Google rules,but Atlas coming from behind; where does that leave DigiTrust and publishers?
This is a sidebar to a longer piece found at the Reynolds Journalism Institute website.
BY BILL DENSMORE
The Identity Battle is underway – is the news industry in the game?
To understand what’s at stake, let’s go back to the mid 1990s.
Once upon a time there was an epic battle between Microsoft Corp., in Seattle, the reining king, and Netscape Communications Corp., the pugnacious upstart in Silicon Valley. It was called “the browser wars.” In the 1990s, Microsoft was undisputed king of the desktop computer and it wanted to extend that position off the desktop and across the World Wide Web.
To access World Wide Web pages at that time, you needed a “web browser.” Microsoft made Internet Explorer, and bundled it for free with its operating system. Netscape made “Netscape Navigator,” tried to sell it, then realized it had to be given away for free, too. For awhile, it looked like Netscape was winning, but the two battled to more or less a draw, Netscape was acquired by AmericaOnline Inc. and eventually faded.
In the meantime, Google began making its own free browser, Google Chrome, and a not-for-profit, the Mozilla Foundation, took over the job of maintaining and update the Netscape browser after being given those rights by AOL. It renamed it Firefox. In the Apple world – which became increasingly independent under Steve Jobs, the Apple Safari browser shipped with all Apple products. Finally, in 2012, Microsoft bought 800 patents from AOL for over a billion dollars, including the Netscape browser and cookie patents.
The lines of battle in the browser wars continues to be how each entity manages the tension between a browser that has enough common functionality to render pages and multimedia gracefully to users, and yet have enough proprietary “hooks” to be able to steer the user toward other services of its owner/sponsor.
Today the four major browser makers have more or less battled to a draw. They have their supporters and users; all four do basically the same thing. Google Chrome may work best with Google’s services and Internet Explorer with Microsoft services. But they are no longer products that differentiate and draw customers for other services.
The connection between the browser wars and the looming Identity Battle begins with a patent and a little bit of code invented in 1994 by Lou Montulli, then an engineer at Netscape Communications. It described a “state object” – we know know it as a browser “cookie.” It described how a website server could send a tiny bit of information to a users web browser (IE, Firefox, Chrome, Safari) and have that browser store it. On the next visit to the same website, the browers software would be directed to send the code back.
This elegant approach solved a problem for the World Wide Web – how to tell that a particular computer has visited a website in the past, and give it some sort of a customized experience (such as recognize a user name). “Cookies” rapidly became a mainstay of the online advertising industry, used to track who was looking at ads and where. Wikipedia has a comprehensive entry about cookies which notes that “as of 2014, some websites were setting copies readable for over 100 third-party domains. On average, a single website was setting 10 cookies, with maximum number of cookies reaching over 800. U.S. Patent No. 5774670 for cookies (“persistent client state”) was sold by Netscape to AOL and, in 2012, by AOL to Microsoft.
What’s the connection between cookies and a new identity battle? It involves Google, Atlas and DigiTrust – and the face that, for identity and marketing purposes, cookies are getting stale.
- Privacy regulators and non-profit watchdogs are wise to their actual and potential abuse.
- Newer versions of the Apple Safari browser block third-party cookies vital to tracking unique users across today’s advertising ecosystem. Mozilla has talked about blocking third-party cookies, also.
- Microsoft ships a version of Internet Explorer with “do-not-track” enabled, what amounts to an instruction by the browser’s users to advertisers not to apply or use third-party cookies.
- Because almost everyone visits a Google page, Google has first-party cookies on almost everyone’s device. This makes Google’s AdWords network and ADSense technologies fairly independent of third-party cookies. And anyone using an Android phone has a Google Ad ID.
- More than half of news consumption is now on mobile devices, and mobile devices do not all handle cookies. As many as half do not in Europe. Some carriers are also said to block cookies, preferring their own network tracking mechanisms. Apple, like Google and Android, has its own proprietary identity system for Apple iIOS devices called “identifier for advertisers, or IDFA.
“There are very many companies -- from Google to Facebook to many others -- who are using information about people in ways they will never reveal,” says Tom Rosenstiel, of the American Press Institute. “They say they don't share it -- but they share it dramatically within their own systems -- packing it into algorithms and essentially making money on it.”
All of these cookie challenges have lead industry analysts to predict both the decline of cookies as useful and common identity-tracking mechanisms, replacement with other sorts of identifiers, and, beyond that, the death of third-party ad servers. This is leading elements of the advertising industry to search for a new common identity platform. Some thing the best identity system if all is registration – and connecting a registration with known information about the user. Facebook isn’t waiting, unveiling a service which Don Mathis, writing at Advertising Age, says “solves two big problems for marketers.”
Facebook and Atlas: A shot across Google bow?
As Facebook grows and seeks to overtake Google’s online advertising dominance, this was it’s challenge: How to duplicate Google’s network of ads placed on millions of other websites? Facebook until now has mostly only placed ads on its own service and applications. But it has “Like” buttons on millions of websites, and each Facebook Like button provides the ability to track --- to set a cook on that user’s computer and associate the same user across millions of sites – “crack rock for publishers in need of page views,” wrote MIT Technology Review author Chirstopher Mims in June 2011.
So in 2013, Facebook acquired for $100 million from Microsoft a decade-old company called Atlas Solutions LLC. In October, it unveiled to the advertising industry an ambitious plan to likely try and replace cookies with a central database of user identity information culled from interactions with Facebook’s website, apps and “Like” buttons – allow advertisers, in theory, to use a common identifier for a target viewer/reader across all placement locations. Atlas will use Facebook ID instead of cookies to track and report user actions to advertisers across all devices.
