The Dark Inky Mystery That is PPM

There’s a certain irony that Arbitron would roll out  PPM 360, Arbitron’s “sleeker, smarter audience measurement solution,” almost exactly a year after the FCC announced their inquiry into PPM.

While the responses to the announcement were mostly enthusiastic, the upbeat stories seemed disproportionately positive given the limited improvements. This new generation of PPM addresses a few superficial and cosmetic issues–the sort of thing that gives the illusion of progress.

Unfortunately, this roll-out continues Arbitron’s efforts to put a happy face on PPM while keeping broadcasters in the dark about far more critical issues, all more important than whether the meter should have a sculpted look.
 
Soon the top 50 markets will all be PPM markets, yet we know no more about PPM than when it was first announced. Even today, important technical aspect of PPM that directly impact your ratings are shrouded in dark inky mystery, hidden from view.

Broadcasters are expected to accept virtually all aspects of PPM on faith, with nothing more than Arbitron’s assurances that everything works like they say it does.

One year ago we highlighted a long list of PPM issues that Arbitron had been unwilling to address publically. Today, one year later, nothing has changed.

As we wrote in May 2009:

Radio has not been served by blind acceptance of Arbitron's PPM. Arbitron's rush to monetize PPM combined with radio's unquestioning naivete regarding PPM has stifled the debate on many PPM issues. It is never too late to start asking the important questions.

So in honor of PPM 360, we are revisiting the many issues Arbitron has refused to address.

Arbitron talks a lot about PPM granularity. According to the company, we can make programming decisions based on minute to minute changes in listening detected by PPM.

When we look at the minute by minute, we find something odd.

Arbitron claims that most listeners don’t switch from one station to another. They listen briefly to a station, and (according to PPM), simply turn off the radio. The vast majority of listeners do not try even a second station let alone surf the dial. This is the so-called drive-by listening that seems to pervade PPM ratings.

According to Arbitron VP of Programming Services Gary Marince, the average PPM radio listening span is ten minutes, and the most frequent length is only two minutes!

We have been studying radio listeners for more than thirty years. We know that real listeners do surf the dial. They switch from one station to a second, sometimes to a third to find something interesting. They don't listen to a station for two minutes and then turn off the radio.

If PPM says they do, then there is something wrong with PPM.

What is the true real-world ability of  PPM to accurately capture all listening. Arbitron has never released test results regarding capture reliability of the meter. Can Arbitron prove that panelists are really turning off the radio, or is there a problem with the meters?

On a related issue, we observed that unlike American broadcasters, European broadcasters insisted on testing PPM, and the testers did something curious. They had panelists carry the meter on the outside of their clothing.

The tests went so far as to provide the participant special slings and belts so that clothing wouldn’t impair reception. See photos from one test here.

Why did panelists wear their meters around their necks or in pouches? Why didn’t they wear them as normal people would?

Is it because if they did slip it into a pocket or purse the meter wouldn’t register? What about a business man wearing the meter under his jacket?

How loud does the station have to be played for the meter to register in a purse or under a jacket? Should every panelist dress like Mr. T shown above to make sure their listening gets registered?

How much listening is lost just because people aren’t wearing the meter around their neck?  How often does the meter fail to count what the user is listening to because ambient noise drowns out the station?

Arbitron isn’t saying. 

Then there is the issue of Arbitron’s refusal to let broadcasters see listening patterns, so called respondent level data. Diary market stations can go to Arbitron headquarters and look at the diaries. There is no equivalent with PPM, yet Arbitron could easily create a PPM Mechanical similar to the diary mechanical.

The only explanation is that Arbitron doesn’t want you to see what’s really going on with panelists.

Arbitron promised that the creation of panels would solve many of the problems of the diary. Wobbles would disappear and extreme weighting would be a thing of the past with PPM. We showed that wobbles haven’t gone away. In many cases wobbles are even greater with PPM.

 What went wrong?

It is all well and good that Arbitron has given the meter a more modern sleek look. Enabling panelists to connect with a cell phone instead of a land line is progress, and Arbitron is to be commended.

But the cloak of secrecy that Arbitron hides behind regarding the most critical aspects of PPM should be a red flag to broadcasters that there are far bigger issues impacting your PPM ratings.

 

Comments

[Randy Kabrich]

Radio might not understand why they do not see the button pushers, but its easy to understand once you understand that Arbitron cannot track anything less than 30 seconds in length.

Code Organization: CBET codes will be transmitted as often as every 4.8 seconds in a continuous
audio broadcast. While a panelist’s Meter is constantly capturing exposure data, the Meter stores data approximately
every 15 seconds. Thus, a Meter has up to three “chances” every 15 seconds to detect a complete code.
Arbitron’s media data processing system organizes stored Meter data in 30-second increments, which reflect the characteristics of the last code of the highest quality received after combining the two 15-second segments into one.
This data organization structure provides
further redundancy in the editing process.

So Arbitron combines two 15-second segments into one 30-second segment, but they roll it up and only expose the minute-level data.

However, that misses button pushing entirely - combined with missing 41% of the listening, according to the RAJAR Valadation tests - and OF COURSE it will appear that a listener turns OFF the radio instead of going to another station!