Shocking statistics

[bierkenstock]

Yesterday the New York Times published an article on editorial job cuts at newspapers around the country. In making comparisons to other media, the article pointed out only 21 percent of adults listen to morning drive radio. The number has to be significantly lower in other dayparts but in radio's prime time, it only reaches one out of five people.

The 12+ AQH numbers posted online don't reflect this. These numbers report shares - the size of the slice each station receives; they don't show that the pie keeps getting smaller.

The article also reports that advertisers are expected to shift even more of their budgets to "new media" (possibly up to 20% of total spending three years from noew and some major advertisers going with even higher percentages). That means less money for "old media" - including radio.

Radio management appears to respond to the current situation with cost-cutting strategies (automation, satellite-delivered programming, job cuts, consolidation). In many declining industries, price increases are also used to off-set declining sales revenues (and thus hasten the decline). This does not appear to be an option for radio. However, Clear Channel is pushing the industry to reduce sales inventory (possibly intending to decrease audience erosion and push up spot prices).

Radio reinvented itself 50 years ago. It looks increasing like it won't be able to do so a second time.

 

[954]

market niches they currently ignore

> Yesterday the New York Times published an article .... In
> making comparisons to other media, the article pointed out
> only 21 percent of adults listen to morning drive radio.
> The number has to be significantly lower in other dayparts
> but in radio's prime time, it only reaches one out of five
> people.

Could part of this be because they are ignoring the
over 54 audience -- and they're listening to iPods,
CDs, cassettes, Sirius, or XM instead? (Maybe even
a few 8-tracks!)

After all, there are still some advertisers
interested in over 54s -- in addition to Geritol,
Viagara, cemeteries, condominiums, chiropractors,
and Medicare supplements.

I'll soon be in that category and I still eat,
drive, read, write, watch TV, and a few other
things I'm not yet too feeble and decrepit for.

If total listenership is declining, maybe they
should quit cherry-picking and try to appeal to
the market niches they currently ignore.

What would it cost them? Only their pride.

73s from 954
<P ID="signature">______________
<center>South Florida Radio Pages</center></P>

 

[davideduardo]

Re: Totally and Miserably Wrong Data.

> Yesterday the New York Times published an article on
> editorial job cuts at newspapers around the country. In
> making comparisons to other media, the article pointed out
> only 21 percent of adults listen to morning drive radio.
> The number has to be significantly lower in other dayparts
> but in radio's prime time, it only reaches one out of five
> people.

This is simply wrong. The cume rating of radio, across the country, is around 80. That means that 80% of Americans listen to radio in Morning Drive. A significant portion of those who don't (75% of the rest) do listen, but in other dayparts.

Roughly 95% of Americans listen to radio in some daypart. Those that do not listen, for example, in mornings, are nearly all people who have diffeent work shifts and are not available in mornings.

75% of all Americans listen in 10-3. 78% listen in the afternoon.

Here is what I think the Times got wrong. In most markets, the total percent of people listening to radio at any one time during 6-10 is around 21 or 22%. That means that at a given time during the morning, about 1/4 of people are listening. But, during the course of the morning, over 80% listen at some time. So the Times was using share instead of cume!

Radio reaches 4 out of every 5 persons in morning drive, but at any given moment, the average is about 21% who are listening. This is because nobody listens all 4 hours of morning drive, but nearly everyone listens to part of it.
>
> The 12+ AQH numbers posted online don't reflect this. These
> numbers report shares - the size of the slice each station
> receives; they don't show that the pie keeps getting
> smaller.

But the cume share numbers that Arb itron gives as "Market Totals" in each metro disprove these numbers... totally.
>
> The article also reports that advertisers are expected to
> shift even more of their budgets to "new media" (possibly up
> to 20% of total spending three years from noew and some
> major advertisers going with even higher percentages). That
> means less money for "old media" - including radio.

No, it means less share of the growth. Radio has moved to slow grwoth, as most mature businesses have. Still, we see some markets, like LA, up 4% this year.

>
> Radio reinvented itself 50 years ago. It looks increasing
> like it won't be able to do so a second time.

