Can personal taste power podcast discovery?

How can human curation help audio discovery at scale?

beyonce-earbud

We spend a lot of time thinking about searchable audio, and we know that nothing is more daunting than a blank search bar. For every person on a hyper-specific quest for a certain audio story, there are many more who can’t articulate what they’re looking for — or aren’t even looking at all. When it comes to audio stories, we’re picky about the voices we’ll let yammer through our earbuds.

We don’t want recommendations that are robotic or anonymous. The most exciting recommendations have a “right place, right time” serendipity. What’s the best way to model that discovery experience for thousands — even millions — of users?


Earbud.fm showcases podcast episodes handpicked by listeners

Screen Shot 2015-11-03 at 10.43.33 AM
This week, NPR released a project that taps into the taste of thousands of podcast listeners. Earbud.fm, a web app spearheaded by NPR’s Arts Desk and built by NPR’s Viz team, curates hundreds of episodes into a “friendly” podcast listening sampler.

Recommendations come courtesy of audio “pros” like On The Media host Brooke Gladstone and celebrities like Matthew Mcconaughey, but also from regular listeners like Connecticut mom Liz Matthews, who wrote in to an NPR listener survey last Spring. A small panel of power listeners from the audio industry (full disclosure: the panel included Pop Up Archive co-founder Anne Wootton) helped sift through the responses to pick the best gateway episodes to get people hooked on podcasts.

Earbud.fm users can navigate audio picks by category, including standards like “Comedy” and more cheeky mood- and format-based categories like “Tug At My Heartstrings” and “Tell Me A Story.” You can even sort audio by who recommended it, e.g. “Pro & Celebrity Picks.” Users who aren’t sure what they want can cycle through random picks with the “Show Me Another” button.

chvrches-earbud

By including podcasts from non-NPR networks like Earwolf and Radiotopia, Earbud.fm is a resource catered not just to NPR’s existing listeners, but broader digital audiences — including, crucially, those who aren’t listening to podcasts yet. As listening increasingly takes place away from the radio dial, NPR has the interest and expertise to position themselves as podcast evangelists and tastemakers.


Product Hunt Podcasts: upvote episodes from all time

The product recommendation platform Product Hunt takes a more populist approach to podcast curation with “Product Hunt Podcasts.” Users submit and vote on podcast episodes every day. In the style of Reddit, the episodes with the most upvotes float to the top. Though hypothetically the podcast picks belong to every genre, the tech-heavy Product Hunt userbase shows bias toward technology and business podcasts.

ph-topph-day

Product Hunt’s Podcasts interface also features podcast-related products, live chats with podcasters, and featured episodes. Even the recommendation format is experimental: in addition to daily lists of upvoted podcasts, Product Hunt curates a podcast feed with one audio pick each day.

Others have dabbled in the idea of providing recommendations in the podcast form itself. Most recently, Nick Quah (author of podcast industry newsletter Hot Pod) started a podcast recommendation feed of his own. Instead of full episodes, Quah highlights featured clips, accepting submissions from listeners. His first pick was submitted to Product Hunt last week, where you can read more about it.

Other things to note about these curatorial approaches to podcast recommendation:

  • Neither Earbud.fm nor Product Hunt Podcasts link to external listening apps. Instead, they offer in-browser listening via a persistent player docked at the bottom of the page. This means discovery is limited to the content already featured — you can’t go browsing an index of thousands of shows.
  • Both privilege “evergreen content,” or audio that’s relevant but was created in the past.
  • Both treat podcast episodes as the essential units of listening, bucking the precedent of show-level subscriptions set by apps like iTunes Podcasts and Stitcher.
  • Though Product Hunt and NPR have different editorial processes for which podcasts are highlighted, both leverage the insights of the listening masses to create discovery experiences for diverse audiences.

For listeners, it’s exciting to see platforms like Earbud.fm and Product Hunt focus on podcast discovery. From the aggregate of these recommendations, we’re collecting meaningful data to power recommendations at scale in the same way that Spotify recommends music through Discover Weekly playlists, or Netflix recommends different TV shows based on each user.

tastemaker-twitter
A tastemaker pick from Twitter, available at Audiosear.ch and through our API

At Pop Up Archive and Audiosear.ch, we’re interested in how we can use recommendations to categorize aspects of “taste” and turn them into a layer of data to recommend audio from an index of thousands of podcasts, catered to different listener preferences. The tastemaker pages on Audiosear.ch are a start — stay tuned as we grow and categorize recommendations in new ways.