Podcasting Shows the Value of an Open Internet

Photo: Patrick Breitenbach. CC-BY.

I’ve listened to podcasts pretty heavily for about 13 years now, since Apple added a podcast directory to iTunes. (I probably manually downloaded a few podcasts even before then, but the process was pretty cumbersome.) But I’m not here to brag about my podcast cred, which would be incredibly nerdy even by my standards, but to outline how the structure of podcast distribution is almost ideal for people who are concerned with private platforms having too much control over speech. It can serve as a model for how cool internet services don’t have to come at the cost of enabling monopolistic private platforms or giving up your privacy.

I find when talking to people about podcasts that many people don’t exactly know how they work. Apple still runs the largest directory of podcasts, but many people think that Apple actually hosts podcasts on its servers, in the same way that it hosts iOS apps, or iTunes music and movies. But it doesn’t! Podcasters themselves have to host their own files, or pay some service to host them for them. They then generate a special web page, called an RSS feed, which contains a machine-readable list of all the different available episodes. Podcasters then can submit a description of their podcast to directories like Apple’s, along with the URL of their RSS feed. Having a single place to search for or browse podcasts is obviously very convenient, and subscribing to a podcast in the directory just requires the push of a button.

However, because podcasting is a fundamentally open medium, even if Apple decides to remove a podcast from its directory, or just not list it to begin with, that doesn’t mean the podcast is unavailable. Most podcasting software, including Apple’s, still allows users to manually enter a URL to subscribe to, for instance. This is very different from platforms that host content themselves, such as YouTube or Spotify.

The open nature of podcasting also means that there are lots of third-party podcast apps out there for the major platforms. Now, here is a possible weak point. Apple runs the largest podcast directory, and very few independent apps have the resources to run their own (nor would most podcasters bother to submit to some small directory). As a result many independent podcasts apps just copy Apple’s directory. It’s good that Apple allows this to happen! But it does mean that Apple has an even larger impact on what podcasts are easily visible to users than even its own (very large) user base would suggest.

However, until fairly recently, many large non-Apple “podcasting” services weren’t really providing podcasts at all, since they insisted on hosting the audio files themselves–in Stitcher’s case, reducing their audio quality, and adding new ads. But Stitcher no longer does this, and Google has seemingly given up on its proprietary Google Play-hosted podcast strategy in favor of a more open approach with Google Podcasts. Spotify still insists on hosting audio files itself.

There are a few reasons why a service might prefer to be more closed. First, it does mean that it can control the quality of the experience. Because podcasters themselves are responsible for the bandwidth of distributing files, in some cases their servers might be slow, or other errors can happen. Podcasts themselves might be encoded poorly–for example file sizes might be 10x too large, or the files themselves just might not play. While I don’t think it’s worth the tradeoff, handing over control to a big tech platform can solve those problems. (It can also provide a more seamless way to handle paid-subscription podcasts.)

Another reason, however, why there is always an incentive for companies to try to close off podcasts has to do with data collection. In the open world, by virtue of hosting podcast files themselves, podcasters necessarily know how many times a file is downloaded or streamed, and the IP address it’s downloaded to. But they don’t get to see what other podcasts you’re subscribed to, whether you actually listened to an episode, or what parts you skip. That information, if it’s available at all, has to come from the developer of the listening app itself. Apple, for instance, provides some metrics to podcasters. But other podcast apps do not share this information and even block certain other data collection methods. A closed-off system takes away even the ability to easily opt out, and makes podcasters dependent on the platforms even for basic information like the number of downloads. Also, in a closed-off podcast ecosystem, it would be easier for podcasters to insert personalized ads into each downloaded episode, based on granular data about the listener that is just not available under the status quo.

All things considered I think the positives of an open podcast system are overwhelming:

  • While large platforms such as Apple can make discovery of podcasts more difficult, they can’t actually remove podcasts from the internet. (Actually removing podcasts themselves from the internet is of course possible, but much more difficult.) Listening to a podcast just requires a URL. Contrast this with YouTube taking down video, which might not be available anywhere else.
  • The amount of data that it’s necessary to give up to listen to a podcast is minimal–just what you’d have to share to download a file in any circumstance. If a popular podcasting app starts to get too invasive, listeners at least have the option of using a different app.
  • Third-party apps are possible, which might protect privacy better than apps from major platforms, offer different or multiple podcast directories, and so on–not to mention offering superior user interfaces and audio playback controls. And even if Apple or Google were to make it impossible to subscribe to a podcast via its URL, third-party apps would likely still offer this option.

I think the open nature of podcasting is one of the few success stories of the decentralized internet since the heady days of Internet 1.0. Most internet services used to be decentralized–Usenet and IRC, for instance. Non-internet computer services like FidoNet and UUCP were, as well. But most of these have declined in popularity, with just the web itself and email in wide use. Now people spend a lot of time on private social networks like Facebook and Twitter, chatting using private apps like Slack, or on centralized media sites like YouTube. All of these are subject to the whims of private companies, which have an outsized influence over online speech, and which typically collect lots of user data. Podcasting shows that another way is possible, and its popularity and scale shows that truly decentralized and federated online services are possible and may one day thrive, even if the internet is currently dominated by private platforms run by massive companies with no responsibilities other than shareholder value.

Source: Public Knowledge.

Creative Commons License
Except where otherwise noted, the content on this site is licensed under a Creative Commons Attribution 4.0 International License.