• 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

KOST Goes All Christmas Friday

DavidEduardo said:
At the end, we would remove such "outliers" as they were not typical and their high scores on what, to everyone else, was a dud, distorted the results. You and Oldies76 and Super and other occasional posters are definitely outliers: radio knows that you can't be pleased and, further, we know that trying to do so puts our core in peril.
Click to expand...

Why in the world would you do that?? It's like throwing away votes in an election. So what if it skews the results. You have to average that in there, to get ACCURATE results. Yeah, the median result will yield a similar outcome you are hoping for. But the outlier will allow you to have that extra song or two that THEY want to hear. By eliminating the outlier, the final outcome is doctored and playing songs the outlier has little interest in and cater to the others nearly 100%. Another piece to the radio puzzle solved.
 
Last edited:
oldies76 said:
You program to researched individuals. 99.9% of the population remains untested and they want different things, so yes, you can't please them all and never will. You will always have the difficulty in pleasing listeners til the end of time. That's a given.
Click to expand...

That's not exactly true. Anyone who listens to music is a "researched individual" in some way. Every song you buy, stream, or search is known. You don't have to participate in a test for us to know. Some researchers use streaming charts that track what people stream on their personal devices. None of the people know they're being researched. But they are. There is a lot of information and data that's available for research.

But the best research is the result. We know that if we play a certain group of songs, it will lead to a certain result. And guess what: It works.
 
oldies76 said:
By eliminating the outlier, you have doctored the final outcome and you are playing songs the outlier has little interest in and cater to the others nearly 100%. Another piece to the radio puzzle solved.
Click to expand...

Are you familiar with the "bell curve?" If not, look it up.
 
oldies76 said:
You program to researched individuals. 99.9% of the population remains untested and they want different things, so yes, you can't please them all and never will. You will always have the difficulty in pleasing listeners til the end of time. That's a given.
Click to expand...

Just look at the polls of last week's mid-term elections. The good ones were off by less than 5% in nearly all cases. And there we were talking about, in some races, samples of a thousand persons in elections with many millions of voters.

Of course, in an election there is only one winner. In radio, many stations can "win" as long as they deliver an audience that advertisers want to reach. So a couple of points, on a scale of 1 to 100, on individual songs is irrelevant and causes no change in the overall results and ranks of songs.

There is a process called a replication study that is done to see what the optimum sample size is for research. "Replication" means, in this context, finding out the minimum sample size that will yield essentially the same results over and over.

One of the ways of doing this is to test far more people than we think are needed to get an accurate sample. Then we see what happens if, at random, we divide the sample in half. Is the result of each half the same? What if we take equal thirds of the sample? Equal quarters? Or at random eliminate 25% or 33% and so on.

We have done this for music tests, and know that a sample in the vicinity of 100 properly recruited station or music type partisans is more than adequate, and doubling, tripling or quadrupling the sample will not change which songs "make it " and which do not as well as which ones are deserving of power rotations and which need less play.

And, generally, the sample size for music tests is larger than the sample for our kind of format in the Arbitron and, later, Nielsen surveys. We use very safe, reliable samples with listeners recruited professionally and verified. That is why a music test can cost $25,000 up to as much as $50,000 per test because they are competently done.

We've been doing this kind of testing in radio for about 35 years now. The research techniques are even older, and broadly accepted in commerce and politics. Yet you criticize sound statistical sampling techniques because you are at a loss to explain why your "outlier" tastes don't mesh with the vast, overwhelming majority of listeners.
 
TheBigA said:
That's not exactly true. Anyone who listens to music is a "researched individual" in some way. Every song you buy, stream, or search is known. You don't have to participate in a test for us to know. Some researchers use streaming charts that track what people stream on their personal devices. None of the people know they're being researched. But they are. There is a lot of information and data that's available for research.

But the best research is the result. We know that if we play a certain group of songs, it will lead to a certain result. And guess what: It works.
Click to expand...


