Are Stations Driving Their Audience Away?

(By Larry Rosin) On November 4, 2022, at a campaign rally four days before the election that could have made her governor of Arizona, Republican candidate Kari Lake addressed a crowd at a rally in Glendale.

“We don’t have any McCain Republicans here, do we?” she said, invoking the name of longtime Arizona Republican Senator John McCain.  “If you are, get the hell out!”

This story came to mind when I was recently discussing, of all things, music test screening criteria.  Increasingly, music radio stations are thinning their music testing targets down tighter and tighter, all but telling large parts of their own potential audiences to ‘get the hell out.’  And I worry that these stations are setting themselves up to suffer the same fate as Ms. Lake.  As most readers will know, Lake narrowly lost her election, likely by the margin of ‘McCain Republicans’ she asked not to vote for her.

While music radio stations are not being anywhere near so aggressive in telling their own listeners who don’t fit the target to get lost, the ever-constricting music research designs may be having functionally the same effect.

Having been involved in radio music research for nearly 35 years, I have witnessed these changes over time.  When music research was first developed, the basic idea of a music test screener was to create a sample that looked like the target audience.  So first came the ‘perceptual study’ or ‘format finder’ where we determined that listening base.  If the core was, for instance, 25-44, that’s what the age range for the screener was.  A twenty-year span was common, because stations wanted to shoot for a lot of listeners.

If the research said the station would appeal 70% to women and 30% to men, the music test was recruited that way as well.  Younger programmers might be stunned to learn that in the 20th century, there were men in the Adult Contemporary music test and women in the Classic Rock tests.

The final screening factor was some combination of P1 and ‘cume-but-not-P1’ listeners, often the same 50/50 that prevails in many cases today.  There were no music montage screeners for established stations (there were for new formats or flips), because the point was to find out what music people who listen like, not to presuppose.

And to be clear, that research worked.  What now may seem like an absurdly broad approach to music test screening led invariably to higher ratings once the results of the music test were installed.

For quite some time now, the ascendant ethos we are encountering for music test screeners is to keep paring away listeners and potential listeners and instead to focus on targets so narrow, that even if they are hit with precision I worry that they don’t represent enough listeners for the client station to win.

Stations are increasingly asking for age targets of just 11 years or so, sometimes fewer.  All women or all men is simply de rigueur.  Even many country stations now target all women in their testing.  By eliminating people outside these narrow age ranges or people not of one sex, every station is leaving tons of P1s to their own station out of their tests – in essence they become each station’s ‘McCain Republicans’ – not told to ‘get out’ but no longer being represented.

In complete violation of the concept that the sample should look like the audience, stations whose audience is mostly one race or ethnicity, will often simply bar listeners not part of that majority.

And now the latest development in the increasing circumscription of the target is playing sample hook montages.  In order to tell us what songs you like, we start out by telling you what songs you must like.  As stated earlier, this makes sense for a format flip situation, when you have defined the target around a sample montage.  And perhaps in very specific situations where for instance you own both the Classic Hits and Classic Rock stations in town, and you want to make sure you have enough ‘true’ fans of each station. But at some point, you are just creating a feedback loop around songs you have already tested.

Most problematically, we are regularly asked to find people who like the target music but who dislike another kind of music.  This one we almost always have to talk the clients out of, mostly because of the absurdity of dinging people for liking something else, since almost everyone likes many kinds of music, but also because at this point the client has attenuated the target beyond the scope of possibility.  We simply can’t fill the test with people in a ten-year age range, who are of only one sex and ethnicity, who are P1 to a three-share station on a platform (FM Radio) that has less listening in the first place, and who both like one kind of music but dislike another.

And the above paragraph speaks to the essential problem – by attempting to zero in on that idealized ultra-core target, you end up setting your station to appeal more and more to fewer and fewer people.  At some point, the target you are aiming at just won’t yield nearly enough people with similar tastes to achieve the ratings your general management demands.  We need to push back against the increasingly held philosophy that the ‘answer’ to higher ratings is to keep narrowing the targets further and further.  The commonly-held assumption that everyone you are screening out will just ‘come along for the ride’ no longer seems to hold.  We are supposed to be in broadcasting, after all.

While even casual followers of American politics know that Kari Lake lost, few are aware that another Republican in Arizona won easily.  In her race for state treasurer, Kimberly Yee ran an inclusive campaign, happy to accept all Republicans who might be voting as well as any Independents or even some Democrats who might be inclined.  She won her race by nearly 300,000 votes.  In both politics and music radio, there comes a point where appealing only to the base becomes counter-productive.

Larry Rosin is President of Edison Research, which he co-founded in 1994. He can be reached by at [email protected]