Rob Piotrowski
Star Participant
Fashionable enough to earn her a spot in the National Radio Hall of Fame. Induction ceremony in November at the Museum of Broadcast Communications in Chicago.
-
Get involved.
We want your input!
Apply for Membership and join the conversations about everything related to broadcasting.
After we receive your registration, a moderator will review it. After your registration is approved, you will be permitted to post.
If you use a disposable or false email address, your registration will be rejected.
After your membership is approved, please take a minute to tell us a little bit about yourself.
https://www.radiodiscussions.com/forums/introduce-yourself.1088/
Thanks in advance and have fun!
RadioDiscussions Administrators
Bloomberg interview with Delilah
- Thread starter Rob Piotrowski
- Start date
Regardless how anyone might feel about Delilah, this is a well-written article with an excellent explanation of how PPM works, and what has happened to the radio business due to this insanely inaccurate system of measurement. It is obvious that no one involved in the development of PPM thought about asking a statistician about sample sizes and randomization. Diaries weren't great, but PPM is worse. Making major format decisions based on a few strapped-on samples from the same few listeners for long periods of time is absurd.
For some obscure reason, Arbitron sent us encoders for four of my stations several years ago, which I dutifully installed. In my unrated market, there are no PPMs, only a miniscule number of diaries. When Nielsen recently decided to "improve" their boxes (you know, the ones that they said didn't need improvement), I made a decision to remove my encoders. It took over six months to even get a reply on my multiple requests to return the 12 units, and now I've been waiting another two months for the empty boxes so that I can return them. One would hope that at some point Nielsen will admit that they were snookered.
For some obscure reason, Arbitron sent us encoders for four of my stations several years ago, which I dutifully installed. In my unrated market, there are no PPMs, only a miniscule number of diaries. When Nielsen recently decided to "improve" their boxes (you know, the ones that they said didn't need improvement), I made a decision to remove my encoders. It took over six months to even get a reply on my multiple requests to return the 12 units, and now I've been waiting another two months for the empty boxes so that I can return them. One would hope that at some point Nielsen will admit that they were snookered.
Actually they did. There was a very long process, and a very long discussion about it. The subject of sample size was brought up, lots of statisticians weighed in on the subject, and like everything, there are lots of conflicting results. The conclusion is that Arbitron was perfectly willing to increase the sample size, but it would cost more money, and that cost would have to be paid by the subscribers. So if you or anyone else wants a larger sample size, you have to pay for it. How do you feel about that?
The development of the PPM began in the very early 90's at the behest of the advertising community which wanted faster and more granular results. But it was not until the Internet became widely popular that the buyers pushed more vigorously for electronic measurement for traditional media which was being "left behind".
In 2008 when the rollout of PPM began for real, the change cost stations 40% to 60% more than their deal with Arbitron at the time. Sample size was discussed, but the idea of more than doubling the cost of ratings was acceptable to no one, particularly in the first year of a recession.
The meter, which measures only time spent listening and cume, is much better than the diary, which measures TSL, cume and memory.
And that is Delilah's problem in PPM. Listeners in the diary world "remember" listening to Delilah "all evening" and write down long continuous periods of time, such a 8 PM to 11 PM.. In reality, that listening may have started at 8:35 and ended at 10:12 and been interrupted by phone calls, and other normal activities. So, in reality those three hours were really about 70 to 80 minutes of actual listening. The PPM shows that, while the diary does not.
Last edited:
Maybe this has been brought up before, but one concern I have with the PPM is that it doesn't know whether you're actually paying attention as far as I know. For example if you have the radio on while doing other stuff around the house, the meter will think you've listened all day, when in reality the radio was on but you may not have been paying attention at all. How does the meter measure that and how do you program to such an audience? My concern here is this is part of where the more music less talk approach got started, but that may not be true at all. Or how about when, for example, I have the meter but I'm in the car with my parents, who change the station at every break, and they listen to stations that I'll tolerate but wouldn't have picked?
Ratings are done to sell advertising, Advertisers pushed for the PPM because they don't care why a person listens, but, rather, how many impressions their ads make.
In other words, there is no concern by advertisers as to the choice of station. They just want to know how many pairs of ears heard their spots.
In the early years of the PPM, I was on a committee of ad agency buyers and radio researchers that Arbitron put together to determine if an engagement metric was needed. The conclusion was that such a measure would only make ad buying harder for the agencies and would not benefit radio.
Stations making programming decisions, if they don't do proprietary research, generally look only at the "productive" cume of a station. We know that, on average, about 50% of a station's cume gives us about 90% or more of the listening time... so we can look at the data for the "useful" cumers only in the Nielsen software. As much as half a station's only listen an hour or less a week... so we don't make decisions based on them.
PTBoardOp94
Walk of Fame Participant
They sent you encoders because you're close enough to the Seattle-Tacoma market that a few panelists might listen to your stations. Probable that those listeners will never make you any money though.
So, in other words, the subscribers were not willing to pay for a sample size that would offer a higher degree of accuracy? And on that basis, Arbitron was willing to produce a report that was not accurate?
Like I said, an obscure reason. Likely that one of my stations showed up once in Thurston County. The signals don't cover other areas. KBKW received the encoders, it's a 1 kw AM station that doesn't under any conditions cover more than the western end of Grays Harbor County.
Randomization does not exist when the participants keep the PPMs for long periods of time. They are only measuring the small subset of the population blessed with the meters. At least the diaries measured a different random sample each time, and one could see trending from different random samples. True however that diaries tended to be filled out just before mailing.
The PPM sample turns over about 8% a month. THe most a household can remain in the sample is two years.
Agencies wanted electronic measurement, fast data delivery and a panel like the meter panel for TV. The reason for a panel is that it can be nearly perfectly balanced for proportionality on multiple stratification variables. This eliminates the uncertainties of the diary method, where different age, gender and ethnic subsets can be significantly undersampled, thus requiring significant weighting.
The advantages of much better proportionality far outweigh the randomless nature of diary placement and return; a true random probability sample would be proportional... but the diary sample never turns out that way.
Keep in mind that increasing "accuracy" by one standard error requires a quadrupling of the sample size. So adding a few hundred to a sample of 3000 (LA's sample) would have little effect, but would increase costs by around 10%.
The accuracy of the PPM, with a proportional panel and electronic monitoring of compliance, is greater than that of the diary. And with the rapidly increasing difficulty in recruiting diary participants in the big, highly stratified markets, the PPM was the only way that Arbitron could have gone. Keep in mind, too, that the PPM daily sample is about five to six times larger than in the diary sample... enabling things like measurement of one-time events like sports events or feature weekends and specialty shows... even topics on a talk show.
The reason why PPM is not in any but the top 50 markets is that it requires non-stop measurement and a minimum number of panelists. Some markets 41-100 do not even have continuous diary measurement due to cost, and local markets that size just can't afford the technology.
How much would YOU be willing to pay for more accuracy? The information they have now is acceptable to both the stations and the advertisers, and that's all that matters. They arrived at the current system after much discussion. Several companies threatened not to participate. In the end, they did because the advertisers demanded it.
Anything is a function of how much someone is willing to pay. I could have a much nicer house, but the one I live in serves the purpose. Same thing with ratings.
Bongwater
Walk of Fame Participant
Sooooo.....Are you saying this thing knows WHEN you're on the phone?
No, but if you turn down the radio to take a phone call, it no longer detects a station being listened to.
- Status