There couldn’t have been any better time to publish this article right after Spotify purchased the exclusive distribution rights for the hugely successful Joe Rogan Experience (JRE) podcast, costing them back by $100M.

I honestly think it was a pretty smart deal by Spotify, despite the amount.

A recent report conducted by NPR and Edison Research showed that in 2014, 80% of audio listening in the US was music, and 20% spoken-word. But in 2019, the numbers were 76% and 24% respectively. This suggests that podcasts may be cannibalising music streams. If you are a podcast fan, when was the last time you streamed music on your daily commute to work instead of listening to podcasts?

No company would try and push for a new product that competes against its existing core product, but I believe this is the case for Spotify so that it can re-negotiate better deals with recording publishers and owners to its advantage. Only that now, Spotify has leverage.

🎶 Fixing the Broken Record

But can one podcast show alone save Spotify? There are many things still broken with the podcast industry:

Monetisation - Is sponsored ads the only way for podcast content creators to generate revenue?
Analytics - How can marketers know if podcasts as a marketing channel is valuable to them when the only standard analytics provided from podcast distribution platforms are number of downloads and number of followers?
Discoverability - How can someone discover a podcast based on its content? How can someone easily share moments of content with others?

In order for Spotify to strengthen its leverage as discussed earlier, it would make sense for it to focus on fixing discoverability - the more people discovering new podcasts to listen, the less people will be listening to music streams. Product features around discoverability will be the focus of this post.

📗 Judging a Podcast By Its Cover

Podcasts is a very different medium from videos or music. An average podcast duration is about 43 minutes. It is therefore not something I would consider as ‘fast media’ - information that can be consumed rapidly, unlike endless viewings of cat videos on YouTube or listening through an entire music playlist on Spotify.

To devote that amount of your time to a podcast without any inkling of how that particular listening experience will be is akin to asking someone to watch a movie in the theater based on its poster and without any trailers. And this has become a huge problem for podcast listeners who currently have to rely on show notes or personal recommendations from friends. In a 2020 survey conducted on podcast listeners, when asked what was the first they did to find a new podcast to listen to, the results were pretty spot on in highlighting the problem:

40% searched their podcast listening app’s directory
18.3% directly asked someone they knew who likes the same stuff
15.2% asked on social media or an online community
13.6% searched on Google
13% browsed through a podcast chart of ‘featured’ section

In the same survey, the podcast’s description came out as the most important factor when a listener judges the podcast’s potential.

Two things came to mind when I saw these results. Firstly, Spotify’s search algorithm only takes into account the title of the podcast episode and does not return matches for search queries found in show notes or descriptions. That implies that the search result quality is not high, yet 40% of people still use search, possibly for a lack of a better alternative.

Secondly, podcast description is not the most accurate proxy to determine the quality of a podcast. Audio quality, narration style and pace do affect podcast quality and these cannot be captured in the podcast description.

Something better must be available for podcast listeners.

🔎 Taking a Peek Into Podcasts

That something better starts with previews. Like how Spotify has included 30-second song previews for its non-logged in users, the company should build a feature around podcast previews that would allow users to have a glimpse through a podcast before committing to listening to it fully.

But unlike the song previews, podcast creators get to decide which 30 seconds of their episode to upload to Spotify as the preview clip - along with a full audio transcription of that audio clip.

This solves the problem of discoverability in a few ways:

Listeners can browse through these previews in a feed-like UI and decide very quickly which podcasts to continue listening to without having to listen to an entire podcast
Content creators are the best to know which section of a podcast is most interesting rather than relying on the distribution platform to decide. Even if the creators don’t know best, updating the clip to experiment, like how you would for A/B testing a landing page copy, would eventually lead to an optimised preview sound bite.
By receiving uploaded audio transcripts, Spotify has access to more metadata to best categorise the podcast and return high-quality search results.
Sharing of the previews as URL links on social media can be enabled to allow for easier word-of-mouth recommendations, as like Spotify playlists.
Additionally, Spotify could also include these previews in between transitions to another episode listen, which Spotify potentially could charge podcast creators as a part of a marketing service.

All of these requirements within the feature aren’t new to Spotify as it has already implemented them into its core music-streaming product. I believe it would not be absurdly complex for the company to make this feature a reality.

🥊 Competition

Many startups have tried to solve the problem of discoverability via podcast directories. One particular and fairly young startup based in SF, Shuffle, does exactly that of the feature outlined above, ~~except that it does not link the previews to the actual podcast (as at time of writing)~~ except that it can only play full podcasts that are included in the Listen Notes API service.

Which leads me to the point of there being so many disparate podcast listening services. Some podcasts are on certain platforms and not on others. This has led to users downloading mutiple podcast listening apps and podcast creators having to upload their RSS feeds to as many apps as possible (although made easier by tools like Transistor.fm), in a bid to achieve high discoverability for their podcasts.

Creating a solution for discoverability solves this, as these creators would flock to the app which give them the highest discoverability.

The end-to-end user experience of listening to podcasts in general is not ideal, having to jump around multiple apps just to complete a single journey of discovering a podcast and listening to it. Major competitors like Apple Podcasts have also not ventured into solving the discoverability problem yet, giving Spotify some time to own the solution to this problem.

🏁 Conclusion

Podcast discovery is broken and the only way to solve it is by allowing listeners to de-risk the possibility of committing tens of minutes to a poor (according to their taste) podcast, much like how you would flip through a few pages of a book at the bookstore, before parting with your money.

As highlighted early on in this post, existing solutions for podcast discovery are sub-optimal. Building a feature around podcast previews in my opinion can solve this, and the solution was starring at Spotify the whole time.

Cheers.

*last updated on 25 May 2020

Thanks for taking the time to read this post. I hoped you enjoyed it and if you did, it would mean a lot to me if you’d share this post with others (or maybe some folks at Spotify?). I’d also love to hear any feedback as well which you can shoot to me via twitter at @markjrobert.