"More marketers are using [marketing-mix analytics] not just to look at media, but to look at individual ads," said Mike Hess, director-global research and communication insights at Omnicom Group's OMD Worldwide, New York. And one of them is Procter & Gamble Co., said Ted Woehrle, VP-North American marketing. "We can now look at individual television spots," he said, "and see which ones performed."
Indeed, not looking at the return-on-investment performance of individual ads can lead to bad media decisions, he said. "You don't just draw a blanket conclusion that print works or TV works or radio works. You have to dig deeper." Likewise, he said, using marketing communications audits based on consumer research are an important check on marketing-mix analytics. A weak ROI for print might lead to a conclusion to cut spending there. But if the communications audit shows consumers respond to print, or the brand needs print to convey a complex message, cuts to spending "would be a bad choice," he said. "You should instead go back to the agency and say, `What can we do to make print more engaging?"'
Another finding from P&G's growing use of marketing-mix models: Marketing-mix results correspond to upfront copy-test scores. This may be bad news for agency creatives, who for eons have been hoping to make marketers less reliant on copy testing, but marketing-mix models are just confirming their usefulness to marketers. "If anything, [marketing-mix models have] increased my confidence in a lot of the predictive tools," Mr. Woehrle said.
John Kastenholz, a former Unilever VP for consumer and market insights and now a consultant, has been critical of copy testing measures such as related recall, saying they miss the emotional aspects that build brands. But he agreed Unilever's marketing-mix results generally have corresponded with copy-test scores.
Still, he believes copy testing that's misused can hurt advertising creative. He believes in using it later in the process, and without researchers prescribing creative solutions. "The challenge," he said, "is how to do a better job of using research to help advertising rather than as a stick to beat agencies with."
He noted that varying ROI for different ads also points to an inherent weakness of marketing-mix analytics-that it doesn't broadly account for the quality of marketing inputs. Even if marketers can do that with their own TV ads using their copy-test numbers, they can't readily do so for all the competitive ads or other media that don't get tested. Yet impacts of all marketing programs by marketers and their competitors figure into marketing-mix results.