We can not fix things that are not broken...
>

 

[davideduardo]

Re: market niches that are unprofitable

> Could part of this be because they are ignoring the
> over 54 audience -- and they're listening to iPods,
> CDs, cassettes, Sirius, or XM instead? (Maybe even
> a few 8-tracks!)

No, that is not the case because the reach, or cume, of radio is not down in this demo or any other adult demo... in fact, it is hardly down in teens! The Times, simply, got it all wrong. They confused share in mornings with cume, and printed the share number as if it were cume.

> After all, there are still some advertisers
> interested in over 54s

Not enough, in larger markets, to make money. Take WDUV in Tampa. #1 12+ and below #15 in billing... in a market with a higher percentage of retirees than nearly any place in the USA.

>-- in addition to Geritol,
> Viagara, cemeteries, condominiums, chiropractors,
> and Medicare supplements.

You named them all. There is nearly nothing else, besides early bird specials
>
> I'll soon be in that category and I still eat,
> drive, read, write, watch TV, and a few other
> things I'm not yet too feeble and decrepit for.

It has nothing to do with being feeble. It has to do with it costing too much per sale to get new business from people with more established consumer behaviour.
>
> If total listenership is declining,

Which it isn't

> maybe they
> should quit cherry-picking and try to appeal to
> the market niches they currently ignore.

How can radio serve audiences when no advertiser will be interested in that particular audience. Go complain to P&G, not radio.
> What would it cost them? Only their pride.
>
> 73s from 954
>

 

[SaltyDog]

Re: Totally and Miserably Wrong Data.

> Here is what I think the Times got wrong. In most markets,
> the total percent of people listening to radio at any one
> time during 6-10 is around 21 or 22%. That means that at a
> given time during the morning, about 1/4 of people are
> listening. But, during the course of the morning, over 80%
> listen at some time. So the Times was using share instead of
> cume!

EVERY time the press gets its hands on a subject I know something about, they get it wrong. It makes me wonder what I read about which I have no knowledge that I am mistakenly believing.. but that's another thread for another board.

Leave it to the NY Times to measure radio using AQH when THEY sell ads using cume, or as they call it, circulation!

I'd rather have our problems than their problems.

<font size=4>CNN: New York Times to cut 4% of workforce</font>

Newspaper publisher to lay off about 500 employees as part of effort to improve efficiency.

September 20, 2005: 5:33 PM EDT

NEW YORK (Reuters) - New York Times Co. said on Tuesday it plans to cut about 4 percent of its work force and warned of lower-than-expected profits as it faces weaker advertising across its core newspaper operations.

Full Story HERE<P ID="signature">______________
"A man is about as big as the things that make him angry" - Winston Churchill

<a href="http://saltydog.5gigs.com">
The Salty Dog</a>
</P>

 

[davideduardo]

Re: Totally and Miserably Wrong Data.

> > Here is what I think the Times got wrong. In most markets,
>
> > the total percent of people listening to radio at any one
> > time during 6-10 is around 21 or 22%. That means that at a
>
> > given time during the morning, about 1/4 of people are
> > listening. But, during the course of the morning, over 80%
>
> > listen at some time. So the Times was using share instead
> of
> > cume!
>
> EVERY time the press gets its hands on a subject I know
> something about, they get it wrong. It makes me wonder what
> I read about which I have no knowledge that I am mistakenly
> believing.. but that's another thread for another board.
>
> Leave it to the NY Times to measure radio using AQH when
> THEY sell ads using cume, or as they call it, circulation!

If the Times sold ads based on thesame AQH standard, which would be "page reads" they would be seriously screwed.

 

[Keith_Lake]

Re: market niches that are unprofitable

What we have here is an amorphous wad of numbers that everyone interprets differently. You show the same ratings to three different people and you'll get
three different interpretations. Part of the "art" of being in sales or management is taking rotten numbers and putting a positive spin on them.

Mark Twain said it best: "There are lies, damned lies, and statistics."

Hear hear.


KL

<a href="http://home.nc.rr.com/gttyson/lastradio.html">The Last Radio Station<a>

 

[davideduardo]

Re: The Times just used the wrong figures...