So, basically you are saying that if enough people playing / streaming xxx song and tracked by these research methods, that song could end up in rotation, solely based on those results?
 
oldies76 said:
Why in the world would you do that?? It's like throwing away votes in an election. So what if it skews the results. You have to average that in there, to get ACCURATE results. Yeah, the median result will yield a similar outcome you are hoping for. But the outlier will allow you to have that extra song or two that THEY want to hear. By eliminating the outlier, the final outcome is doctored and playing songs the outlier has little interest in and cater to the others nearly 100%. Another piece to the radio puzzle solved.
Click to expand...

Radio wants to reach a mass audience. When perhaps 0.5% of listeners want something quite different, we go with the 99.5% because we know that the folks in that tiny group have tastes that are not the same as the infinitely larger group.

We know why we "cleanse" the data to remove extreme outliers. It is to avoid having a tiny group distort the big picture. I've also seen cases where, in a test, we noted a person with odd behavior... indicative of drug or alcohol, perhaps. We check that person's individual scores, and if it seems like they were scoring at random, just so they would get their $100 to $200 payment, we delete them, too.

We do not want to have that "extra song or two". That's because such songs, only liked by a small group, will kill the station for the big majority group.

We have a choice: play to the outlier and get a 0.2 share or play to the majority and get a 6 share. It's an easy decision.
 
oldies76 said:
So, basically you are saying that if enough people playing / streaming xxx song and tracked by these research methods, that song could end up in rotation, solely based on those results?
Click to expand...

No, there are other factors involved. The first is whether the on-demand streamers are in our market, and the song is right for our format.

On demand does not give us precise information on the demographics and radio usage of streamers, but when we evaluate music we look at all kinds of factors including streaming to make sound decisions.

Of course, streaming stats are really only useful for current music, and not so much for stuff more than a year or two old. But the date is very valuable to CHR, Urban, Latin Hits, Country and other formats that are predominantly current based.
 
TheBigA said:
Are you familiar with the "bell curve?" If not, look it up.
Click to expand...

From google...

"Bell curve system of performance appraisal is a forced ranking system imposed on the employees by the management. Through this system the organization tries to segregate the best, mediocre and worst performers and nurture the best and discard the worst."

So essentially you discard the radio outliers. I agree with David's reason that the ones out to get the 200 bucks should be nullified, but the outliers who sincerely choose their favorites, even if they give high approvals to all of them, should not be. But there is probably not a system in place to determine which outliers are truthful and the ones out to get the bucks.
 
oldies76 said:
But there is probably not a system in place to determine which outliers are truthful and the ones out to get the bucks.
Click to expand...

We've explained this to you a million times. We're aiming for the center of the curve. That's where the mass is. The outliers are not part of the mass.
 
David and Big A,

While it may seem differently from the tone of my posts, I actually rarely disagree with you both and am fascinated by the things you know that I have been glad to learn over the years regarding an industry I have nothing to do with have been intrigued with over the years. I present the position of, as you correctly point out, an outlier music lover that radio cannot possibly serve and generally gripe about the way things are, but I definitely understand that radio is a medium for the masses and a business that has to be run to satisfy them.

The arguments you have made this evening are not new in that they focus on the same themes you have spoken of over the years, but you have both made your points very eloquently this evening. Bravo. I especially like Big A's theme of free terrestrial radio being a free sample for listeners to buy later. That makes a lot of sense. David's discussion of testing methods and results are also very illuminating. In any case, since it always seems like I am taking an "opposing view" I just wanted to express my thanks that you both take the time to continue to illuminate blowhards like myself.

Thanks guys!
 
I have learned an extensive amount of information as well. I know that I tend to challenge some points presented, but overall the information presented over time has taught me a great deal, especially with points I can agree with. Channel Flipper's summary says it well. And thanks again.
 
No problem. Having said all that, as you know, there ARE radio stations in LA that target your taste in music.

