain maximum value, developers also need to generate channels on an individual basis — and that requires that users identify themselves when they sit down in front of the screen.

Some see that as a substantial roadblock. “Most people have no interest in logging in to say who they are,” said Eric Freund, director of business innovation and development for SureWest.

The success of personalized video channels also will depend on how extensively service providers embrace the concept — and to date, major players have shown little interest. Although a high level of intelligence and searching would underlie a personalized channel offering, the passive nature of the experience may simply be at odds with service providers' current emphasis on two-way interaction.

When asked about the idea of a personalized channel, Ambeault said, “Widgets are focused on information feeds — graphic and text. There are plenty of other places to get music and video.” He mentioned Verizon's new search capability as a good way for people to find video content.

A key competitor for Verizon is the Yahoo!-Intel team, which is working on a widget platform with a wide range of partners, including Comcast. Russ Schaefer, senior director of product marketing for Connected TV for Yahoo!, expects recommendation engines to be a key area for developers, but he envisions them being more of a “lean forward” experience.

Using a recommendation engine as a starting point, Schaefer said, “you could keep saying, ‘Go to the next video.’ It could be automated.”

Like Verizon, Intel and Yahoo! have opened their widget platform to third-party developers — and with so much in the works, perhaps someone will think of expanding a recommendation engine to support personalized channels. There is no evidence yet, though, that anyone is taking it to that level.

Perhaps it is simply too early for personalized video channels. Until people have more experience with recommendation engines, it may be too much of a stretch to even imagine what a personalized channel would be.

As president of The Diffusion Group, Michael Greeson has done extensive research on advanced video concepts. He believes personalized channels would be “highly desirable,” but that they “might be one of those things consumers don't know they want.”

The success of personalized channels, he said, “depends on the power of the recommendation engine and having really good metadata.“