> What we have here is an amorphous wad of numbers that
> everyone interprets differently. You show the same ratings
> to three different people and you'll get
> three different interpretations. Part of the "art" of being
> in sales or management is taking rotten numbers and putting
> a positive spin on them.

There is noting ambiguous about rinning the market totals and showing that the NY Times piece was, simply, wrong.

All the numbers I gave in this thread are total market averages around the US.

Morning drive radio reaches about 80% of all Americans sometime in the 6 to 10 AM window. On average, about 22% of all Americans are listening at any given average moment in the daypart. There is no other way of interpreting this.

Different markets vary by a percent up or down. That's about it.

And, still, the problem is that the Times should have used cume figures, sicne they were talking about cume conceptually.

 

[Keith_Lake]

Re: The Times just used the wrong figures...

>
> There is noting ambiguous about rinning the market totals
> and showing that the NY Times piece was, simply, wrong.
>
> All the numbers I gave in this thread are total market
> averages around the US.

Emphasis on the word "average".
>
> Morning drive radio reaches about 80% of all Americans
> sometime in the 6 to 10 AM window. On average, about 22% of
> all Americans are listening at any given average moment in
> the daypart. There is no other way of interpreting this.

Again, that's an average. Some stations own the market in morning drive time, some don't. And saying "there is no other way of interpreting this" illustrates my previous point rather well.
>
> Different markets vary by a percent up or down. That's about
> it.

I would be willing to bet it's much more than plus or minus one point up or down.
>
> And, still, the problem is that the Times should have used
> cume figures, since they were talking about cume
> conceptually.
>
What you read was *their* interpretation of the numbers. Show the same set of numbers to another group of people and you will get an entirely different interpretation. To attribute it to "conceptual cume", whatever THAT is, only further clouds the issue. Radio management today is too hung up on numbers that can be too easily influenced by many different factors (diary placement for one) when they should be concentrating more on what's going out over the air. Program what the people in your market want to hear and the numbers will take care of themselves.

KL

<a href="http://home.nc.rr.com/gttyson/lastradio.html">The Last Radio Station<a>

 

[JohnnyMorganWXJX]

Yet more evidence:

A) The NY Times' editors are idiots and/or buffoons; and

B) You can take data and make it say whatever you want (even if it's wrong).

 

[davideduardo]

Re: The Times just used the wrong figures...

> >
> > All the numbers I gave in this thread are total market
> > averages around the US.
>
> Emphasis on the word "average".

You are trying to make a case where none exists. I already specified that every market is about the same, and the difference have to do with commute patterns, etc., and are less than a percent up or down.
> >
> > Morning drive radio reaches about 80% of all Americans
> > sometime in the 6 to 10 AM window. On average, about 22%
> of
> > all Americans are listening at any given average moment in
>
> > the daypart. There is no other way of interpreting this.
>
> Again, that's an average. Some stations own the market in
> morning drive time, some don't. And saying "there is no
> other way of interpreting this" illustrates my previous
> point rather well.

We are NOT talking about stations. The NYT did not talk about individual stations. Why did you bring that extraneous issue into this discussion, which is about the usage of the medium?

Again, we are talking about usage of the radio, not stations. The morning usage of the medium was specified by the NYT at 21% when, in fact, it is around 80%, give or take a percent, in every market. The average is 80%, the range is about 79% to 81%.


> > Different markets vary by a percent up or down. That's
> about
> > it.
>
> I would be willing to bet it's much more than plus or minus
> one point up or down.

It isn't. I looked at 20 random markets using a 5 book average, all markets in the top 100 and got the lowest at 78.6 (Austin, TX) and the highest at 81.2% cume rating. The individual books range form about 76 to just over 83, but the average is in the 79 to 81 range. Remeber, this is a random probablilty sample, and there is a margin of error. All the variances are within a margin of error centered at 80% usage of radio by persons 12+ in all the markets I looked at.

