Following The Bouncing PPM Data

Todd Wallace, DJ/PD/Programming Consultant/Researcher/Publisher known as the 'Father of Callout Research', which he pioneered in early 1970s writes to correct my semantics. In yesterday's post, I termed fluctuations in PPM weeklies "wobbles." Todd points out that a better word for it may be "bounce."

While the target sample-size in meters for most major markets is larger than the average number of monthly diaries were, it is much smaller than the quarterly total of diaries and unique households sampled over the 84 days of the quarterly survey in a typical Arbitron diary “book.”

What this means: while there will be less sampling “wobble” than we would have seen with diary monthly-isolations, there may be more such “bounce” when compared to a full quarterly diary report.

Ratings Bonuses In A PPM World

PPM Strategist and longtime radio research analyst Charlotte Lawyer and I had an exchange 60 days ago on how bonus structures may change for on air talent in PPM:

"I had a program director ask me about using aqh ratings. I always used AQH persons rank or share goal in diary, and/or cume for some stations depending on format and goals. With ratings so compressed now you can have maybe a .3 rating separating #1 and #10. I suggested he use aqh persons because of the compression and consider adding TSL. This PD's stations (spoken word) have great tsls in ppm. Do you have any thoughts?"

At the time we talked about it, but I didn't blog it, thinking that it was a pretty narrow subject area for universal interest. Then came "KRBE, Houston’s Ryan Chase (says) a temporarily messed-up sample due to Hurricane Ike cost him “a sixth of my income.”) and the online chatter about Ryan Seacrest and Kevin & The Bean.

What to do when the change wrought by a new method of measurement is no longer motivating, but actually frustrating, demotivating, to talent?

Samples are one-third of what they used to be, shares are too compressed, hair-raising wobbles appear to be weekly events and in some markets format shares for whole groups of stations and styles of personalities seem to be turned upside down.

I agree with the idea of bonusing on AQH person rank rather than share, but even that can seem unfairly capricious, given that it only takes a handful of panelists to dislike something to doom a station or talent.

So, how about trying this on for size: bonus talent on a metric they can control, comparing to their past trend rather than other stations or personalities on the metric which drives PPM ratings, the number of listening occasions in their daypart?

Get your panelists to come back to the station more times, earn a bonus. Have panelists listen for fewer occasions this month compared to last month, lose a bonus.

To be motivating, a bonus needs to reflect something talent can control. To do well, you incent the behavior you want.

Do you have a better idea? I'd love to hear it.