This weekend there was a nice interview in The Observer Magazine with Chris and Rosie Ramsey, who make one of the UK’s most popular podcasts Sh**ged, Married, Annoyed. It’s undoubtedly very popular, and the interview notes how they rapidly sold out venues like Wembley Arena and the O2. The couple have a new BBC TV show coming soon, and Chris is featuring on the latest excellent series of Taskmaster. These guys are definitely in the ascendancy.

But the piece opens with the claim that they have “more than 100m listeners.”

Hmm.

Further on, the piece notes: “Launched in 2019, [the podcast] immediately zoomed into the iTunes Top 10, where it stayed: it’s the only podcast that has ever stayed in the Top 10 for a whole year. As of last month, it has 100m listeners.”

To be clear, I don’t doubt the podcast’s massive success.

The line about its pervasiveness in the iTunes Top 10 supports that because the iTunes Top 10 is not like the Top 10 movies, or Top 10 albums. Apple uses an algorithm that isn’t limited to overall listens, but also favours new listeners. In other words, a popular podcast with a large number of listeners, but few new listeners jumping on board, is likely to be overtaken by a slightly less popular podcast that is seeing significant growth. So if you stay in the iTunes chart, it suggests big numbers and continued growth.

But what I really wanted to dig into here is the “100m listeners” line, because I’m pretty sure that’s wrong.

My suspicion is that the number is actually 100m listens.

Different platforms name this slightly differently, but it can be referred to as downloads or listens in some podcast analytics platforms. It does include streams, and it’s notable that Apple calls it plays in the Apple Podcasts Connect platform, while Spotify calls it streams (they also have a separate metric, starts, for streams that begin but don’t make it to 60 seconds play-time).

Downloads, listens, plays and streams are all basically the same thing, and I’ll mostly use downloads here just to avoid confusion between listens and listeners.

Why is it not 100m listeners?

Well, let’s try something called the Fermi estimation which writer David Epstein talks about here. Essentially, you’re trying to either work out a number, or check whether a number could be true, when you have no real prior knowledge.

In the example in the linked piece, people were asked to estimate the number of piano tuners in New York City – those people having no prior knowledge of either how many people own pianos, or of the New York piano-tuning industry. By making some sensible estimates, you can get in the ballpark with your guess.

In the UK, the same strategy is fun to try in the Distinctly Average round of Richard Osman’s House of Cards where contestants are asked to come up with estimates of generally unknowable numbers like how many cups of tea are drunk daily in the UK, or how much does the average British person spend on beer in pubs in their lifetime.

But let’s use a Fermi estimation to get a feeling for whether the Ramseys’ 100m listeners could actually be right.

We’re pretty sure Chris and Rosie Ramsey are very popular – they’ve sold out massive arenas after all, as well as having had book deals and their now upcoming TV show. But the population of the UK is only 67m, and that includes children. And only 25% of the UK (adult) population listen to podcasts every week. Let’s be generous and say that 20m people in the UK have heard their podcast.

We still need to find another 80m to get to that 100m listeners number. But podcasts are global, and I suspect the Ramseys, like other British comedians, are popular in Australia too (population 27m). Another generous 10m listeners there.

What about the US? I’m sure nobody there would find a Geordie accent a challenge (I’m not being rude – my dad’s a Geordie!). A decent proportion of the 330m US population would get us over the line…

But I’m not sure that really holds up. If they were that big in the US, their new TV show would be on Netflix or a US broadcaster.

And anyway, let’s look at the person who most think probably is the biggest podcast (and podcast comic) in the world – Joe Rogan.

Spotify doesn’t publish stats, but this Verge piece reports that Rogan himself was claiming 200m downloads a month back in 2019, prior to going exclusively to Spotify. While podcasting itself has continued to grow over the last couple of years, it seems likely that going exclusively on a single platform curtailed his numbers a bit, and indeed, the tenet of the Verge piece was about his waning influence since he went Spotify-exclusive.

But even if we just consider the 200m number, and take account of the 18-20 episodes a month he was doing at the time, that represents, say 11-12m downloads an episode.

Over a month, not everyone listens to every episode, so while 11-12m people heard one episode, perhaps 20m heard any episode. In other words, Rogan may have had roughly 20m listeners in a given month back in 2019.

That’s for the biggest podcast in the world.

Now if there is one thing to take away from this entire piece, it’s this:

DOWNLOADS ≠ LISTENERS

(That’s the not equals to sign in the middle!)

But for a single episode of a podcast. downloads are a reasonable proxy for listeners.

However, we do have to be a bit careful because any of the following could be true of a single download:

• I downloaded a podcast episode but didn’t listen to it. One download, zero listens.
• I downloaded a podcast episode and forced my family to listen to it on a long drive. One download, five listens.
• I downloaded a podcast episode and inflicted it on the entire workforce of my factory where I’ve also banned personal headphones. One download, 1,000 listens.

Any or all of those could be true of a single download. But research suggests most people listen to podcasts alone, so counting one download (or listen or stream or play) as one person listening seems reasonable given the lack of other available evidence.

However, when you add up downloads – or listens or streams or plays – over time, you are not necessarily increasing the number of listeners!

Think of my fictional niche podcast, Adam Pontificates. It’s so niche, I have an audience of just ten people who listen religiously. Nobody else ever downloads or listens to it – just those ten. I produce one episode a week for a full year, and my fanbase never misses an episode!

52 episodes x 10 listeners = 520 downloads (or listens)

But there are still only 10 actual listeners.

If I ratcheted up my production schedule and made two episodes a week, and my dedicated loyalists stayed with me, I could get to more than 1,000 downloads!

2 episodes a week x 52 weeks x 10 listeners = 1040 downloads (or listens)

But still only 10, very long-suffering, listeners.

In real life, podcasts grow and decline over time at different rates, and not everyone listens to every episode. So the maths is never quite this clean cut.

But it is important to differentiate between downloads (or listens) and listeners. They are different, and if you’re going to use either number, then you should really accompany it with a period of time. Is that a weekly number, a monthly number or an all-time number?

Going back to The Observer Magazine piece, I think the important things here are that:

• It probably meant listens and not listeners
• It should really have included a time-scale. Is it a per-week, per-month, per-year or all-time figure?

The number of episodes you publish in a given timeframe also matters. The number of downloads for a daily podcast will be massively bigger than the number for a similarly popular weekly podcast. That daily podcast will have more inventory for selling ads in, but both podcasts might reach roughly the same number of people.

Everyone loves a big number, and the internet is full of what the Entertainment Strategy Guy calls “datecdotes” – those numbers which sound big and impressive, but don’t actually give you anything too helpful, and can’t easily be compared with any other numbers.

I’m not expecting anyone to become a stats expert. But if you hear a big number, just stop and think. Does it pass the sniff test? Do you actually believe it?