There was one exception. Las Vegas. At 77% average. Of course, the other dayparts are much higer than the national averages, since the market operates on a different work schedule due to the local economy... there is a drive time at 2 AM whent he casinos change shifts!
> >
> > And, still, the problem is that the Times should have used
>
> > cume figures, since they were talking about cume
> > conceptually.
> >
> What you read was *their* interpretation of the numbers.

No, what I read was not an interpretation. They said that 21% of people used radio in Morning drive, and that figure is the AQH average, not the cume rating. In other words, they used the wrong figure. 80% of Americans use radio in the morning, and at any given time in the 16 quarter hors of morning drive, about 21% are listening (higher in 7 AM quarter hours, lower in 9 AM quarter hours.)

There is no interpretation here. They used the wrong figure because, obviously, they did not find out what each number means.

> Show the same set of numbers to another group of people and
> you will get an entirely different interpretation.

Wrong. Ask about "how many people use radio" and you are specifically asking a cume question. Ask about how many are listening at any given moment, and you have an AQH persons question.

> To
> attribute it to "conceptual cume", whatever THAT is, only
> further clouds the issue.

There is nothing "conceptual" about the very easy to find market cume rating. It means, "how many different people used radio in the time period being inspected." That figure is an average of 80% of all 12+ persons, nationally.

> Radio management today is too hung
> up on numbers that can be too easily influenced by many
> different factors (diary placement for one) when they should
> be concentrating more on what's going out over the air.

Diary placement does not affect market cume. It may affect the cume and shae of stations that have more or less signal in a particualar area, but not the market usage. There is proportionality in the Arbitron survey, and variances are almost entirely due to statistical wobble and margin of error, not the survery itself. If we could afford bigger samples, the wobble would be reduced.

> Program what the people in your market want to hear and the
> numbers will take care of themselves.

As they are.

 

[Keith_Lake]

Re: The Times just used the wrong figures...

Looks like we're going to have to agree to disagree on this one. I maintain that ratings are notoriously inexact and are subject to all sorts of outside influences on their way to the final figures. You could get similar results with a dart board. Unfortunately, until someone comes up with a scoreboard-type device that shows exactly how many people are tuned to your frequency at any given moment, they're all we've got. I think it's ridiculous that formats and careers live and die based on estimates and extrapolations of vague categories of numbers that are only seen by people inside the walls of the stations. All too often this creates a disconnect between the station and the audience because the programmers are obsessed with these nebulous numbers while the listeners just want to hear their favorite song. If they don't hear it within a few minutes, off they go channel-surfing.

Most markets nowadays are saturated with stations that all sound more or less alike. This leads to a considerable margin of error in filling out the diaries that the station owners pretend does not exist. As long as the Average Listener
hears his/her favorite song, they couldn't care less whether it's on Z-92 or Mix 106 or whatever.

KL

<a href="http://home.nc.rr.com/gttyson/lastradio.html">The Last Radio Station<a>

 

[davideduardo]

Re: The Times just used the wrong figures...

> Looks like we're going to have to agree to disagree on this
> one.

That is an easy cop-out when you want to ignore the facts. Fortunately, America's advertisers, for whom Arbitron is designed, don't ignore the validy of the ratings.

> I maintain that ratings are notoriously inexact and
> are subject to all sorts of outside influences on their way
> to the final figures.

Which is, then, why ad agencies and advertisers use them as a metric for over $22 billion dollars in radio advertising each year?

Any poll has statistical variation. The degree of such variation, often called "wobble," is considered small enough to make Arbitron ratings usable as the standard for pricing models and ad buys all over the US, Mexico and Puerto Rico, and the diary method acceptable in about 50 countries all over the world.

> You could get similar results with a
> dart board.

No, you couldn't. The Arbitron survey is, additionally, audited by the MRC, a group composed of statisticians, researchers and ad agency people and they are audited at length each year and the methodology has been rigorously certified.

> Unfortunately, until someone comes up with a
> scoreboard-type device that shows exactly how many people
> are tuned to your frequency at any given moment, they're all
> we've got.

The Portable People Meter does that, with daily results downloaded to Arbitron. Just like TV meters. The interestingthing is that shares do not change between the diary method and the PPM. One is just faster and is better on achieving constant proportionality in smaller time frames.

