The so-called Apollo system was first proposed in September 2004 by Arbitron and VNU and endorsed mightily by the biggest of all advertisers, Procter & Gamble--but has yet to get any other backers due to the exorbitant costs the companies were asking marketers to pay for the service. Arbitron and VNU have subsequently retooled their pricing strategy for Apollo and have been pushing marketers to commit to a national rollout of the service, which would equip 30,000 households with ACNielsen's product scanner technology and 70,000 people with Arbitron's portable people meters (PPM). ACNielsen is a unit of VNU, and is the sister company of Nielsen Media Research, which has an option to develop a media ratings service based on the PPM technology, but so far has been loath to do so.

Because of the extraordinary costs of rolling out a national single-source measurement system--estimated to be well in excess of $100 million--neither Arbitron or VNU has been willing to commit to deploy the service on the assumption that if they build it, marketers will actually come. In fact, Arbitron sank more than $100 million into the development of a ScanAmerica, a similar system it created in the 1980s, but shut it down when advertisers, agencies, and media companies refused to pay for the service.

The terms of the pilot are still being devised, and Arbitron and VNU plan to get back to the industry in 30 days with a proposal that would outline its scope, including a timetable, the geographic coverage, the size of the pilot sample, the project's deliverables, and its costs.

By empirically linking exposure to advertising across a wide array of media with the products purchased in the households of people exposed to those ads, supporters--especially P&G's marketing team--believe Apollo can answer the age-old question in the advertising business: What was the return on my advertising investment? Depending on the way it is ultimately structured, Apollo might be able to give marketers access to such data in real-time, or at the very least close enough to actual advertising patterns so that they could alter their creative strategies, media schedules, and promotional offerings on the fly--fundamentally changing the way media is planned, bought, and sold.

There are many implications of such a system for marketers, agencies, and the media alike. For marketers, it might mean little or no waste in their advertising executions, making TV and radio advertising more akin to the super-efficiency of online media, especially search.

For agencies, the emergence of such a powerful single-source measurement system would create a new level of accountability, forcing agencies to revise advertising messages and media strategies on a much more dynamic basis and with much greater proof of results. Some agencies fear this might make the process too scientific, and that the art of advertising and media strategies could be lost.

In fact, Apollo is just one step in a progression toward greater scientific proof and accountability of advertising and media buys. Marketing Management Analytics (MMA), a unit of Aegis Group that is closely aligned with Carat, recently developed a real-time application for its marketing mix modeling system that theoretically could give marketers and brand managers the ability to look at advertising and media results on the fly and make changes accordingly. MMA was one of the leaders of such modeling, and its pioneering work with major packaged goods marketers such as Kraft transformed the way many of them and their agencies advertised and used media. But the original version of the system required a great deal of planning, development, and data-gathering, making real-time analytics and decision-making impractical. It was used more for long-range planning and fine-tuning ad decisions after the fact.

The new Web-based system, dubbed Avista, will give marketers the ability to do that dynamically, each day, by looking at results and playing with alternate scenarios on their desktops, says John Nardone, the former media director at results-oriented interactive agency Modem Media, who joined MMA last year to develop the new product. That product is out of beta, and MMA is pitching it to marketers now.

For media, all of these developments have huge implications that could literally reshape the advertising marketplace--making media outlets more accountable to advertising results, whether they like it or not. On Wednesday, during an interview at the ANA forum, Viacom's Leslie Moonves said he did not believe media outlets like TV networks were accountable for the effectiveness of the ads that agencies and advertisers place in their programs.

The development of more dynamic measurement systems that can track those results could change all that. And if the media doesn't take a direct role in shaping such ROI, advertisers and agencies likely will do it themselves--buying the media that work for them and steering away from those that don't.

One obvious implication of such data that could also transform the media marketplace is the way marketers and agencies target media audiences. Historically, media sold--and occasionally guaranteed--audience delivery based on age, sex, income, and educational demographics, or geographic indices. With product usage-based data, marketers may ask the media to do it on the basis of product preferences. Arbitron's ScanAmerica system even coined a name for such audience breaks, dubbing them "buyergraphics."

Marketers who have been using online media already have had a taste for such targeting, and probably have an appetite for more. Another Internet-based form of targeting that is expected to have a major impact on traditional media, especially television, is behavioral targeting, or a method that targets consumers based on their media usage behavior. Instead of paying through the nose to buy ads reaching viewers of CBS' pricey "CSI," a marketer utilizing behavioral targeting would identify the other programs those "CSI" viewers watched at other times.

Leading online behavioral targeting companies such as Tacoda and Revenue Science have plans to move into TV when that data becomes available, while some TV technology players, including Invidi, are in the process of developing an infrastructure that could enable that.

"I think behavioral [targeting] is the ideal for television," said Patrick McKenna, manager of marketing communications for BMW North America, during Wednesday's ANA forum. However, he added: "I like the analogy of crawl, don't run."

Meanwhile, the advent of a single-source measurement system like Apollo could have another profound implication for planning and buying media. Because it would measure the effectiveness of more than one medium at a time, it would allow marketers to look at results across an array of media options. In other words, they could begin targeting consumer behavior independent of a specific medium such as TV or radio, but based on what the consumers they were seeking to reach actually do.

Initial plans for Apollo include the measurement of both TV and radio, although Arbitron has been aggressively developing ways of measuring other media--including print, outdoor, and the Internet--via the system.