Cracking the ROI conundrum

By Published on .

It could be the Holy Grail of ROI-but, just as with the legend, many skeptics doubt that such a measurement miracle will ever come to pass.

VNU and Arbitron unveiled grand plans last week for a long-elusive tool that will marry media exposure and purchase data to measure return on marketing investment.

Procter & Gamble Co. is the first to sign on to provide marketer input. Greg Ross, director-North American media and marketing, said the tool has "potential for a breakthrough in consumer understanding about how and when consumers are most receptive to brand information."

But while other marketers and agency executives told Advertising Age they love the concept, they also found numerous devils lurking in the details, including unsolved issues around the introduction of Arbitron's Portable People Meters and long-standing questions about the reliability of research panel data.

"Conceptually, it sounds like a nirvana tool," said Larry Koslow, VP-global lens care marketing, Novartis' Ciba Vision. "But there are so many variables that go into that kind of analysis, the question is will it be credible when it comes out?"

The proposed venture, code-named Apollo, would combine consumer media usage data from Portable People Meters with purchase data from VNU's ACNielsen HomeScan consumer panel and other sources, such as survey research. A "small national pilot" test is already underway, an Arbitron spokesman said, and a broad rollout could come next year.

Mr. Koslow said that prior to 2001, when Wal-Mart Stores still provided scanner data to syndicators, he often found significant differences between retail scanner data and data provided by HomeScan. That could indicate a margin of error in the panel data exceeding the sales impact of a given ad.

The Arbitron spokesman acknowledged Apollo's data is likely best suited to big brands and categories shopped by larger numbers of consumers, to reduce the margin of error.

Noreen Simmons, director-strategic media planning at Unilever, said her biggest concern is that asking HomeScan panelists to take on an additional responsibility will cause some to drop out. "Part of the value of this panel over time," she said, "is the fact that people tend to stay in the panel and their purchase habits can be tracked over time."

The two companies said an unspecified number of the 30,000 Apollo panelists would be drawn from a HomeScan panel that more than doubled its size to compensate for the loss of Wal-Mart data. As part of their HomeScan duties, for which they're paid, panelists are asked to scan everything they buy from a store when they get home; and some submit to additional surveys. The burden involved in scanning purchases already leads some researchers to suspect that panels under-represent purchases.

The Arbitron spokesman said the two companies are exploring the panel burden factor in their pilot.

Agency executives also noted technical issues with PPMs, which require that a special signal be encoded in audio tracks of TV or radio programs. The ads themselves would need to be encoded, and not all TV networks have yet agreed to transmit the signals.

other applications

While PPMs are used primarily to track TV and radio usage, they have the capability of monitoring anything that has an audio track, including cinema ads and some retail displays, the Arbitron spokesman said. He said the companies are exploring ways to track magazines and outdoor ads.

Mr. Hess said similar "single source" initiatives failed in the past two decades because clients simply weren't willing to pay their considerable cost. "Maybe this time around it will make it because of the added pressure" on accountability, he said. "Everybody's talking ROI these days."

In this article:
Most Popular