STITCHER MOBILE APP

Podcast playlists, reimagined

540757+iPhone+7+mockup.jpg

Stitcher is a popular podcast app known for offering users exclusive access to coveted content. My team was tasked with creating a new playlist feature for Stitcher that reimagined how users interact with their podcasts. 

The problem

There are over 100,000 podcast shows on Stitcher. According to Salesforce, podcast consumers subscribe to an average of 6 podcasts at a time. What are the chances you’d listen to all those different podcasts in the same sequence?

Stitcher is an app that leans on its playlists feature to help users organize their listening and automatic downloads. The problem is that Stitcher’s playlists can only queue the episodes in a given playlist in one sequence at a time (e.g. newest to oldest), when in reality listeners will be moving through different podcasts in different patterns.

Stitcher users need a way to aggregate, prioritize, and download their next (or unfinished) episodes from various podcasts, because they’re typically progressing through a diverse set of shows simultaneously.

By enabling users to define their preferred listening sequence for individual podcasts and favorite their top shows, we ensure that they would have the most relevant content at their fingertips in the “Up Next” playlist. They'd also be able to conveniently auto-download this playlist for offline listening.

Stitcher lets you reorganize all the episodes in your playlist using a sort feature. If you could pre-define your preferred sequence for individual shows, though, playlists would queue smarter.

Stitcher lets you reorganize all the episodes in your playlist using a sort feature. If you could pre-define your preferred sequence for individual shows, though, playlists would queue smarter.

What we heard from users

We approached our research with a few questions in the back of our mind. What are podcast apps getting right? What are common user pain points? Where are the opportunities?

Interviews with 13 podcast listeners revealed that users consume different podcasts in different sequences.

Users often used the default “Recently Added” playlist as an easy way to find newest episodes from their most valued shows, but what if they were listening to a narrative podcast like S-town in chronological order starting from the beginning? They’d have to manage this podcast separately.

We also heard that users listened to only a small portion of the podcasts they were subscribed to. While this was not necessarily a problem, it suggested that there were opportunities to help users manage a large library of shows not all of equal value or priority to them.

Making sense of qualitative data through affinity mapping.

Making sense of qualitative data through affinity mapping.

Quotes from user interviews

“I wish I could tell the app to queue the 3 most recent episodes of a news podcast, and the 3 next episodes in sequence of my ‘backlog’ shows that I’m working my way through.”

“I have 76 different podcasts which appear in my feed, but I would describe myself as actively listening to 6 or less.”

“News and politics shows take priority when there are new episodes...hearing an old news episode is like reading a newspaper from five days ago.”

Competitive analysis

We took time to explore several top podcast apps, and developed a competitive analysis document to organize our findings. We were struck by several feature gaps, including:

  • None of the apps provided the option to auto-queue podcasts in chronological order or random order - only most recent.

  • None let users favorite their podcasts, despite the fact that listeners had clear “go-to” shows amongst their many subscriptions.

Screen Shot 2020-03-20 at 3.26.31 PM.png

Our proto-persona

We summarized our user research in a persona.

We summarized our user research in a persona.

Goals & Needs

  • Stay on top of topical/news shows that will quickly become obsolete

  • Make steady progress through content that is not time-sensitive, such as narrative shows

  • Never miss an episode of favorite shows

  • Need easy access to fresh content from second tier shows when favorite shows exhausted

Preferences & Pain Points

  • Actively listens to just a small fraction of shows currently subscribed to

  • Frequently uses “Most Recent” playlist

  • Finds it cumbersome to scroll through episode lists to find place in chronological sequence

  • When bandwidth for second tier shows is available, selects shows haphazardly

Design studio ideation method

Each team member sketched 6-8 solutions to our user’s problem, which we then critiqued and refined.

Each team member sketched 6-8 solutions to our user’s problem, which we then critiqued and refined.

We used the design studio method to rapidly generate, critique, and iterate on concept sketches of possible new features that would improve the UX of Stitcher.

We evaluated our ideas based on how well they a) met the needs of our persona and b) aligned with findings from our competitive analysis.

