Apollo aimed to determine once and for all how exposure to a wide variety of media and marketing tactics influence purchases by tracking consumers' combined media and purchasing habits in a single-source database.
In and out of home
The effort used Arbitron's Portable People Meters to track a wide variety of media exposure in and out of home among member's of Nielsen's household panel, who scanned all of their purchases. But although the service attracted seven major package-goods backers -- most notably charter client Procter & Gamble Co. -- Apollo was unable to win sufficient client commitments to foot the substantial cost of a national rollout.
Arbitron and Nielsen sank more than $45 million combined into the program in the form of losses and upfront investments, based on financial disclosures by the publicly held Arbitron, which owned half the joint venture. That includes what Arbitron Chairman-CEO Steven Morris said on a conference call Feb. 14 would be a $4 million charge for his company's share of the costs to shut the program down.
P&G, the client at whose urging Apollo was initiated, was believed to have fronted another $20 million to launch it in 2005, according to one person familiar with Apollo. A P&G spokesman called that number "way high," but declined to specify the contribution.
"We respect the decision," the P&G spokesman said, "but we're disappointed Project Apollo will be discontinued, and we remain committed to finding new methods to help us make effective and efficient consumer-centered media decisions. We had seen encouraging benefits from Apollo during our involvement, but we were still in the process of making our final decision."
The tens of millions spent on the experimental program is a princely sum for market research. P&G, believed to be the world's biggest spender on market research, spends about $200 million annually worldwide for all of its survey research.
First scaled back in 2006
The move comes about a week ahead of the March timetable Mr. Morris has laid out for a final decision on Apollo. Backers already had scaled back plans in 2006 to a pilot in about 6,000 households from original plans for a 50,000-household national program.
Apollo was ambitious enough that coverage of it jumped beyond the often insular world of market research, garnering coverage in The New York Times magazine and Business Week.
The program was a prime example of P&G's "connect and develop" approach to innovation through outside partners. Its kickoff in late 2004 was one of the rare times, also including a 1998 summit on internet marketing, when it hosted competitors at its Cincinnati headquarters.
Ultimately some other stalwarts of package-goods marketing, including Unilever, Kraft Foods and SC Johnson, also became Apollo clients.
At one point in early 2005, backers were seeking commitments of 0.5% of clients' marketing budgets for the program, according to people familiar with it. That would have amounted to seven-figure commitments for any marketer spending more than $200 million annually and eight-figure commitments from the biggest spenders.
"Everyone recognized from the outset what an ambitious effort we were undertaking," Mr. Morris said in a statement.
Susan Whiting, exec VP of Nielsen, said, "We have learned a great deal from Project Apollo."
Continue to mine data
And, indeed, the learning can continue for at least a while as clients will be allowed to continue to mine data generated by the program since its pilot began two years ago, said an Arbitron spokesman.
In an effort to build interest among a broader range of clients, Apollo backers had released data in recent years that had piqued curiosity, if it didn't ultimately open wallets sufficiently.
A project last year demonstrated conclusively what researchers long had believed based on prior research -- that TV ads really can reduce consumers' sensitivity to price and price promotion.
Another project in 2006 for a P&G brand, believed to be Tide, showed that concentration of a brand's target consumers can vary widely even within different programs in narrow media classifications, such as women's cable programming. Then again, it also showed the "Brand X" in question was already doing a remarkably good job of identifying those media buys.