> I think it's ridiculous that formats and careers
> live and die based on estimates and extrapolations of vague
> categories of numbers that are only seen by people inside
> the walls of the stations.

The ratings are supposed to be mostly a sales tool. They are copyright. Anyone who works with them knows they are accurate within the constraints of the sample size and methodology.

> All too often this creates a
> disconnect between the station and the audience because the
> programmers are obsessed with these nebulous numbers while
> the listeners just want to hear their favorite song. If they
> don't hear it within a few minutes, off they go
> channel-surfing.

There is concordance. Stations determine best songs with other kinds of research, and measure effectiveness of the whole station by Arbitron. Arbitron does not measure the like and dislike of songs. It is not supposed to.
>
> Most markets nowadays are saturated with stations that all
> sound more or less alike. This leads to a considerable
> margin of error in filling out the diaries that the station
> owners pretend does not exist.

There is very little margin of error in the diary process, except that, as proven by the PPM, very short incidents of listening are missed... People are very accurate in putting down what they listened to... about 80% of diary mentions include the exact frequency, so they must know.

> As long as the Average
> Listener
> hears his/her favorite song, they couldn't care less whether
> it's on Z-92 or Mix 106 or whatever.

Not true. People listen to stations that do the best job of continuously entertaing, whether with amusic mix, talk, or personality. Since only about 30% of listening is in cars, where push buttons make changing easy, 70% of listening is on radios where tuning another station is not worth the effort.

 

[SaltyDog]

Re: Totally and Miserably Wrong Data.

> If the Times sold ads based on thesame AQH standard, which
> would be "page reads" they would be seriously screwed.

They are already screwed. More bad news just came out.

<font size=4>Average age of newspaper readers: 55</font>

Eric Black, Star Tribune

October 11, 2005 MEDIA1011

Newspaper readership is down. Fewer young people are picking them up, and the average age of a newspaper reader is now 55, according to a Carnegie Corporation study. Many papers have been losing circulation at alarming rates across all age groups. Full Story HERE<P ID="signature">______________
"A man is about as big as the things that make him angry" - Winston Churchill

<a href="http://saltydog.5gigs.com">
The Salty Dog</a>
</P>

 

[RockManAC]

I'd believe it. I know more people who're listening to their iPods, CDs, Satellite Radios or Internet webcasts more than they are traditional AM or FM stations.

-A<P ID="signature">______________

</P>

 

[Keith_Lake]

Re: The Times just used the wrong figures...

> > Looks like we're going to have to agree to disagree on
> this
> > one.
>
> That is an easy cop-out when you want to ignore the facts.
> Fortunately, America's advertisers, for whom Arbitron is
> designed, don't ignore the validy of the ratings.

What makes them so above question? So far I've heard nothing that suggests ratings are less than largely inexact and have a bigger margin of error than they're willing to admit.
>
> Which is, then, why ad agencies and advertisers use them as
> a metric for over $22 billion dollars in radio advertising
> each year?

Well, they have to have something to convince ad buyers on the local level that many thousands of people are actively listening when the actual number is realistically in the high hundreds. On the national level the numbers they claim are in the millions but undoubtedly with the same exaggeration.
>
> Any poll has statistical variation. The degree of such
> variation, often called "wobble," is considered small enough
> to make Arbitron ratings usable as the standard for pricing
> models and ad buys all over the US, Mexico and Puerto Rico,
> and the diary method acceptable in about 50 countries all
> over the world.
>
> > You could get similar results with a
> > dart board.
>
> No, you couldn't. The Arbitron survey is, additionally,
> audited by the MRC, a group composed of statisticians,
> researchers and ad agency people and they are audited at
> length each year and the methodology has been rigorously
> certified.

They're all in it together. One hand washes the other. Most local management folks will believe anything that comes out of a tony NYC or LA bureau where the people wear expensive suits, have opulent offices and maintain some sort of connection with the stock exchanges.
>
> > Unfortunately, until someone comes up with a
> > scoreboard-type device that shows exactly how many people
> > are tuned to your frequency at any given moment, they're
> > all we've got.
>
> The Portable People Meter does that, with daily results
> downloaded to Arbitron. Just like TV meters. The
> interestingthing is that shares do not change between the
> diary method and the PPM. One is just faster and is better
> on achieving constant proportionality in smaller time
> frames.