Ultimately, we discarded some less inspiring possibilities (e.g., an automatic unsubscribe prompt for unused podcasts, like Gmail’s), and came to consensus around our favorites.

Sketching the listening order modal

As a group, we had coalesced around 2 important features to iterate on further.

When a user subscribed to a podcast, a modal could pop up asking them what sequence they’d like to move through the show.

When a user subscribed to a podcast, a modal could pop up asking them what sequence they’d like to move through the show.

When users subscribed to a podcast, a modal would pop-up modal for users to specify a listening sequence, so that Stitcher could auto-queue episodes more accurately. A user’s preference might look something like this:

  • S-town : chronological sequence

  • This American Life : random sequence

  • The New York Times’ The Daily : most recent sequence

We also sketched how users could “favorite” their top podcasts, so that you could quickly toggle to only display content from your favorited shows.

Refining our user flows

Early feedback on one of our user flows suggested that requiring users to define the listening order for all their podcasts was bad UX. In response, we offered default sequence settings, which users could easily adjust if needed.

Below is the final iteration of our second user flow. Here, the podcast was set to queue in random sequence by default, but the user decides she would like to listen in chronological sequence (first to last) instead.

Screen+Shot+2020-03-22+at+6.47.42+PM.jpg

Our early navigation was confusing

Podcast+Pages+later.jpg

We asked users to interact with paper screens based on a user scenario, and found that we needed to make the queue settings more visible.

Test scenario: You have some chores to do around the house, and you want to listen to a new podcast your friend recently recommended to you. On the podcast detail page, you see that it has been queued in chronological order start with the first episode, and you look for a way to change the sequence to newest to oldest.

Test results: Confusion during usability tests made it clear that locating the listening order feature in a show’s "Settings” wouldn’t work. We needed to give it its own button. The same went for the option to “Favorite.”

Establishing visual hierarchy in the UI

After working in low-fidelity and mid-fidelity, it was time to move our ideas in Sketch and then InVision. Additional usability tests helped us improve the legibility of our interfaces.

When developing our high-fi style guide and screens, we used many elements of the existing UI of Stitcher, but adapted certain graphic patterns and buttons where there was clear room for improvement.

User feedback on our first clickable set of Sketch screens told us that we could do a better job of establishing visual hierarchy through contrasts in color and proportion.

We also refined the “Listening Order” button in line with Norman Nielsen’s visibility of system status heuristic: rather than simply “Listening Order,” the button itself would state what listening order was currently set to (i.e. “Playing: Random”).

 

Between these two screens, we made many small but critical improvements to the UI, such as scaling up and left-aligning the blue “Play” button of the “Up Next” queue.

Between these two screens, we made many small but critical improvements to the UI, such as scaling up and left-aligning the blue “Play” button of the “Up Next” queue.

Our final prototype

Our final InVision prototype demonstrates how a user could:

  1. Subscribe to a podcast and see Stitcher’s recommended listening order for it.

  2. Customize its playing order to match their preferences, if they differed from the default setting.

  3. Discover all their next desired episodes from various podcasts in one smart, master playlist called “Up Next.”

  4. Toggle the “Up Next” playlist between a) episodes from all shows and b) only episodes from their favorited shows.

Reflection

Throughout the project, our team struggled with vague and unstable terminology related to podcasts.

Did “queue” refer to a list of episodes (playlist), or did it refer to the sequence in which episodes would play? On multiple occasions, it became clear that the three of us were not on the same page. So, coming to shared definitions of terms like "playlist," "queue," and "sequence" within our team and for our users was critical to our success.

Illustration courtesy of humaaans and Pablo Stanley.

Illustration courtesy of humaaans and Pablo Stanley.

Next Steps

If we had more time, we would have integrated offline listening into our designs, given that multiple users wanted to auto-download episodes for listening without using data. Finally, we would have explored the idea of an automatic unsubscribe prompt for neglected podcasts (similar to Gmail’s), because our research showed that users only actively listened to a small fraction of the podcasts they were subscribed to.

 

Thanks for taking a look at the Stitcher project!

Previous
Previous

Suvie: Making Sense of Meal Deliveries

Next
Next

City Feed: E-commerce for Urban Shoppers