And lo, a new Leviathan was born. Boasting 500m detailed consumer profiles, Xaxis must be the biggest marketing database in the world and, in the words of Jakob Nielsen, managing director, interaction, at GroupM, 'a paradigm shift' in many respects.
Xaxis, with its terminology of audience versus media-buying, marks a fundamental shift in power from media owners to data owners. Once, the former were powerful gatekeepers to the audiences advertisers desperately needed to reach. By tracking individuals as they travel from one website or page to another, however, Xaxis is simultaneously unbundling the media owner's carefully constructed package; it can tell it exactly which pages are working.
At the same time, it is obliterating the sanctity of media-owner properties by buying audiences across multiple online media properties and distribution platforms such as the web, paid-for social media and mobile. Meanwhile, it remains agnostic about media owners' claimed audience USPs.
This shift 'from placement to audiences' is fundamental, notes Nielsen, and, as more media (especially TV) goes digital, this shift in power can only accelerate, he argues.
At the same time, Xaxis is allaying advertiser fears of a parallel shift in power by creating multilevelled data-crunching.
At the core lies data aggregated from all WPP's campaigns for clients (probably about a third of the global market). This database is enhanced by information purchased from third-party data harvesters such as BlueKai. These are then matched against the client's own data which, crucially, remains separate and proprietary. In this way, Xaxis promises advertisers the best of both worlds: the richness and reach of a vast internet-wide database, plus the unique details from each brand's insights into its own customers.
Not surprisingly, even WPP's goliath 'frenemy' Google is reported to be rushing its own response. Before online behavioural advertising (OBA) ventures such as Xaxis can be declared a success, however, they must clear two hurdles: public acceptance and the underlying economic case.
Privacy campaigners have already dubbed the database 'the Xaxis of evil'. Nielsen counters that consumers need more educating. It insists that Xaxis holds 'no personally identifiable information', yet this information is identifiable enough to serve up ads targeting an individual. This blurring of the distinction between anonymous and personally identifiable - the grey area of 'pseudonymity' - is increasingly exercising the minds of regulators.
A week after WPP announced Xaxis, the European Data Protection supervisor, Peter Hustinx, addressed an eprivacy conference. He raised two pivotal OBA questions: what is 'personally identifiable data', and what is user 'consent'?
Opt-out falls short
Take the way publishers pass on details of customers' online behaviours to ad networks. Many of these customers are indeed personally identifiable: they have logged in to a service. Now details of their behaviours are being passed on with neither their knowledge or consent.
Nielsen argues that this has no bearing on Xaxis because 'it is publishers' responsibility to decide what they do with their data'. In his speech, however, Hustinx stressed that the EU's eprivacy directive considers the storing of information and accessing of information stored in a user's terminal as an intrusion into the user's privacy; and that it requires the user's consent before information is stored or accessed.
He went on to warn that the IAB's current self-regulatory response - a clickable icon on behaviourally targeted ads with an option to opt out - 'fails to implement the new consent requirement'. This is because when consumers opt out under the IAB scheme, they cannot opt out of being tracked: the tracking continues unchanged. The opt-out relates only to ads served up on the basis of this tracking.
Describing current developments in OBA as 'an erosion of fundamental rights and a market failure', Mr Hustinx called for the implementation of 'privacy-by-default' settings, by which third-party cookies are rejected unless the user decides otherwise.
Nielsen admits that this would kill his business model. According to Dr Ian Brown of the Oxford Internet Institute, however, this exposes a contradiction in the industry's stance. 'If consumers really want behaviourally targeted advertising as much as advertisers claim, they cannot plausibly claim that working on an opt-in rather than an opt-out basis would kill the industry. They can't have it both ways,' he says.
This opt-in/opt-out debate 'is now clearly very high on the EU regulators' agenda' adds Dr Brown. In Europe, at least, 'business models that rely on advertisers using technology to collect information about people without their consent are not sustainable. Being able to opt out does not count for consent'.
The second key question is the economic case. A slide (see above right), provided by DoubleVerify at a recent industry conference, maps current players in the OBA industry. Perhaps this vast complexity and cost is necessary. Perhaps it's just the first flowering of a new industry before rationalisation leaves a few leviathans like Xaxis ruling the roost. Or perhaps it illustrates an underlying problem.
In the 11th century, the Chinese took water-clock technology to new heights. The clocks were vast edifices, sometimes 10m tall, requiring exquisite engineering to manage their complexity. They told the time, but not very accurately and at great cost.
Back to before square one
OBA is the internet's Chinese water clock. It has a fundamental design flaw. If customers simply tell you what they are interested in, or want, you can serve up the right information at the right time, which in turn builds trust. It was this insight - in the form of the search term - that propelled Google from zero to hero.
OBA, however, jettisons this insight. Instead of involving customers, it tracks them to make guesses about what they are doing. This creates the need for all manner of data-crunching to reduce the errors and waste created by the original guesswork.
If customers were volunteering information about their intentions, goals, preferences and purposes, the industry might not need all the demand-side platforms, data management platforms and related data collection, aggregation, verification, modelling, analytics and optimisation activities identified by DoubleVerify.
Right now, the industry is making bold claims about the uplifts it can deliver. Yet doubling response rates doesn't look so wonderful if you are doubling from 0.25% to 0.5%. What about the other 99.5%? As Nick Jones, director of interactive services at the COI points out, the same breakthrough claims were once made for direct mail and the plain old clickthrough ... before the novelty wore off.
Henceforth, brands will need to consider two questions. First, will stalking customers ever build trust? Never mind the technicalities and legalese, just consider the emotions. Second, as we enter deeper into the information age, will consumers become less or more active in managing their own information?
If the answer is 'more active', then, strategically, the challenge for brands is to develop knowing, active, participatory, trust-based information-sharing relationships with customers. Yet OBA drives brands in the opposite direction. Or, to put it another way, in the emerging environment, the very notions of 'audiences' and 'targeting' - with all their connotations of subject/object relationships; of active doers and passive done-tos - could be reaching their sell-by date. If so, Xaxis might just be an elegant solution to the wrong problem.
Alan Mitchell is a respected author and a founder of Ctrl-Shift and Mydex.