How PPM Could Rescue Radio

I’m not one to worry about the distant future. Remembering to pick up my laundry is challenge enough. Then Steven Spielberg casually mentioned civilization’s inevitable move from a carbon to silicon base-his matter-of-fact way of saying when robots take over the Earth. By then, data is not only the new creative, he is the new creative director.

Man vs. Machine: There are many things machines do far better than people. In media, the robotic radio PPM compared to the all-too-human diary is a good example. The vagaries of the diary measurement may be costing radio millions each year in a simple but hidden way. Diary reporting is no longer adequate for how audience estimates are used to select media.

Marketing Mix Modeling: Today, many advertisers leap over conventional measurements like audience, demos and CPMs, and go directly to consumer response to make their media decisions. The tool of choice is complex Marketing Mix Modeling. Advertisers take the pieces of last year’s brand marketing spend and match that to brand sales to see how they track.
For media, the deciding measure is its contribution to total brand sales, minus the cost of goods, divided by the cost of the medium. It is the equivalent of advertising-delivered profit before taxes or “Payback.” You can’t argue with the goal or the model. Both seem to work. It’s the marketing input data that need attention, especially the radio data.
Why Radio Should Win: Years of marketing mix studies have uncovered two planning truths. All marketing expenditures show diminishing marginal response. Each additional dollar spent in a medium usually pays back less than the one before. This argues against media concentration and supports media mix. The second truth is each week added to a schedule usually pays back more than the week before. This recommends continuous advertising.

Both findings suggest brands should shift marginal TV dollars to other media — 20 percent to radio for example — to improve total media payback. The dollar shift works in three ways:

• Reducing TV dollars should increase TV-generated payback per dollar. (Remember, each added dollar pays back less.)
• Radio’s lower spending level should generate payback at a higher point on radio’s payback curve, making it more efficient than other media used more heavily.
• Radio’s lower cost will buy additional weeks, which should improve total campaign payback.

There is supporting data for this theory.

An MMA study of multiple brands analyzed by John Phillip Jones in The Ultimate Secrets of Advertising shows a medium’s rank in payback was the reverse of its rank in spending. In Jones’ examples, radio, with the lowest share of dollars, produced the highest payback. Some marketing mix studies also show this higher radio payback pattern, but many others don’t. There could be several reasons for this, ranging from the radio creative to the inadequacy of the data used to represent radio. I think data may be the problem.

Modeling Tracks Change: Marketing Mix Modeling works by linking changes in advertising weight to changes in brand sales. In the case of media, if the audience data fed into the system are overstated or averaged rather than time specific, the causal link between changes in media exposure and changes in sales can be lost. Diary recall tends to exaggerate listening to leading stations and the data is reported as audience averages from the 12-week survey. Weekly data is available, but seldom used.

The use of diary data or the alternative radio dollars spent tend to flatten the audience delivery highs and lows marketing mix models need to work effectively.  In contrast the PPM, now in most major markets, can provide measured week-by-week schedule audience delivery over 52 weeks of the year. It will be interesting to see whether radio’s marketing mix payback numbers improve as more PPM markets are installed.
RAB may throw a welcome party for robots.