Ofcom has just published its latest Audio and Podcasts survey data for 2023. They tend to just publish the raw data tables, so I like to dig into the data and get into what the data actually means. I’ve done this for the last couple of years, and you can expect many charts and graphs.

(2022 analysis here | 2021 analysis here)

Ofcom has changed things up a little this year, and published the Podcast and Audio surveys together. As a consequence, there are some small changes to both the surveys to avoid duplication.

Note that I’m only looking at the Podcast survey in this blog. I will dig into the Audio survey separately, which gets much more into ways of listening to the radio, including sections on some key apps.

The Podcast survey involved using research agency Yonder’s online panel between 3rd and 12th March 2023, and used a 1,884 nationally representative sample of which 518 were non podcast users, 1,006 were regular podcast users and 360 were occasional podcast users. A boost was used to ensure that there were at least 1,000 regular users, to provide robustness when asking questions of regular users. The full technical report of the sampling is on Ofcom’s site.

The full data tables prepared by Yonder are available on Ofcom’s website. Note that they do take the form an Excel spreadsheet with 100,000+ lines of tables. You have been warned! Nonetheless, some will probably want to get into the weeds of this data.

One other reminder is that different surveys have subtly different methodologies and that means you will get different numbers. So be cautious when comparing a number from this survey with numbers from other surveys that almost certainly were compiled in different ways.

This first chart is a little crowded, but it details overall listening patterns. If any sections are unclear, please hover over the element in the chart and a pop-up should help clarify things.

This simplified version of the chart is perhaps easier to see and puts the respective audio formats into perspective. Note that “Online Music Services” means things like Spotify, while “Music Video Websites” largely means using YouTube videos for background listening. “Social Audio Services” means things likes Clubhouse or Discord. “Other Audio” could include Apple Fitness+ or various wellness audio services.

Specifically looking at podcasts, the data shows that 25.5% of people listen to podcasts weekly and 36.4% listen monthly.

50% of people never listen to podcasts.

By way of comparison, last year Ofcom reported 25% of people listening weekly. The 2023 number is probably within the margin of error, so we can call the percentage unchanged.

Looking at sex, gender and socioeconomic charts show to what extent each group listens “regularly” to podcasts. In this instance “regular” means “weekly”.

A few things to note here:

• A surprisingly heavy skew towards men amongst regular (weekly) podcast listeners
• Younger people are more likely to listen to podcasts, but not the very youngest group (note that this survey did not measure under 18s)
• Podcast listening skews towards ABC1s

A perhaps more useful way to look at this listening is to look at the breakdown of regular weekly podcast listeners.

These pie charts show overall regular weekly podcast listeners broken down by gender and age.

To give a point of comparison, here is that same data for monthly podcast listening.

The key thing to note here is that the gender gap closes substantially, now 56% male and 44% female. By way of comparison, Edison Research in the US reported in 2022 that 53% of US monthly podcast listeners were male and 46% were female. The recently published 2023 data does not include a gender slide in the freely published report. Again, note that Edison Research will be using some different methodologies so this comparison should only be made broadly. (Note also that Edison did report 1% as non-binary/other. Ofcom’s survey did ask allow respondents to answer non-binary or other, but the responses are not tabulated. The data suggests 0.8% of respondents answered in a such a manner.)

Something perhaps not widely reported is the increased likelihood of those of non-white ethnicity to be podcast listeners.

What this chart shows is that white listeners are basically “average” podcast listeners, with 24% listening weekly, on a par 25% of the population overall.

But this rises sharply to an overall 38% for any ethnic minority.

The figure for black respondents is especially high at 49% of respondents saying that they listen to podcasts weekly. However, I would note that the base for both black and mixed ethnicities is considered very small. Indeed, the research company only considers the net ethnic minority group to have a significant large enough sample not to warrant noting.

In other words, use the data cautiously. The 38% figure can be safely used, but be wary of digging into specific ethnicities. It’s a shame, as I think there’s a really interesting story in here that needs better data to substantiate.

