Can You Work Out Your Listening Figures From Twitter?
June 14, 2016
There's no doubt that radio audience figures are pretty difficult to work out.
In some parts of the world, radio audience figures are estimated by giving a percentage of people a little portable listening spy device, which listens to everything and attempts to spot a secret hidden code in the audio. These things are called PPMs, Portable People Meters, and are in use in markets in the US and Canada as well as in countries like Denmark and Norway.
In other places, researchers knock on doors and ask people to fill in a diary - either a paper diary or online - detailing their listening habits. Those people, carefully chosen to reflect the country's demographic, provide numbers from which you can estimate total audiences.
In other places still, people use the telephone to call people and ask them what radio station they listened to yesterday. Again, these numbers are used as a base to estimate total figures.
All of the above research costs money. Quite a lot of it. So, it's only natural that small radio stations are quite keen to work out a different way to determine how many listeners they have. Preferably, a free way.
The question is... is it possible that social media plays an answer?
I scoffed at this as an idea when someone suggested it. Could the total followers to a radio station Twitter account really be matched to an audience figure? Can you translate Likes to Listeners? I was a little dubious. No: I was massively, 100%, adamant that these figures measure different things. It's a fool's game to try to equate them.
That said, I have a website with - for the UK, at least - total listening figures (in 'cume')... and social media figures. So, against my better judgement, I compared cume with "followers" on Twitter, on Facebook and on TuneIn.
The result for Facebook looks, charitably, like someone has sneezed over the monitor while exclusively eating blue M&Ms. Even so, there's some math stuff that means that it has a 0.41 Pearson's Correlation Coefficient. That is apparently supposed to mean a "strong" positive relationship between Facebook likes and total listeners. And my experiment appears to show that if you have 2,000 Facebook 'likes," you have about 10,886 listeners. It's math. I don't believe it, but it's math.
The Twitter figures return a Coefficient of 0.53, so the relationship is a bit stronger for Twitter followers. 2,000 Twitter followers means 11,948 listeners, so the math tells me.
Frankly, I don't believe either of the above.
However, the TuneIn figures actually produce a graph that I believe. You can see there's a connection between total listeners and total "followers" on TuneIn - "followers" is TuneIn-speak for "presets," because, hey, why make it easy for listeners by using 'presets' when you can use 'followers' instead? The Coefficient figure is 0.78, which apparently means the relationship is a very strong one. So, math says 2,000 TuneIn followers means a cume of 11,010.
As I'm at pains to point out: this is probably a silly thing to do. Social media has little to do with actual listener figures. But I could be persuaded that if you have a preset on a radio app, like TuneIn, that is comparable to audience figures in some way. And it is at least possible that people would stop 'following' a station that they no longer listen to.
Naturally, there are plenty of flaws. Some stations aren't on TuneIn; others who promote TuneIn almost exclusively on-air as a listening option will be over-represented. I'm not seriously suggesting selling advertising on these figures. They're, as a late night UK radio presenter would call it, "bullshine."
But, anyway, science.
Next week, I'll be rifling through your kitchen bins to discover whether the total amount of coffee you drink is related to the amount of Justin Bieber on your playlist.