Monday, April 19, 2010

Why Google can’t Out-Open Facebook with XAuth


This post was also published in VentureBeat.


XAuth, Google’s attempt to head off Facebook’s domination of online content sharing, is fraught with problems. It appears to be built with good intentions, allowing smaller social services to persist in a Facebook- and Twitter-dominated world. But unlike OAuth, the standard many of those services use today to link publishers’ websites to their services and which allows any website to work directly with any identity provider, XAuth actually stands in between the two and directs traffic. And that spells trouble. I should know. I’ve tried what they’re doing before.

I was the CTO of Sun Microsystem’s Liberty Alliance, where we invented the concept of federated identity, and am the inventor on the patent that covers federated identity. We first attempted to do the exact same thing –stand in between the websites and the identity providers in order to provide a seamless interaction.But no one wanted a single third party in that position — least of all privacy advocates. Instead, we delivered a specification that is very similar to how OAuth works today. Trying to evolve OAuth so that it becomes a service that intermediates direct relationships is a nonstarter.

The thesis of XAuth is that there are too many login and sharing services for a publisher to choose from, and that XAuth would only show the ones that are relevant to a user. This is the same principle as Google’s OpenSocial platform for developing social applications. OpenSocial’s creators posited that there were too many social networks for developers to choose from, so Google would provide a single open API that would access multiple social networks. We all know how that played out – there is only one social network that matters anymore.

The point of logging in and sharing is to share with friends. If people are all on one social network, why do they need to choose from a list of 20 providers to share? Most publishers offer logging in and sharing via Facebook and Twitter, which are both very easy to implement, and they get to service the vast majority of their customers.

I believe it is no coincidence that XAuth is launching right before Facebook’s F8 conference, its annual gathering for developers. The aim of XAuth is not to make it easier for users to login and share information, but to curb the inroads that Facebook is making into Google’s core advertising business. Facebook’s $500 million in estimated annual revenue is still small compared to Google’s $23 billion. But Facebook has steadily progressed over the past year, passing Google first in terms of minutes spent online, and recently in terms of unique visits.

Why should Google be so worried about Facebook? It really boils down to an obvious truth. Facebook is about people, and Google is about the algorithm. And who uses computers? People. And the algorithm doesn’t even work anymore. Search for “futon” and take a look at all of the results that have been gamed by advertisers. Who can tell you what a futon is the most comfortable? Not Google.

By spreading Facebook features into websites, Facebook will learn who is using which sites and what content is popular for which types of people. Websites will increase registrations, increase pageviews, increase sales, and most importantly be able to display much more highly targeted ads. The ads that Facebook displays next to profiles are way more targeted than the ads that Google thinks someone might like based on the content of a page.

To better understand the urgency of XAuth, I’ve included a table laying out current and upcoming Facebook features that publishers will love to integrate into their sites and that will displace some of Google’s most lucrative features. The people at Google are not stupid, and they wil
l have to pull more tricks out of their hat than just XAuth.

Current/Future Facebook Feature
Replaced Google Feature
Benefit to Publisher
Benefit to Facebook
Facebook ConnectGoogle Friend ConnectIncreased registrations and logins.Learn which sites are popular and how they are visited.
Facebook SearchGoogle Site SearchReturns what products and content is popular within a social network, increasing pageviews and sales.Learn what products and search terms are popular.
Facebook LikeGoogle BuzzPosts content to newsfeeds drawing traffic from friends.Learn in real time what content is popular and viral.
Facebook Banner AdsGoogle DoubleClickHigher CPMs since social and profile targeting are better than content or behavioral targeting.Google’s banner ad revenue.
Facebook Text AdsGoogle AdSenseHigher CPMs since social and profile targeting are better than content targeting.Google’s text ad revenue.

Can the algorithm evolve to beat the people? If it doesn’t, Google could face a Facebook-dominated world. Here’s an illustration of what one popular publisher, the Huffington Post, might look like after Facebook takes over. It’s Google’s worst nightmare.

Thursday, April 15, 2010

TV 2.0: Hulu’s Flatlining, and the Networks are Ready to Innovate


This post was also published in VentureBeat, where there is feedback from the founders of TiVo and Brightcove.


Hulu — a joint venture of NBC Universal, News Corp., and Disney — has had a good thing going. The ad-sponsored video site carries numerous TV shows through agreements with the broadcast networks. And for the past year, it’s made good money selling ad space on network shows and luring in viewers with its quality streaming. But that’s all about to change.