Ofcom does ask respondents who’ve listened to podcasts in the past, why they no longer listen.

I would suggest that many of these problems are issues of discovery rather than a failure to provide podcasts that meet the needs of these consumers. Time is likely to be much more of an issue however!

Speaking of discovery, Ofcom asks respondents how they tend to hear about new podcasts.

Word of mouth and social media remain key to discovery, but the podcast apps themselves are major drivers as is radio. Perhaps most concerning here is the 30% of people who claim never to hear about new podcasts…

But there are some notable differences in how important different discovery mechanisms are dependent on your age.

Perhaps unsurprisingly, social media is much more important to 18-34s. But word of mouth is more important to this group too.

And it’s this age data that reveals that it’s over 55s who are simply not learning about new podcasts.

Respondents were asked when they listened to podcasts. Note that they could obviously choose multiple categories here.

One of my favourite questions is how many individual podcasts do respondents listen to in a week.

This year, the mean number has gone up from 5.1 in 2022 to 5.4 in 2023 (rounding up from 5.35).

It can be useful to examine the spread.

I suspect that many reading the analysis will be in one of the two right-hand groups, listening to more than 11 podcasts a week. But we’re not normal!

I often remind people of this number because it shows how hard breaking into someone’s media habits can really be.

The good news here is that it is a number that is edging upwards, but it could be growing faster.

I’ve not plotted the data here (too many charts!), but younger podcast listeners are more likely to hear fewer titles (4.11 average for 18-24s), whilst the age group with the highest mean number of podcasts listened to is the 45-54 year old age group with 5.99 average podcasts listened to.

There are no noticeable differences between the number of podcasts that men and women listen to, with women (5.4) slightly ahead of men (5.33), but within the margin of error.

We get a similar story with the number of podcast series subscribed to. The mean is just 6.4 podcast series (a number that is down from 7.1 last year).

Perhaps the most significant number here is the 10% of regular podcast listeners who do not subscribe to any podcasts at all. Looking at the data, these individuals are much more likely to be older – 22% of 55+ respondents say that they don’t subscribe to any titles.

Podcast genres are always interesting to examine, and Ofcom’s research asks about them.

Comedy takes the top spot, overtaking Entertainment compared with last year. But the overall picture is pretty consistent with previous surveys. For clarity, the three bars represent a podcast listener who ever hears a title of that format, listens monthly to podcasts in that genre or weekly. Invariably more people ever listen than listen monthly, and more listen monthly than weekly.

But that chart does show that people who listen to news podcasts are more likely to have an increased frequency of listening compared to other genres.

Note that the full datasets do get into the profiles of every single one of these genres, so the full data tables are useful for those who want a deepdive into, say, the profile of business podcast listeners.

The research captures podcast apps and websites that people use. This is quite a complex picture and Ofcom is quite comprehensive in trying to measure different apps and even different versions of apps. The only thing I would note is that some of the sample sizes are necessarily small, and sometimes you might note that a particular platform doesn’t actually have that many podcasts on it. Nevertheless, the numbers are measured.

Respondents are also asked which they use most often.

My main takeaway from this is that Apple doesn’t do quite as well as other surveys would suggest.

Overall, Ofcom research notes that regular podcast listeners use an average of 3.04 different platforms to listen to podcasts.

Respondents were asked to rate the importance of the different factors when thinking about the ways in which they access podcasts. The chart below shows those who consider those factors important or very important.

As ever, individual apps are rated on these measures, which if anyone is interested, they can dive into the data tables to look at themselves!

Finally, the research asks about some specific podcast formats, and asks about frequency of consumption.

I hope this is useful. Questions, comments and thoughts are welcome!

All data collated by Yonder on behalf of Ofcom. Data collected in March 2023.

Disclaimer: These are my views alone and do not represent those of anyone else, including my employer. Any errors or omissions are mine. Please let me know if you find any mistakes (I hope there aren’t any!)

The “Podcast Survey” image above generated using Adobe’s Firefly Text Effects AI software.