When I started writing this series of columns, my expectation was that two or perhaps three would suffice. This is number four, the final one in the series, because Nielsen ratings are always a complicated subject.
If you’ve read the previous three columns on the future of radio/audio audience measurement, first, thank you, and second, you’re probably as astounded as I am that when you finish this column, you’ll have read around 4,000 words on the topic.
Last week, I pitched a revised cume metric as the new currency. Bigger numbers, simple to understand, and with the suggested data collection system from two weeks ago, it should also be far less expensive. But programmers will be at a loss when it comes to evaluation, as few look at cume and instead parse AQH to its limit.
It won’t be the same, but think about it this way: minute-by-minute cume gives a very good picture of what’s happening with audience flow. Check my column from April 15 of this year, and you’ll see how to do minute-by-minute meters with PPM in Analysis Tool. Using minute-by-minute cume will be nearly the same. In addition, as streaming data becomes more important, you can overlay it and see how well the two match up.
Will TSL go away? You’ll have to calculate it differently, but a long listener will contribute to cume in more QHs than someone who listens for just a few minutes a week. In other words, what you’re doing now will remain the same under this system.
What I’ve proposed over this four-week sojourn is one option, but think about this: if you’re a Nielsen subscriber, do you feel the cost of the service is a good investment in 2025? If you want Nielsen to charge less—and I’m assuming you want to continue having a third-party ratings service—then the cost of data collection must be reduced. Computers cost far less than people, and the fewer respondents that must be recruited to wear a PPM watch or fill out a paper diary for a week, the lower Nielsen’s costs and, ideally, the lower the fees to you. In this case, I believe we’re looking at a pretty large reduction after the initial investment.
What you’ve read is one scenario. I’m sure others can come up with interesting ideas, but the real question is: “What do we want the radio/audio audience measurement system to look like in 2030? In 2035?” We have a system today that was pretty good for its time and, for better or worse, has served the radio industry well, but it is now outdated.
The industry needs to work with Nielsen and any other potential participants that want to be part of a solution to build a system fit for the digital age. Whether this means an industry group that would be independent, much like the JIC (Joint Industry Committee) started in the TV industry, or one that could operate under the aegis of the NAB, RAB, or other organizations, it needs to happen—and happen soon.
Whenever the word “existential” is used, I’m typically skeptical. When public broadcasting lost its federal funding, I questioned the use of the word based on an average loss of 14% of revenue, per a public broadcasting analyst. Commercial radio revenues were down around 30% on an inflation-adjusted basis from 2019 to 2024, and while a few licenses have been turned in at the FCC, the business is still here.
In this case, we’re close to existential. Radio must continually prove itself to get agency dollars, which have been shrinking. Large market clusters are driven far more by the current ratings system than small-market operators. Yet we’re competing with digital, and buyers don’t seem concerned about bots, fraud, and incredibly low standards for inclusion in audience estimates.
I believe any committee needs some top executives as well as media researchers. I’d also suggest including outsiders with expertise. Many years ago, we were looking for ways to improve response rates for the Arbitron diary service. We conducted a two-day off-site meeting with a group of very smart people from different fields: academia, direct mail, government. The late Ed Shane of Shane Media moderated. Plenty of Arbitron people were in the room, but there was one rule: the Arbitron crew watched and listened but did not participate unless there was a need to clarify or explain methodology, and then only if called upon. We picked up some interesting ideas because we heard from experts with no preconceived notions about how our methodology should work.
In this case, outsiders with expertise in modeling, digital applications, and other fields should be part of the group. That’s going to require a budget, but it’s small potatoes compared to the millions spent each year on an outmoded system.
The time to get started was a couple of years ago, so the sooner the better. If someone has a different idea, let’s hear it, but if we’re still talking about diaries and AQH in 2030, we’ve failed—or maybe worse.
Let’s meet again next week.