The Hulu value proposition as a destination for premium online video was always doubtful given that Hulu is owned by content companies that stream the same content from their own websites. Sure, there was a window of time there when it was very difficult to stream good quality video with a modern player, but that window is long gone. The networks now all offer excellent streaming high-definition 1080p players (although some properties such as AmericanIdol.com could definitely benefit from a phone call to Brightcove or Ooyala for a player and streaming upgrade). On top of that, with Comcast’s acquisition of NBC, the content itself is now owned by the folks delivering the pipe.

CBS never allowed Hulu to syndicate its content, in a very prescient decision on its way to becoming a top 5 video destination. Viacom recently pulled its popular “Daily Show” and “Colbert Report” shows off of Hulu, really letting the air out of the balloon.

Hulu sells ads on the video it streams, meaning that Hulu’s ad sales team competes with the networks’ own ad sales teams. Hulu’s sales pitch to the networks was, “let us compete with you on your new content and we will help you monetize your older assets”. But Hulu hasn’t been able to monetize the older TV shows it runs. Pull up any TV show over two years old on Hulu, and all of the ads are public service announcements. (Although Google, ever the expert on remnant ad inventory, bought up all of this inventory for a pittance over the Christmas holiday season to play precursors to its Super Bowl ad.) So the Hulu trade off is not working. Yes, it’s profitable, and yes, the number of streams served is growing, but the number of unique users has been flat for almost a year.

If Hulu is not going to be the solution for premium content owners, what is?

The TV networks always complain when they are not in control of when and how a viewer watches their content. They complained about VCRs, and then DVRs, both times because users could skip the commercials. It took a long time for DVRs to finally make the networks money. Nielsen had to start tracking who was watching time delayed commercials, and advertisers had to agree to include these views if they occurred within three days. Video streaming, on the other hand, forces users to sit through commercials. The trouble is, there was no system in place to track the number of views, making it hard for network TV’s ad sales teams to pitch the inventory to advertisers.

The good news is, in January of this year, Nielsen announced that it would start combining TV and online viewing of shows into a single rating. With Nielsen’s move, the networks have an independent auditor that can verify that viewers are watching a particular show. This presents an opportunity to integrate Internet distribution into their existing business model, which they have not been able to do even with on demand video on cable. So from here, we’re going to see some real evolution taking place in TV. It’ll no longer be something simply ported into the online world. Instead, we’ll see the networks fully move online and develop new business models for broadcasting to an internet-based audience.

Following is a prediction of five steps we’ll see these networks take in the coming months:

(1) Allow content to be streamed online at the same time as or within an hour of when it is broadcast over the air. For three days these streams are counted by Nielsen just as if a viewer watched the show live or on a DVR.

(2) If the video is played in fullscreen mode, show all the same commercials as in the broadcast version, just like a DVR, but also force the viewer to watch the commercials. If the video is played within a browser window, use shorter, online-style, 15 second commercials, and place sponsored engagement features such as quizzes and polls below the video player.

(3) After three days, replace the commercials with different commercials that are sold outside of the upfronts. This presents an opportunity to sell popular “season catchup” packages and other such products, so the online ad sales teams continue to have a unique product to sell.

(4) Allow the network’s streaming players to be embedded on any site, so that a thousand Hulus like Boxee, Windows Media Center, AppleTV and TV.com can blossom and help increase ad views for a small cut of the revenue. Since the network’s commercials are now being propagated, companies like Boxee can be harnessed as part of the syndication solution.

(5) And the thorniest issue: affiliate stations. It used to be that if you wanted to watch CSI in San Francisco, you could only see it on KPIX, CBS’ local affiliate. Now you can wait a few days and see CSI online from CBS.com. The local affiliates will need to be cut in on the Internet action, and they will also need to adapt to changing times. With cable, satellite, and now Internet streaming, there is no real compelling reason to have the same content broadcast over the air. In fact, it is a disincentive — Disney makes far more money per subscriber from cable systems on ESPN than on ABC, since they are simultaneously broadcasting the ABC content for free. Network affiliates are now differentiated by local ad sales and local programming, not last-mile content distribution. Although it will require some infrastructural changes for both the networks and the affiliates, the networks can extend their ad platform so affiliates can place ads for viewers within their geographic areas that stream shows. In addition, broadcast affiliates should be able to stream the shows from their own websites with additional, local engagement features surrounding the show.

This model I’ve outlined above will enable the networks to sell ads over the air, on cable, via DVR and online at the same time and have them all measured by Nielsen in a single rating number. Affiliates are included in the model since they can also distribute their ads to their local markets. Online ad sales teams can sell additional ads after the initial viewership period. And all the content is embeddable and monetizable across the web. The unique aspect of web content is that it is easy to create engagement — and particularly social engagement — around that content, which accelerates content uptake.