You know where they are, so just be happy with what you've got. No need to change the world.
 
oldies76 said:
From google...

"Bell curve system of performance appraisal is a forced ranking system imposed on the employees by the management. Through this system the organization tries to segregate the best, mediocre and worst performers and nurture the best and discard the worst."
Click to expand...

The bell curve, used in applied statistics, is demonstrated when you look at naturally occurring things like IQ. That's where we find things like "68% of the population falls between 85 and 115 IQ, with those from 70 to 85 and 115 to 130 being each, 14%" and so on... a declining and matching curve on either side that flattens out like the rim of a bell.

In a music test, we see similar phenomena where 80% and better of the scores are within a narrow range, with highly divergent scores from the normal range, if repeated by the same person over many, many songs indicating an outlier... one whose tastes differ radically from the norm nearly every time. This person is of no use in determining mass appeal and is thus an outlier and eliminated.

So essentially you discard the radio outliers.
Click to expand...

Yes, just as we don't let dragsters into the Monte Carlo Grand Prix. They are out of place and don't "fit". In fact, "fit" is often discovered when we analyze the range of divergent scores on each song. It tells us if maybe we have a good song, but it belongs on a different radio station, not ours.

I agree with David's reason that the ones out to get the 200 bucks should be nullified, but the outliers who sincerely choose their favorites, even if they give high approvals to all of them, should not be. But there is probably not a system in place to determine which outliers are truthful and the ones out to get the bucks.
Click to expand...

After doing a thousand or so music tests and spending nearly $20 million of one company's money on research, I can say that there are both statistical and observational ways of telling outliers. Since they will actually do more harm to the station than the money moochers, we are very conscious of the need to keep their results out of the tests.

The good thing is that we only find one perhaps in every 200 to 300 participants, and if they were to get through, their effect on the total sample is around 1% when one is present... well within the total project margin of error.

The purpose of spending so much money on testing is not to accurately measure taste. It is to accurately measure the tastes of our listeners and/or potential listeners who will also participate in radio audience surveys and who will spend the most time listening. For example, we also don't test people who might listen 1 hour a week...

We're mining for gold. If, in the process, we find trace amounts of tin, we ignore it. It's not what we are after, and it's not economically viable to mine trace amounts.
 
Last edited:
TheBigA said:
No problem. Having said all that, as you know, there ARE radio stations in LA that target your taste in music.

You know where they are, so just be happy with what you've got. No need to change the world.
Click to expand...

Mr '76 is somewhere in the Pueblo / Colorado Springs area.
 
TheBigA said:
No problem. Having said all that, as you know, there ARE radio stations in LA that target your taste in music.

You know where they are, so just be happy with what you've got. No need to change the world.
Click to expand...

You are correct, Sir!!! It's been a long time coming for me, but I find myself almost exclusively listening to ONE radio station for music (88.5 FM KCSN/KSBR). I do listen occasionally to an out of town oldies station (WDJO Cincinnati). Beyond that I have little interest in any other format or commercial radio station that plays music. At the office I use an old cell phone that I retired 3 years ago BUT still serviceable to play the Tune In Radio app and it is connected to an old Logitech boombox. On the app itself is a laundry list of maybe 50 stations I have sampled or listened to over the years, a lot of the stations on this list are out of town OTA commercial stations playing the same format as 88.5 FM. BUT for over a year now, I find almost all my music interest needs are met by this ONE radio station and find myself rarely sampling the others very much anymore. AND of course I want them to continue so I donate twice a year $100 to help keep them going. Since I spend my money this way, I don't buy CDs because my new car, as many do now, has eliminated a CD player. BTW one of the COLs is Mission Viejo where I live so it feels like it's a home town station I am supporting.
 
You must log in or register to reply here.
Status
This thread has been closed due to inactivity. You can create a new thread to discuss this topic.
Back
Top Bottom