PPM is a game-changer in many respects. One of the least discussed aspects of this is radio’s loss of history. We have 40 years of diary based listenership data that we are essentially throwing away in this transition.
It might seem a minor point, given the industry’s limited attention span and our tendency to look no more than a book or two back. After all, you’re no better than your last book, but the fact remains that we are losing perspective and our history.
We were reminded of this recently by a blog post that misused some recent Arbitron numbers to caution radio about a decline in cume, writing:
There is a slow but steady slippage in weekly Cume Ratings. While radio rightfully touts itself as a reach (or Cume) medium, we shouldn’t be complacent about weekly Cume Ratings levels slipping from 95.3% in Arbitron’s Fall 1998 figures to 92% in the most recent RADAR study. The younger demos are cause for particular concern.
The problem is that the blog was comparing Arbitron’s 1998 American Radio Listening Trends to the latest RADAR study. You can see the latest Arbitron report here and the RADAR press release here.
There are multiple errors committed in the analysis. The American Radio Listening Trends are based on listenership in 98 diary measured metro Arbitron markets. RADAR uses a 330,000 person DMA sub-sample of Arbitron’s diary and PPM measured markets.
The differences between the two sets of in-tabs may seem minor, but the analysis expresses concern about a 3% drop in cume over a decade. A small drop like this could easily be a result of mixing methodologies and have nothing to do with an actual decline in cume.
We know that PPM numbers differ from diary numbers. Perhaps the decline is a result of including RADAR PPM data. Perhaps the decline is a result of drawing a separate sub-sample for RADAR. Perhaps the declines are real. The bottom line is that we have no way to tell.
As more markets transition to PPM, it will become even more difficult to draw meaningful conclusions about trends in radio listenership. And it couldn’t come at a worse time.
We recently pointed out here that media usage is not a zero-sum game. People can increase their usage of new media without listening less to radio. Despite that, there is an assumption and belief in many quarters that radio is in decay. New-media pundits are absolutely convinced that people are abandoning radio in droves. “Nobody listens to radio anymore!” Even many people who work within this business believe it.
If we had consistent Arbitron data, we could refute these assertions, but we don’t. And it isn’t just PPM that is causing the problem.
Arbitron’s changes to the diary method are also causing problems. Over the past decade, Arbitron has been changing the rules. To lower costs, diaries that were once rejected are now ok. We have new editing rules that might have a minor impact for individual stations, but have significant implications in our ability to trend national numbers over time.
While Arbitron changed some rules to save money, they kept other rules to save money. In 2003, only 3% of households were cell-only households. The number has steadily grown since then, and now the estimate is something like 20%, considerably higher than that in young and Hispanic households.
This creates a problem for Arbitron and their recruitment of participants. They call households using land-line telephone lists supplied by a third-party company. That means Arbitron is systematically ignoring a growing segment of listeners.
The inability to recruit young cell-only households made Arbitron listening estimates more and more suspect as the years went on. It is quite possible that radio’s apparent declining cume and TSL is related to the failure to include cell-only households.
Nielsen’s diary test in Lexington, Kentucky suggests this might be the case. According to Nielsen:
The Lexington study found that:
- Cell-phone-only homes logged nearly 23 hours of radio listening per week compared to just over 19 hours for the total sample.
- Listen to 3.5 radio stations compared to less than 3 stations among total sample
- Has an AQH total radio rating of 17.3% versus 14.3 rating for the total sample
- Skews younger, primarily between the ages of 18 and 34.
So we have to wonder about Arbitron numbers over the past several years. What would the national listenership numbers have been had Arbitron had a representative number of cell-only households? Maybe the “loss” of younger listeners is related to the absence of cell-only households.
Arbitron wasn’t going to address the problem until Fall 2009, but pressure caused them to accelerate the process. We’ll start seeing the impact this Spring, and Arbitron has already released preliminary estimates that confirm Nielsen’s findings.
We have the transition to PPM, which distorts long term trends. We have changes in diary editing rules that distort long term trends. We have Arbitron’s unwillingness to recruit a proportionate number of cell-only households. All these things make it impossible to draw meaningful conclusions about radio listenership trends.
If you are in a PPM market, the transition is like starting over. You might as well throw those diary based books away, because all the numbers that preceded your first PPM book are meaningless.
From this day forward, when you read about “trends” in listenership, be very skeptical. We have given up our history.
Comments