I began developing audience evaluation methodologies during my time working simultaneously at 98 FM Natal and 89 FM Natal. As early as 2010, I was already identifying significant inconsistencies between Brazil’s leading research institutes and their methodologies, especially when compared to approaches in other countries, where there is a stronger emphasis on accuracy and aligning results with reality.

It is no coincidence that even major institutes in Brazil have missed electoral predictions by wide margins. This only reinforces the idea that, even with genuine intent to deliver accurate results, their methodologies are fundamentally flawed.

Listeners online were exact numbers that radio stations had, while institutes came up with completely different ones

One of the most striking examples I encountered involved an audience figure I knew with absolute certainty, yet the institute’s published data was drastically different, off by an astonishing magnitude. This occurred when they reported the online audience for a given radio station during a specific time slot, while I had direct, real-time access to the actual audience numbers from that station’s Shoutcast panel. The same pattern emerged whenever I checked other stations’ Shoutcast data and compared it to the institute’s published figures, the discrepancies were undeniable.

I am not suggesting deliberate falsification, but for some reason their numbers were severely inaccurate. This is particularly concerning because obtaining these figures requires nothing more than extracting them directly from the streaming server, providing an exact, verifiable count, rather than relying solely on sample-based interview surveys, which inherently introduce margins of error.

Methodology

Radio audience research data in Brazil is measured in a declarative way (people respond where they say they listen to radio the most), and this can create a mismatch with the perception we have in everyday life for several reasons:

Data collection methodology – Surveys are conducted via telephone or in-person interviews, and the question usually allows for multiple answers (home, car, work, etc.). Many people select “at home” even if it’s just for a few minutes in the morning, and that already counts toward the statistics.

Listener perception – For many, having the radio on while doing something else at home (cooking, cleaning, having coffee) is “listening to radio,” even if attention is minimal. This increases the weight of the “home” category in responses.

Heavy in-car listening – Although the car is indeed an ideal ecosystem for radio (little direct competition, built-in audio, driving habits), it usually ranks as the second most common listening location because not everyone drives daily, but almost everyone spends time at home.

Difference between “passive” and “active” listening – At home, radio often plays in the background. In the car, attention is usually greater because it’s the main companion during travel.

Interestingly, if we measured the average continuous listening time, the car could surpass home in many cities, but traditional surveys measure reach (how many people listen in each location), not intensity.

International overview: the car is king for radio listening

United States

A Nielsen report shows that 65% of all AM/FM radio listening happens out of home, and 44% occurs in the car.

Another study, Edison Research Share of Ear, reveals that while on-demand audio (like streaming services and podcasts) has already surpassed linear formats in other environments, in the car approximately 76% of listening time is to linear radio (AM/FM, SiriusXM, etc.).

Another interesting Edison finding is that 74–75% of daily in-car audio time is dedicated to radio, regardless of the vehicle model.

Empire State: consistent behavior

Nielsen also confirms that AM/FM radio dominates the in-car environment, accounting for 85% of ad-supported audio listening, back to pre-pandemic levels, while other environments see a greater presence of streaming.

A 2024 study highlights that 86% of ad-supported in-car audio time is dedicated to radio, and 62% of drivers say they would rule out buying a car that lacked AM/FM capability.

Other key markets

In Futuresource’s international report, the U.S. shows that over 40% of radio listening occurs in the car, while the U.K. records 24%. The report also indicates that globally, 30–40% of radio listening happens in the car, with an even higher share in the U.S.

In Australia, the Infinite Dial study reports that 88% of respondents listen to radio while driving, confirming radio’s dominance in the car.

Why your environment is so revealing

Studies show that the car is the environment with the highest concentration of attentive, engaged radio listening, especially in countries like the U.S. and Australia, fully aligning with our practical perception.

Home tends to appear more prominently in declarative surveys because people consider listening to the radio as part of their daily routine, even if it’s passive or fragmented.

In the car, radio faces little to no direct competition (such as television or the internet) and is usually the main, sometimes the only, source of information and entertainment, especially during commutes.

How can one not be surprised by such mistakes?

I make use of telemetry whenever possible, including for radio. Streaming is the main tool for assessing listener behavior, even if it doesn’t indicate the absolute audience size, and it has been evolving into an important sample of the overall profile of what listeners value with their attention, as well as what ends up driving them away from a station.

What surprises me is that telemetry is readily available and easy to implement. You can check where listeners are connecting from, how long each of them listens to the station, whether they stop listening when there’s a change in program or host, or if a commercial break runs too long. This level of precision is remarkable.

What shocked me most is that audience research institutes in Brazil have no real notion of how many listeners stations have online, that number is practically a mystery. A station can check in real time exactly how many listeners it has and since when they’ve been tuned in. It’s hard to trust radio audience metrics in Brazil after realizing just how badly wrong they are.