Spotify plays are one more nail in the download’s coffin
We have always been very vocal about the fact that downloads are a poor way to measure podcast success. Bumper’s biggest issue with downloads: they do not represent actual podcast consumption.
A download is not a listen
If you’re new to this conversation, this is what you need to know. Downloads represent file transfers – MP3s copied from a hosting platform to someone’s device. Downloads do not confirm whether anyone actually listened. Many downloads happen without anyone listening. In fact, data collected from hundreds of thousands of podcast episodes through the Bumper Dashboard shows that most downloads are never listened to. I’ll say it again: the majority of downloads do not reach people’s ears.
Despite those shortcomings, the podcast industry continues to use downloads as a key metric. But the tide is turning. We’re noticing a change in the conversation over the last year – more and more people that we meet at conferences or over Zoom, as well as clients that we work with on a daily basis, realize that we need to replace the download as the default way to measure podcasts.
Downloads are declining while audience sizes increase
We have seen downloads decline for many clients for quite some time. At the same time, and confusingly, audience sizes are actually increasing. More people are generating fewer downloads.
This shape of download graph is common for many of our clients. The below are monthly downloads for a podcast we are working on since January 2022:
Downloads have been decreasing for many successful podcasts for a number of reasons, including:
The introduction of iOS 17 in the Fall of 2023. This started an ongoing download decline for most of our clients as the auto-download behaviour in Apple Podcasts was changed.
An increased focus on YouTube as a distribution and discovery platform. Views on YouTube don't yield any downloads.
Adding video episodes to Spotify. Great for discoverability but those videos do not generate downloads.
The IAB updated their podcast measurement guidelines in May 2024, requiring compliant podcast hosting companies to adopt more rigorous filtering of downloads.
At the same time we see corresponding charts that look like this. The below are playback initiations (measured through our play metric in the Bumper Dashboard) for the same shows as above, over the same time period. The number of plays have clearly been increasing:
Again, downloads are going down while plays are increasing.
So why do we focus on the number that keeps going down?
We are not oblivious to the fact that the current business model of podcasting is the biggest driver of a continued focus on the download. Most podcast advertising and sponsorship is sold based on impressions that are triggered by downloads. We fully acknowledge that moving away from this long established system is extremely complex. But the time for a better system is now.
Spotify’s impact on the number one currency in podcasting
Recently Spotify announced they are making play counts publicly available. This will be the first time Spotify is making podcast consumption data available to the public, only publishers were able to see that data up until now. Although it’s still somewhat unclear how exactly Spotify defines “plays” their documentation states that “plays” measure the “number of times a Spotify user actively watched or listened to your content on the Spotify app.”
For Bumper this marks the start of the end of the download era.
As anyone can see the number of plays on Spotify, communicating the success of a show through downloads will not be feasible. All stakeholders in our industry will all look at those play counts and use those numbers to gauge the success or failure of a show. And when the download number associated with Spotify will inevitably not match the play numbers, the reliability of the download data will deservedly be called into question.
Advertisers will inevitably ask to pay for impressions that correlate with the Spotify play number. Buying impressions that don't match the number of plays on Spotify will cause confusion and erode trust.
And yes, download data captures more than Spotify. There are many other podcast players that will generate downloads including Apple Podcasts. But Bumper believes we are only months away from advertisers starting to ask for Apple Podcasts play data as they are evaluating campaigns. (Side note: Apple Podcasts plays are tricky).
The parallels to YouTube in this development are undeniable. YouTube has of course always made their view count available to users of the service and it looks like Spotify is emulating that user experience.
A new currency is needed
The download is still the de-facto currency of podcast monetization. Let’s call it the world's dominant reserve podcast currency. Turns out we picked the wrong currency and it’s time to move on.
We need to agree on a currency that is rooted in the consumption of our shows, and audiences being exposed to the messages from our advertising partners. We need to be able to show our clients that we are charging for ads that have actually been heard/seen by audiences. Spotify introducing a public play metric actually offers an opportunity to better report to our ad partners, and prove that the messages have been heard by audiences.
This means we need to have a broader conversation and collaboration within our industry and agree on what that new reserve currency is going to be.
But what if the new currency is less valuable?
We've been advising clients on better metrics for podcast monetization for the last year. By far the biggest fear is that any new metric is going to yield smaller absolute numbers than the current download. No one wants a smaller number – despite that it might be more valuable or accurate.
The truth is complex and we have seen many versions of this story:
Some Bumper clients indeed see more downloads than our newly introduced Bumper plays. By de-prioritizing downloads, their numbers are effectively going down although a play should be treated more valuable than a download.
Other clients see plays that actually exceed downloads. They see more people hitting play than people downloading their episodes. How? Podcasts that have a big focus on video (whether it's YouTube or Spotify) will see many audience interactions that do not lead to any downloads. A reminder: neither YouTube nor videos on Spotify are generating any downloads.
This is an example of a relatively new show on the Bumper Dashboard. This podcast has seen almost twice as many plays as downloads:
Many shows have an enviable opportunity to increase their numbers, and therefore their available impressions, by switching to a currency that is not derived from downloads but instead rooted in podcast consumption.
Bumper’s approach to podcast measurement
Our recommendation: let’s pay attention to numbers that
measure actual audience interactions with our shows (listening and watching)
keep increasing as more people turn to our medium
We do not know what number should replace the download. But we are working on the concept of a Verified Impression that measures how many times a message was actually exposed to a listener/viewer, regardless of where the ad is placed in the episode. We are able to identify how many times a listener makes it to a sponsor message, whether it’s a preroll or anywhere else in the episode. We are already working with some clients on how a Verified Impressions fits into their business model.
Be part of the conversation
Bumper thinks this is the most important conversation in our industry right now. Next week, Dan Misener and I are at the Podcast Show in London and host an invite-only discussion around the death of the download and the future of podcast measurement. Please let us know if you’re around, we’d love to see you there.
Also, for the rest of the year, we are doing some extensive research with the podcast industry to identify how to best agree on one metric that can replace the download. If you’d like to be part of that conversation, please get in touch.