The data could have a profound effect on the way advertisers and agencies value media, and on how media outlets sell, and even how they are programmed. By showing how advertising run on various outlets, programs and programming genres impacts consumer purchase decisions, the data could create new advertising metrics. Such data, which Arbitron once dubbed "buyergraphics" when it was developing a similar single source system years ago, theoretically could replace conventional age and sex demographics that are now the basis of media deals. Instead of buying TV programs based on the number of adults 18-49 who watch it, for example, advertisers might purchase them on the basis of how many orange juice drinkers watch the show.

"You will receive reports for 20 categories and 50 brands," Arbitron and Nielsen informed its media partners in letters sent to participating networks this week. "These will include consumer packaged goods and other important advertising categories, such as financial services, automotive, and cinema/movies."

Advertisers sponsoring the pilot study have already had preliminary data for more than a month, and are also receiving a much deeper set than is going to participating media outlets. Concurrent with the data being released to the media this week, those marketers will receive a new set of data, according to executives familiar with the plans. Those marketers include Procter & Gamble, Unilever, Kraft, SC Johnson, Pepsi and Pfizer, which constitute Apollo's steering committee, and whom are developing a plan for rolling the system out into tens of thousands of households nationwide beginning next year, assuming the pilot study is successful.

The pilot is based on a sample of 5,400 households and 11,000 persons six years or older that have agreed to utilize Arbitron's portable people meters to measure their media usage, and to have their household product purchases scanned via ACNielsen's Homescan system.

The companies have also been conducting a series of online surveys of those panel members to glean other important information about media consumption, lifestyles and attitudes.

Test results released to date indicate that the pilot has good representation of the U.S. population and that compliance levels have also been good. Top line data indicates profound differences between conventional Nielsen demographic ratings and the composition of actual brand purchasers for most programming genres. Actual brand purchasers, for example, can index three times higher than women 25 to 54 for cable network news programs, according to preliminary results released in June. The data also revealed pronounced differences between certain product categories, and the purchasers of specific brands.

The data being released to participating networks this week will include information for the following 10 consumer packaged goods categories: Diet Cola, Soup (Canned), Pizza (Frozen), Candy (Chocolate), Snacks (Tortilla Chips), Detergents (Heavy Duty Liquid), Coffee (Ground and Whole Bean), Light Beer, Batteries and Pain Medication (Headache). Another set of reports scheduled to be released later this year will contain brand-specific information, followed by non consumer packaged goods information.

This news story is important. I recommend that everyone read it.