Cookies are flawed when used alone, Atlas asserts: “Today’s buyers move seamlessly between devices . . . but the existing advertising technology for ad serving and measurement – cookies – can’t keep up,” the Facebook subsidiary says in its web marketing. “Cookies don’t work on mobile, are becoming less accurate in demographic targeting . . . .”
Picked as a partner by Atlas -- and advertising-agency giant Omnicom -- to test Atlas with two premier brands – Intel and Pepsi – was a little know suburban Washington, D.C., company, Neustar. Neustar has a lot of experience running real-time databases of billions of tiny transactions. It maintains a 680 million-entry U.S. number-portability database for the phone industry. But it may be about to lose that contract. Running a central identity database for Facebook/Atlas could be an attractive addition to Neustar’s existing ad-tech business. “Identity matters to marketers,” Neustar declares on its home page.
Testing with Omnicom and Neustar, Facebook has started an arms race to see who can develop fastest a proprietary across-all-plaforms digital ID that is cookie-independent so that it will work well in mobile and in apps. It already sells unique user data it has to Nielsen Corp., and with Atlas it could have a not more to sell. One ad-tech company is now working to integrate real-time bidding on display advertising with so-called “native ads” – content messages similar to stories, and testing with Omnicom.
DigitTrust – nonprofit ad-tech consortium counters Google, Facebook?
Not to be outdone, a non-profit consortium of 20 independent advertising technology companies has formed DigiTrust LLC and says it hopes to create post-cookie identity technology “that will improve consumer privacy, reduce pixels on publishers’ pages and allow third parties to provide rich, personalized experiences across the web.”
As DigiTrust sees it, there are so many third-party advertising tracking cookies on publisher websites that are slowing down website response times, angering but users and publishers. The idea is to come up with a single universal identifer for a web users – or at least their device – that all advertisers, ad platforms and networks could read and work with.
In the current advertising ecosystem, your phone, tablet or computer may have hundreds of “cookies” on it from different advertisers and networks, each identifying you as a “unique user” to that network. Today, these networks trade cookie attributes furiously back and forth in the background, trying to “match” cookies and figure how many of those hundreds of cookies refer to the same person. It’s worth reproducing a paragraph from DigiTrust’s website to understand their argument: (“Pixel synchronization” refers to placing an invisible piece of code on a web page, so that an advertiser can then use a tracking cookie.)
“Real-time bidding (RTB) technology has been instrumental to the automation of digital ad buying and selling. Pixel synchronization is a broadly-deployed process that enables RTB suppliers to include each RTB buyer’s cookie-based identifier in the bid request so that buyers can determine whether they wish to bid, and how much. However, this process results in excessive third party requests on Internet pages per day. Considering the growing number of Internet-connected devices used by each consumer, and the growing number of RTB partners integrated by publishers and marketers (each of whom uses a different identifier), the volume of pixel synchroization events is growing exponentially.”
Google doesn’t have that problem within its own vast network, and neither will Facebook, if its Atlas play works out. If DigiTrust launches with something (the consortium is still talking with constituencies and regulators, and considering technologies), the idea is that a single cookie (or something that works like a cookie) could be used by all ad networks to uniquely identity a web or mobile user, at least for commercial purposes.
What should news organizations do?
If third-party cookies diminish in importance, to be replaced by three or more competing, overarching market-driven identity systems, (Google, Facebook, DigitTrust) what do news organizations do? Contribute their user profiles to one of them? To all of them? Work with the government-inspired NSTIC initative? (See Appendix K: “Identity Matters” for detail about NSTIC.) Or create something of its own, focused not on advertising, or government logins, but on user privacy and interests?
One longtime news industry researcher says don’t start by going to the big boys.
“You dare not get involved with Google, Amazon or the others because they are so hated,” says Robert Picard, research director at the Reuters Institute. “ It is one thing to say yes we'll work with you once we have it up and going but to bring them in directly -- one of those large ones -- would I think be a problem it would be better to bring in a second-layer player and use them and then the others can try to figure out how to get involved. You still have a lot of hatred among newspaper people about Google. Most newspaper people don't have that same feeling about Facebook because they don't know better yet.”
An industry investor agrees with Picard: “There should be an alternative to Facebook in terms of the keeper of all of the information,” says Mike Wheeler, a media-tech venture investor and RJI interview for this report. “I just don't know who it is going to be.”
But for advertisers, Facebook is already meeting the need, observes Tom Drouillard, CEO, president and managing director of the Alliance of Audited Media. “Because Facebook has all this information.” He continues: “If you look at how the online targetters build profiles, a lot of that is look alike modeling at the end of the day and look alike modeling (EXPLAINED) is less accurate than Facebook if you ask me. And when you think about Facebook's reach, there is nothing better.”
Options for news organizations listed
So for news organizations that need to understand the interests and atttributes of their users in order ot deliver a personalized information service, here are the options:
- Align with Facebook, Omnicom and Atlas, adopting Facebook Connect as a standard for identifying user interests and attributes.
- Align with Google and its Android Google Ad ID, building and managing user profiles from there.
- Piggyback on the identity information collected by Apple’s iIOS IDFA approach.
- Work with the DigiTrust consortium to develop a consortium-owned standard that is not controlled by a single platform and is aligned with the interest of advertising-tech companies.
- Collaborate to develop a new standard which focuses on consumer privacy protection and interoperates with Facebook, Google, Apple, DigiTrust and evolving NSTIC-compliant services, sharing useful aspects of the Internet2 Shiboleth single-signon service.