I've been reading up on the PPM. It does seem capable of providing more exact figures, but does it measure Ipods and other personal playback devices?
>
> > I think it's ridiculous that formats and careers
> > live and die based on estimates and extrapolations of
> > vague categories of numbers that are only seen by people inside
> > the walls of the stations.
>
> The ratings are supposed to be mostly a sales tool. They are
> copyright. Anyone who works with them knows they are
> accurate within the constraints of the sample size and
> methodology.

Which is a cool way to say they don't have exact numbers, or at least not as exact as they'd like to believe.


If the listener doesn't hear their favorite song within a few minutes, off they go channel-surfing. The SEEK button on the car radio is your worst enemy. Listeners to news/talk stations might stay on a station a bit longer, but if the
subject gets boring, or the station goes into a marathon commercial break, that listener is gone too.

> There is concordance. Stations determine best songs with
> other kinds of research, and measure effectiveness of the
> whole station by Arbitron. Arbitron does not measure the
> like and dislike of songs. It is not supposed to.

True. For songs they have such bizarre methods as auditorium testing and focus groups, which is an entirely different area of weirdness we'll save for another thread.
> >
> > Most markets nowadays are saturated with stations that all
> > sound more or less alike. This leads to a considerable
> > margin of error in filling out the diaries that the
> > station owners pretend does not exist.
>
> There is very little margin of error in the diary process,
> except that, as proven by the PPM, very short incidents of
> listening are missed... People are very accurate in putting
> down what they listened to... about 80% of diary mentions
> include the exact frequency, so they must know.

Maybe it works that way in your market, but I've seen more than one instance here of diaries being delivered to the wrong market, wrong county, etc., with the result being terribly skewed numbers. What do they do in such cases?
>
> > As long as the Average Listener hears his/her favorite song, they couldn't > > care less whether it's on Z-92 or Mix 106 or whatever.
>
> Not true. People listen to stations that do the best job of
> continuously entertaing, whether with a music mix, talk, or
> personality. Since only about 30% of listening is in cars,
> where push buttons make changing easy, 70% of listening is
> on radios where tuning another station is not worth the
> effort.

"Not worth the effort". I rest my case.


KL

<a href="http://home.nc.rr.com/gttyson/lastradio.html">The Last Radio Station<a>

 

[KMRichards]

Re: The Times just used the wrong figures...

I'm going to ask a single question.

> > Fortunately, America's advertisers, for whom Arbitron is
> > designed, don't ignore the validy of the ratings.
>
> What makes them so above question? So far I've heard
> nothing that suggests ratings are less than largely inexact
> and have a bigger margin of error than they're willing to
> admit.

As long as the agencies <u>do</u> accept the ratings, what makes you think your questioning them matters?<P ID="signature">______________


</P>

 

[Keith_Lake]

Re: The Times just used the wrong figures...

>
> As long as the agencies do accept the ratings, what makes
> you think your questioning them matters?
>

Okay, you got me there. As said previously, I think the Portable People Meter system is going to go a long ways towards minimizing the margin of error which prompted my original posting. It will be interesting to see how quickly the industry accepts it because it also accomodates webcasts, Ipods and personal playback devices. I'll bet we see some stations accustomed to big numbers get mightily embarrassed when the margin of error is erased.

KL

<a href="http://home.nc.rr.com/gttyson/lastradio.html">The Last Radio Station<a>

 

[davideduardo]

Re: The Times just used the wrong figures...

> > > Looks like we're going to have to agree to disagree on
> > this
> > > one.
> >
> > That is an easy cop-out when you want to ignore the facts.
>
> > Fortunately, America's advertisers, for whom Arbitron is
> > designed, don't ignore the validy of the ratings.
>
> What makes them so above question?

