Podcast Metrics: How to Measure Your Performance (With Such Little Data)

Podcasting has been around for awhile (since the 1980s, to be exact), but in the last few years, the level of interest amongst listeners has grown dramatically.

Today, approximately 67 million Americans over the age of 12 listen to podcasts monthly.

As the audience has grown, so has the number of podcasts and the interest from advertisers considering spending some of their ad dollars on this emerging medium.

Whether you are a podcaster or a brand considering sponsoring or advertising on a podcast, measurement of a podcast’s performance and audience is key to improving and assessing return on investment (ROI).

Unfortunately for marketers (like me!) who are used to relying on robust analytics platforms to deliver consistent key performance indicators (KPIs) for things like website visitor traffic, pay-per-click advertising, and social media engagement, tracking and measuring podcast performance remains challenging.

Because podcasts are downloaded to user devices for online listening or future consumption, it is impossible to measure how they and the ads they contain are consumed.

For this reason, there has been a conspicuous lack of consistency in how podcast performance is measured.

Luckily, that has all begun to change with the release of the Interactive Advertising Bureau’s (IAB) Podcast Measurement Technical Guidelines, updated in December 2017.

Why is Podcast Performance So Hard to Track?

If you are a podcast listener, odds are you use an app like Apple’s Podcasts (available on iOS devices), iTunes, Stitcher, or Google Play to download, store, and play your preferred podcasts.

You might even be listening on your Alexa or Google Home.

Regardless of the platform you use, however, you must first download the podcast.

This download is what makes podcasting (and measuring podcast performance) fundamentally different than other forms of digital media.

Bear with me here, I’m going to get a little technical...

As defined by IAB’s measurement guidelines:

“Podcast content is an on-demand media format that listeners either download to, listen to later, or consume online. Unlike the streaming format more common in video, podcasts continue to be downloaded because of the convenience offered by existing platform and application functionality.

Despite the use of the word ‘streaming’ in podcasting, ‘streamed’ podcast files are progressively downloaded via the standard HTTP protocol.”

In plain English this means that while it might seem like you are streaming a podcast, you are, in reality, listening to a file that has been downloaded from a host server.

Once that file is downloaded, the host itself has no way of knowing how much of the podcast you’ve actually listened to - or whether you listened to it at all!

To clarify, a stream occurs when digital media is delivered via the hosting server at the same time that it is being consumed; the file is not downloaded to the local device. Examples of popular streaming services include YouTube, Netflix, or even Facebook.

Podcast apps, on the other hand, download the media file (your podcast) onto your local device before playing them. The program used to access your media file is called a “user agent.” User agents can include web browsers, iPhone apps, and desktop apps such as iTunes, etc.

So, long story short, there really is no such thing as a stream in podcasting.

But Podcast Downloads Are Also Unique.

As mentioned earlier, podcasts are transferred to your local device via a “progressive” download which makes it difficult to establish clear KPIs.

Unlike a traditional download (for example, downloading a PDF from a website to your computer), in which the file must be fully downloaded before it can be accessed, a progressive download allows playback of a media file to begin before the download of the full file is complete.

The download can actually continue while the file is playing (a.k.a. while you’re listening to your podcast).

This is not only different from a traditional download, it is also different from a stream.

Whereas a stream plays the media directly from the host (think YouTube videos), a progressive download plays the media from the user’s local device.

When you listen to a podcast, the audio file downloads from the provider’s server to a temporary file on your device (think computer, tablet, or phone) in response to your device sending multiple requests to the provider’s server to get a little more of the file each time.

The tracking challenge with progressive downloading is that a download does not necessarily equal a listen - or a listener; it merely means someone’s loaded it onto their device.

A listener can also download a file and not listen to it, or even download the same file multiple times (for example, to both their phone and their iPad for an individual listener).

A downloaded file can also be listened to multiple times.

An application (such as an HTML5 media player or podcast app) used by a listener may request the media from your hosting server several times to fully download the content during playback, or a listener may manually re-request download several times in a short timeframe.

All of these variables and inconsistencies make it difficult to set and assess KPIs for podcasts.

Setting Key Performance Indicators for Podcasts

Given the unique complexities of the ways that podcasts are consumed and downloaded, what is the best way to track and measure their performance?

The good news is there ARE some podcast metrics that can be consistently tracked and provide valuable insights into performance, audience growth, and engagement.

Unique Downloads

Earlier, I explained how podcasts are downloaded, and the difference between traditional downloads and progressive downloads.

Because of these, the number of times a podcast has been downloaded is not a perfect measure of audience size -- but it is the closest approximation and a KPI worth tracking.

The IAB recommends using “unique downloads” to measure the size of the listening audience.

According to its guidelines, a unique download represents a media file that is progressively downloaded (not streamed) using a specific user agent from a specific IP address, all within a 24- hour window.  

In other words, a unique download is a filtered statistic in which multiple individual download requests are attributed to the same overall download request. So, rather than tallying up the raw requests online from a user, those requests are shown as one unique download.

While there is definitely room for error (multiple people listening to a single stream, or a single person downloading the podcast in two different places such as home and work), the number of unique downloads is probably the closest to accurate of all podcast measurement metrics because it most closely resembles your actual podcast audience size.

The IAB also provides for how much of the file (the ID3 tag plus enough of the content to play for one minute) should be downloaded to count as a unique download for reporting purposes.

If the system you use to track your podcast’s performance cannot measure partial downloads, then the IAB guidelines require that the file is downloaded in full to actually count as a unique download.

Subscribers

Just about every podcaster I know has on their wish list the ability to track their number of subscribers.

Unfortunately, this is not currently possible as podcast applications are managed via cloud-based services. This gives the appearance of one subscriber even though there are many more.

It is also nearly impossible to get data on how long people are actually listening to individual podcast episodes because this information is stored on the “client side” (i.e. the server at iTunes, Apple Podcasts, etc.) and most of these user agents don’t report that information to podcast hosting platforms.

So, what do you do?

Look For Consistency

In the absence of accurate subscriber data, the best way to understand the size of your subscriber base is to look at the degree of consistency in download totals between each episode.

The larger the number of consistent downloads, the larger your subscriber base likely is.

More importantly, look at the first 48 to 72 hours after you publish because this time is typically dominated by subscribers.

Now, this method is an approximation and can’t be used to provide a completely accurate number of subscribers because some may deviate from their consumption habits (ex. They might binge listen to recent episodes on a certain day of the week).

Don’t Neglect Context

While it is nice to be able to determine your overall number of podcast subscribers, you have to put these numbers into context. The more niche your podcast, the less relevant these numbers are.

For example, if your podcast is about politics in your hometown of 10,000 people, the potential audience for that is going to be fairly limited and getting 500 subscribers - or 5% of your total potential audience, could be considered a real home run and might garner you the attention of local businesses looking to advertise.

On the other hand, if your podcast is on cooking and you only have 500 subscribers, advertisers are likely to be far less interested because they have so many other podcasts in the same niche with dramatically larger listening audiences.

Don’t Lose Sight of the Big Picture

While having access to a consistently, easily-measurable set of KPIs is every marketer’s preference, it is important to remember that the most meaningful metric of all is audience interest and engagement.

For me, this comes in the form of emails thanking me for airing particular episodes that people found useful or someone posting to Facebook that they used a tip they hear on The Inbound Success Podcast to improve their marketing results.

Sometimes this happens months or weeks after an episode airs, but regardless of the timing, audience feedback is the best indicator of current - and future - success as a podcaster.