The problem with discovering what to watch via over-the-top devices is that you’re really not much better off than you were using a cable electronic program guide. Neither is the content provider.
A new automated content discovery software system from San Francisco-based ColorTV aims to change that, gathering info on individual users’ preferences and steering them to it. Obviously, ColorTV should be pretty interesting to advertisers too.
CEO Giancarlo Maniaci has a useful analogy; digital music services that intuit what a user listens to, and suggests sound-alike tracks.In ColorTV’s case, it suggests look-alikes, broadly. It uses audio from video streams and image recognition technology to discover what you’re watching, and then suggests more of it in the same vein.
Among OTT apps, Netflix’s recommendation engine is hailed for doing something like that. Its analytics have even determined when viewers get hooked on a series, or when and if they get bored.
Broadcast apps leave you out there by yourself. But Maniaci says its ColorTV software, called Spectrum, does the same kind of analysis. Its machine-learning personalization service gives viewers suggestions based on what they’ve watched before, and on the reverse side, it gives programmers real-time data about who’s watching and who’s tuning out.
“We want to solve discovery,” he says, and so does everybody else.
What’s interesting to me is that ColorTV’s system adds the possibility of using promos that show directly after a viewer has completed watching a program, pushing them toward content that, let’s say a network, is promoting. Driving real-time tune in with promos helps create audiences (though I wonder if in practice they become exactly the kind of promos that drive a lot of us crazy).
Right now, apps from broadcast or cable networks just dump a bunch of content in your lap and more or less say, figure it out. The ColorTV deal is more helpful, not just to viewers, but in theory, to advertisers too. The big data dashboard ColorTV shows could be real-time dynamite. It lets broadcasters in on viewer analytics, right down to a single viewer.
That kind of applied data can change things quickly, like color TV sets did when they were introduced. Partly, that’s why ColorTV took that name. More prosaically, Maniaci says, ColorTV is a well-known and easy to remember appliance.
The SDK has been tested via Apple TV and is now available for Amazon Fire TV, Android TV, Roku and Chromecast apps. NBC via Pivmo is an early client; another non-network content provider is seems poised to jump in soon. But while it may seem the most logical users would be established networks, Maniaci says, “It’s not just NBC creating content. There’s BuzzFeed. There are lots of digital providers.”
Certainly, the market does seem mighty large. The slide show ColorTV shows would-be customers quotes Jupiter Research that predicts 332 million OTT customers worldwide by 2019, from 92 million now. Diffusion Group estimates ad revenue from OTT will grow to $41.5 billion in 2018 from $8.4 billion now. Being in early, if ColorTV delivers what it says, can be huge.
In its early use, with one million users able to use ColorTV’s content presentations for 20 apps, Maniaci says he can document a 51% increase in consumption, and a 10% decline in viewer “bounce.”
Maniaci was the co-founder and CEO of Tapit, a pioneering mobile ad tech firm sold for $23 million in 2013. Chris Richards, the chief tech officer and the other founder at Tapit, has the same position at ColorTV, which just started at the end of 2015. Taiwan’s Foxconn, the world’s largest contract electronics manufacturer, provided $1.5 million in seed money. Foxconn’s John Hui is a ColorTv adviser.