They aren't. That is why the Media Research Council, funded by advertisers and agencies, audits and certifies both Arbitron and Nielsen and renews this via a long, long audit every year.

when some of the most brilliant statasticians and researchers, defending the interests of advertisers, sign off on Arbitron it makes me pretty confident we have a good company doing as good a job as possible.

> So far I've heard
> nothing that suggests ratings are less than largely inexact
> and have a bigger margin of error than they're willing to
> admit.

You have not heard that from anyone who knows anything about statistics, polling or research.

Statistics is the only science where "error" is not a dirty word. Any poll has a margin of error. There are several variables that determine margin of error, and Aritron shows you how to calculate them for any given population, sample size and audience size.

> >
> > Which is, then, why ad agencies and advertisers use them
> as
> > a metric for over $22 billion dollars in radio advertising
>
> > each year?
>
> Well, they have to have something to convince ad buyers on
> the local level that many thousands of people are actively
> listening when the actual number is realistically in the
> high hundreds. On the national level the numbers they claim
> are in the millions but undoubtedly with the same
> exaggeration.

There has been no accusation of this in any circle of people in the ad business or the research business or the media... or, even in academia. The only time I ever hear this sort of accusation of exaggerated numbers is in newsgroups where people with no knowledge of statistics and research start pontificating based on no knowledge and a lot of supposition.

The fact is, I have replicated Arbitron's results by conducting a smaller sample survey using a different methodology but the same principles of random probability sampling, and gotten results within the margin of error that statistically match those of Arbitron very closely. In fact, I do that every week in nearly 20 markets.
> >
> > > You could get similar results with a
> > > dart board.
> >
> > No, you couldn't. The Arbitron survey is, additionally,
> > audited by the MRC, a group composed of statisticians,
> > researchers and ad agency people and they are audited at
> > length each year and the methodology has been rigorously
> > certified.
>
> They're all in it together. One hand washes the other.
> Most local management folks will believe anything that comes
> out of a tony NYC or LA bureau where the people wear
> expensive suits, have opulent offices and maintain some sort
> of connection with the stock exchanges.

Arbitron is in rural Maryland, is significantly employee owned, and has no Wall Street focus... and I go there a couple of times a year. the offices are, shall we say, spartan... unless you find cheap industrial carpeting and linoleum floors to be tony.

In any case, ad agencies and advertisers audit Arbitron to make sure the costs for advertising can be applied to some form of reliable metric for cost determination. This is why they have Nielsen, the ABC, etc.

Advertisers and the media are essentially adversarial. Agencies would like nothing better than to find something that would lower the costs of media. Your suggestion that media and advertisers are in cahoots is, simply, immensely laughable. This is like saying that the Yankees trained the Angels' pitchers last night.

> > The Portable People Meter does that, with daily results
> > downloaded to Arbitron. Just like TV meters. The
> > interestingthing is that shares do not change between the
> > diary method and the PPM. One is just faster and is better
> > on achieving constant proportionality in smaller time
> > frames.
>
> I've been reading up on the PPM. It does seem capable of
> providing more exact figures, but does it measure Ipods and
> other personal playback devices?

Ratings meassure media usage. They do not meassure TOT (Time on Toilet) or SOI (Time on iPod) or TOD (time on drugs). Advertisers could care less how much time is spent gaming, on web pages, on iPods or listening to 45 rpm singles.

Media measurement is done to determine how many people are listening to a station when the spot runs. It does not matter what non-listeners are doing. What matters is how many listeners there are, and the rate being charged so the advertiser can determine the cost per impression.
> >
> > The ratings are supposed to be mostly a sales tool. They
> are
> > copyright. Anyone who works with them knows they are
> > accurate within the constraints of the sample size and
> > methodology.
>
> Which is a cool way to say they don't have exact numbers, or
> at least not as exact as they'd like to believe.

They are accurate within very small margins, and those margins can be determined by the formulae Arbitron gives in the "Purple Book" which is available to all subscribers and licensed users.

Advertisers realize that ratings numbers are within a narrow range, and are not exact. Nothing short of a census can be exact, and that is impossible to do.

Nobody says the ratings are 100% precise. They don't have to be. The purpose itself is to assist in buying ad time, not in placing a smart bomb inside a hidden bunker.

