Digital is consuming more of Procter & Gamble's media budget and has become so important that the world's biggest advertiser is fundamentally changing how it creates those ads and applying what it learns to other parts of the marketing mix.
P&G's optimization approach for digital is not unlike the continuous improvement processes long used on factory floors, and some of its brands, including Pantene, are using continuous data on consumers' response to digital ads to tweak media buys and inform creative elements.
For about a year, Pantene has been using such a system, Smart Media, developed with Resource Interactive, Cincinnati. Smart Media analyzes click-through rates and flash surveys on purchase intent across numerous permutations of ads and placements.
The billion-dollar hair-product brand has improved performance on those target metrics by 28% to 90% vs. the previous system of creative pretesting alone, according to Marketing Director Kevin Crociata.
The program evaluates three creative elements (the headline, hair visual and background color), and Pantene makes changes based on how different combinations of each element perform in various media placements.
The brand makes discoveries that inform future creative. Those have included that white backgrounds don't work well on Yahoo, orange is effective on Facebook and blondes get a better response than brunettes on some sites.
"Background color is big," Mr. Crociata said. "It sounds pretty simple, but it can have a dramatic impact on purchase intent."
Increasingly, what P&G finds isn't limited to digital or Pantene. Mr. Crociata is applying Smart Media to the Herbal Essences brand and has shared the approach with P&G's corporate marketing group. He's also using the digital optimization system to assess how various ad claims might work in other media.
"If we take one of the three or four claims we have for a particular initiative and find one is working particularly well in the digital space, it's learning we can apply back to print ads, in-store elements and other places," Mr. Crociata said.
The information also influences Pantene's social-media content, he said.
"I love the real-time element of this and being able to work more efficiently based on what we're learning in digital," Mr. Crociata said. "It's not just to figure out the plan and let it run for six months and do an analysis. Every week we have meetings as a team to see what we're learning from social [and] from Smart Media, and we're working to optimize it."
Smart Media scales some traditional organizational and disciplinary walls in the process. Since it simultaneously deals with both media and creative issues, media shop Starcom MediaVest Group and creative agency Grey Global Group have also been involved, he said.
Agency creatives haven't resisted the approach, Mr. Crociata said, "but we've had to change the way we work" by developing three options for background colors or claims upfront rather than settling on just one before launching ads.
"The number of permutations is great in total, but it's pretty simple for the creative process upfront, because we break it into those three key elements of the headline, the hair visual and background colors," Mr. Crociata said. "We can create about 12 individual units, and then they can be optimized."
The creatives at Resource "are very open to real-time learning and getting the best out of our media vehicles. But it does take some planning up front. You've got to prepare so you have enough variation to change up as you go."
According to Rex Briggs, CEO of Marketing Evolution, Smart Media sounds similar to experiments that P&G conducted eight years ago for Olay and Pantene. Those tests, which involved a variety of permutations to optimize digital-media placements, were part of a broader effort by marketers and formed the basis of Mr. Briggs' book, "What Sticks."
While Mr. Briggs said that he believes the approach is better than applying TV-style copy-testing to digital creative -- as P&G and some other marketers have tried -- he wonders how much further along P&G would be today had it kept doing continuous optimization of digital ads in the intervening years.
Another issue: P&G is trying to restrict how many times any individual consumer gets surveyed by brands in a research-quality effort, and the optimization system can end up generating a lot of surveys. So P&G may need to find another behavioral yardstick. One possibility some marketers are exploring, Mr. Briggs said, is mouse heat mapping, which tracks people's mouse movements as a proxy for attention.