trending Market Intelligence /marketintelligence/en/news-insights/trending/iSAdABKmVgqXqQXicGhYGQ2 content esgSubNav
In This List

Becoming the Google of video


S&P Capital IQ Pro | Powered by Expert Insights

451 Research Podcast

Next in Tech | Episode 41: IoT's Role in Energy and Utilities


Top 10 Tips: Europe's shifting TV powerplays


Insight Weekly: Soaring food prices; bankruptcies reach new low; insurtech M&A to accelerate

Becoming the Google of video

It has been some years since media has been shaken by aGoogle-sized opportunity, and many investors are wondering if there are anyleft.

Xavier Kochhar thinks he might just have one.

"Google didn't just shift advertising; Google actuallyopened up a whole new pie," he said in an interview. "We're doing thesame for video."

The digital video industry is still a shifting andfragmented landscape. There are no Google-esque discovery solutions, nocomprehensive iTunes-like shop and no interactive tools for compiling, sharingand finding a user's favorite videos. But Kochhar and his small Santa Monicateam believe they are developing just that: one video meta-platform to bindthem all.

His solution, called The Video Genome Project, or The VGP,seeks to catalog every video available online, attaching reams of data pointswhere historically there have only been a handful. This includes bothcommercial video, like movies and TV shows, as well as user-generated content,like the hours and hours of streams uploaded to YouTube every second. Theproblem seems unfathomable, unquantifiable in man hours.

That is, if The VGP was using man hours.

This is what Kochhar says makes The VGP different from goingattempts to organize video with structured data: His approach is totallyprogrammatic. There is no one physically tagging "Poltergeist" withdata like "scary" and "possession." Or rather, there is,but they do not work for The VGP.

Source: The VGP
The VGP Founder and CEO Xavier Kochhar

"The world is telling us exactly what they think aboutvideo, but you can't see it, because how they're saying it is notstandardized," the executive said.

The company creates the data using a semantic engine, anautomated, patented four-step process for scraping the Internet of all thesemantic data surrounding a specific video. If people are talking about a videoonline, whether it be short-form user-generated content or a full-lengthHollywood film, The VGP can categorize it based on that semantic information.

In doing so, they realized that there are a finite number ofcategories, or "genotypes," that when combined in differentquantities can essentially describe all of the video content online.

"We're just listening to the world. We're cleaning uptheir data, and the world was saying these genotypes exist at scale in everystory, kind of like the basic building blocks," Kochhar said.

Today the company has identified over 100,000 “genes” thatdescribe stories. This volume of metadata attached to video, as well as theautomated process used to collect it, are unprecedented, the executives claim,and they set them apart from the competition.

"We've created over the last three and a half years ataxonomy that is so rich and so interconnected that no one will ever catchus," The VGP COO Kerry Tucker said.

The result is a system that can sort video in extremelynuanced ways. Common genes in content — granular themes like"outcast" stories, "one-versus-many" plots and "stranded"protagonists — begin to take shape as part of a common preference, like themesof "isolated heroes." Suddenly, it becomes clear why someone likesboth "Star Wars" and "Napoleon Dynamite" but not "StarTrek" or "The Hangover." It can show how the preferences of a21-year-old male gamer are similar to those of a 42-year-old housewife, and howthey can be cross-marketed. It can even show the intensity of a certain gene ina video, like just how "car-chase-y" a certain film is.

Source: The VGP
The VGP maps content based on 100,000 genes

Kochhar and Tucker plan to market The VGP as both a B2B anda B2C service. Distributors can plug the product into their own over-the-topplatform as a search and recommendation engine, and there will also be adesktop and mobile app where users can search the entire web's worth of contentfor platform-agnostic suggestions and discovery. The system can even tell auser if a video is available for rent on one platform, like , but will also beavailable soon on a subscription video-on-demand platform that the user alreadypays for, like NetflixInc.

DIRECTV vice president of product management Sarah Lyons,who has encountered many recommendation engines, said she believes there ispotential with a The VGP-like approach.

"A lot of third-party recommendation engines arelargely doing recommendations based on the same thing, whether it'scollaborative filtering or recommending things based on similar genres oractors," she said in an interview. "It's going to take someone to dosomething unique and break through and approach it from a different way."

Just how the company executes with its product will be avery important determining factor for The VGP, Jordan Levin, chief contentofficer for the National Football League and a member of The VGP's advisoryboard, said.

"The challenge for them is trying to figure out how tostay focused because there are so many different ways to exploit thatdata," Levin said in an interview.

Because of his background in media, The VGP is just thefirst offering Kochhar decided to create with his semantic engine, but thetechnology could also be applied to any corner of the market where semanticdata exists: retail, dating, terrorism, travel and so on. But for now, just astrategy for video is daunting enough.

For example, when Kochhar told Hany Nada, managing partnerfor Silicon Valley venture capital firm GGV Capital, that he would targetdistributors like ComcastCorp. and AT&TInc., Nada said he encouraged him to broaden his prospects.

"It's interesting, but working with [MVPDs] is reallyhard. What I think would be killer is going after the consumer market becauseof all of the problems we have finding the right content," Nada said in aninterview.

Kochhar took the advice and moved forward on the B2Cplatform, integrating marketing solutions and other partnership opportunitieswith media brands that may feel threatened by a third-party recommendationplatform. As Nada pointed out, current content providers want people to returnconsistently to their own platform, so The VGP will have to offer a substantialcarrot to gain their cooperation.

Source: The VGP
The VGP's B2C platform under development

The exact size of the opportunity is unclear, Nada said, buthe is bullish. The worldwide video recommendation market in 2015 was at around$68.0 million, according to SNL Kagan analyst Michelle Abraham's . She forecasts thatrevenue to grow to $206.9 million by 2020.

Those are relatively small numbers, but the competition isalso somewhat manageable. She identified only 11 companies playing in thefield, including big vertical corporations like as well as small startups likeJinni and Spideo.

"I think there is quite a lot further to go. There aremany companies who have not adopted a recommendation engine," Abraham saidin an interview. "The operating companies tend to be in the maturemarkets, so there are still greenfield opportunities."

However, she was skeptical at the Google-like revolutionKochhar is so passionate about. The market has been moving incrementally, evenamong The VGP's direct competitors that claim to use similar techniques, so amassive paradigm shift is doubtful, she argued.

Still, some disagree.

"Really up to now Google has owned browser search. It'sintriguing to me that no one really owns video search yet," L.E.K.Consulting head of media and entertainment Dan Schechter said in an interview.

He pointed to Priceline Group Inc., Expedia Inc. and TripAdvisor Inc., noting that the combined market cap ofthe biggest online travel agencies rivals the value of the entire hotelindustry they parse. Search and discovery can be as big an opportunity as theindustry people want to search.

Furthermore, Netflix itself said its recommendation engine createsabout $1 billion each year, which is about 15x the video recommendation enginemarket in total, based on SNL Kagan data. Big MVPDs and content providers goingover the top have been too focused on getting services established to invest inthis math, Schechter said.

"They haven't responded because they haven't been ableto get out of their own way," he said. "But if you were Comcast youshould have a Netflix recommendation engine that is much better now than thecurrent thing. You should have had that two years ago."

Because of the Comcast-sized possibilities, and consideringthe flight to OTT services and new digital bundles, Schechter sees morepotential for the recommendation market than Abraham.

For Lyons at AT&T, the opportunity is also clear, butfor more obvious reasons than the market dynamics.

"Other industries do it so well," she said. Musicplatforms, travel services, retail websites, even trunk clubs, have become verysophisticated in their personalization process. Why not video? "There'sgot to be something to advance video beyond where we're at today."