Why are you criticizing ratings, which are a form of a poll, for bieing exactly what they are supposed to be: very close estimates of consumer behaviour.
>
>
> If the listener doesn't hear their favorite song within a
> few minutes, off they go channel-surfing.

Actually, talking to real listeners will show they punch in the car if they hear a song they don't like, not the absence or any particular one they do like.

> The SEEK button on
> the car radio is your worst enemy.

Less than 30% of listening is in-car in the USA. 70% or more is in homes and offices, while people are working, getting dressed, having breakfast, etc. People in such situations pick a favorite station and leave it. They seldom jump around.

In general, adults do not do that much punching around. They have a favorite station or two, usually mood determined, and use them when they feel like using each one.

> Listeners to news/talk
> stations might stay on a station a bit longer, but if the
> subject gets boring, or the station goes into a marathon
> commercial break, that listener is gone too.

Not usually. There is huge cosistency in talk programming.

By the way, what are your qualifications to make all these blanket (and wrong) statements about ratings, statistics and listener behaviour? I have never heard so much wrong data in so little time in my whole career.
>
> > There is concordance. Stations determine best songs with
> > other kinds of research, and measure effectiveness of the
> > whole station by Arbitron. Arbitron does not measure the
> > like and dislike of songs. It is not supposed to.
>
> True. For songs they have such bizarre methods as
> auditorium testing and focus groups, which is an entirely
> different area of weirdness we'll save for another thread.

There is nothing weird about AMTs or perceptual research. All are techniques adapted from the giants of marketing like P%G and such. And the techniques work... I have seen a staiton turn from, for example, 8th in a market to 1st with just a music test.

You seem to distrust things you don't understand. Sort of like the way I feel about the IRS.
> >
> > There is very little margin of error in the diary process,
>
> > except that, as proven by the PPM, very short incidents of
>
> > listening are missed... People are very accurate in
> putting
> > down what they listened to... about 80% of diary mentions
> > include the exact frequency, so they must know.
>
> Maybe it works that way in your market, but I've seen more
> than one instance here of diaries being delivered to the
> wrong market, wrong county, etc., with the result being
> terribly skewed numbers. What do they do in such cases?

The diaries are not tabulated if they are sent wrong. However, each diary is coded and I have never, ever heard of diaries being sent to the wrong place. In any event, diaries are tabulated based on the ZIP code of the residence. Wherever the ZIP code is, the diary is tabbed int hat market.

Where, specifically, have you seen diaries tabulated in the market where they do not belong? FYI, when a diary is manyally processed, they look for the names and slogans and program names in that market. If non match, the diary goes to a different level of processing, as, perhaps, the diarykeeper moved or was on vacation and they verify by a call back . There are all kinds of double checks.

You have made a bunch of accusations, with no evidence that any of this is true.

 

[davideduardo]

Re: The Times just used the wrong figures...

> >
> > As long as the agencies do accept the ratings, what makes
> > you think your questioning them matters?
> >
>
> Okay, you got me there. As said previously, I think the
> Portable People Meter system is going to go a long ways
> towards minimizing the margin of error which prompted my
> original posting.

It does not change the margin of error at all. That is a function of sample size, and the PPM uses a smaller sample. What it does is provide data faster, and helps to keep the sample truly proportional from week to week, rather than having proportionality wobbles that prevent more granular data, such as monthly ratings vs. 90-day ratings.

> It will be interesting to see how quickly
> the industry accepts it because it also accomodates
> webcasts, Ipods and personal playback devices.

No, it does not. I rates electronic media, not iPods. It rates TV, Cable and any encoded webcsasts and satellite, too. Nothing else. It can even rate storecasting.

> I'll bet we
> see some stations accustomed to big numbers get mightily
> embarrassed when the margin of error is erased.

As I said, the PPM was tested in Philadelphia and the shares and AQH persons levels were nearly identical to that of the diary. In fact, the big finding of Philly is that the diary works. The Houston tests just started, and only one report is out. The shares are virtually the same, and the AQH persons numbers are the same. NO CHANGE.