Media: The PPM Has "Absolutely No Value"?

TORONTO, ONTARIO - Long-time readers of this blog may recall some discussion of the Arbitron Portable People Meter (PPM), the new device that the radio ratings services company has been intending to replace diaries as a way of deriving ratings for radio stations. When I last explored this subject back in March, there were growing concerns that reduced sample size might be creating a problem for Arbitron.

I'm a bit late in reporting this, but it appears that sample size is indeed a big deal. In an open letter to Arbitron last month, KGO-KSFO General Manager Mickey Luckoff, arguably the most successful station boss of this era, specifically decried reduced sample size. In terms of the promised ability to better refine what caused a listener to change the station, Luckoff stated that the PPM-derived data provided "absolutely no value in making any such decision with any degree of accuracy or reliability." (I suspect any accurate mechanism would just tell them that the frequency was changed at a commercial break anyway.)

How bad is the problem? Harker Research's Radio Insights looked at what they expected to be a typical case. They found that during morning drive time, the key time for radio listenership, the median station in a major market had either one or zero PPM listeners for 70% of that time period. That's not a sample, that's a single listener. Luckoff clearly had a point in his open letter.

Morning drive in general has been an issue with the PPM--more people claimed to be listening to the radio in their diaries than the PPM is reporting. In another Radio Insights posting, they showed a 12% decline in listeners in the top 3 US markets during morning drive, according to Arbitron's publicly-released data. For an industry starving for advertising money, every listener counts, and a 12% drop is significant. It's so significant, in fact, that the ratings service in the United Kingdom dropped the PPM over the decline in morning ratings, which it claimed was not real.

As if the sample size and morning listener decline problems were not enough, there has long been a perception that minority listeners were not adequately represented. While I have yet to see any data on this issue, Reuters has reported that the Media Rating Council has found "persistent problems" with minority representation in the PPM data. Frankly, if sample sizes are as low as they seem to be, then it's not surprising that there are minority representation problems--there are majority representation problems as well.

The really interesting thing about the sample size problem is that it is not fundamental to the technology--it's only that Arbitron is trying to save money by having fewer PPM units in the field. The morning drive problem, on the other hand, may be a technological issue--people may not find the devices easy enough to integrate into their lives in the morning.

It has been noted that in the United Kingdom, the decision was made to abandon the PPM in favor of on-line diaries, which were deemed easier to use than the traditional written diaries. If I were an Arbitron competitor, I'd be rushing to present an on-line diary system to radio stations--the cost savings alone (Aribitron is charging about 50% more for PPM-derived data than diary data) would justify a switch, never mind the other issues.