Giving users the ability to personalize their podcast recommendations with minimal collaborative filtering. My goal is to:
-
Help users listen to podcasts they like
-
Minimal recommendations and notifications
Project overview:
Challenge
How might we prevent unwanted recommendations?
Design Hypothesis
Users want to select what they want to listen to with minimal notification.
Project Duration
4 Weeks
Responsibility
Lead UX/UI Designer, UX Researcher & Product Manager
Software
Figma, Miro, Zoom, Google Meet, Pen & Paper
Collaborators (Remote): UX designers, developers, and Data Scientists
Mike Roth, Stacy Leitstein, CJ Parnell, Aaron Hume, Ashley Poon, and Patrick Cudo
User Research
I wanted to identify
-
What is important to podcast listeners?
-
Which of their needs is not being met?
I interviewed 4 podcast listeners from avid listeners to occasional listeners. Some of the questions I asked our users are:
1. How do they find new podcasts?
Why?
To understand the search process of a podcast listener.
Takeaway:
1/4 job seekers
browse through the recommendations or from word of mouth.
2. What made good or bad recommendations?
Why?
To learn users' interests.
Takeaway:
4/4 podcast listeners
said a good recommendation is if its based on their interest.
4/4 podcast listeners
said it's horrible if it's a random recommendation.
3. The factors that influenced their podcast choices.
Why?
To understand why they make that selection.
Takeaway:
3/4 podcast listeners
listen to a podcast base on a friend or social media recomendation.
02
Interviewing our users
Interviewed 7 podcast listeners vary from avid listeners to occasional listeners. Questions focused on the interviewees' relationship and behavior with and factors that influence them to listen to podcasts such as:
-
How do they decide what podcast to listen to?
-
How would they prefer to be notified of a new podcast?
-
What is their experience with recommendations?
User interview takeaway:
-
Users don't like browsing through lots of podcast options.
-
Listen to podcasts based on the description or recommendation from peers.
-
Doesn't receive quality or recommendations that capture their interest.
-
Users want recommendations they are interested to.
Competitive/Comparative Analysis
I put together a high-level competitive and comparative analysis to learn about our competition. Our top 3 competitors are Spotify, Podyssey, and Listen Notes. I looked to indirect competitors both within and outside of the market.
Our number one competitor is Podyssey, a startup that came to the game in 2019 has grown incredibly fast in the past two years. This validates the demand for podcast recommendations but indicates a very strong competitor. Our goal from this point was to improve upon one feature provided by Podyssey in order to set our application apart.
From our synthesized our data:
Meet our Personas
I wanted to understand our users more in-depth. I create two personas that provide us with information about our avid and occasional listeners based on the information of our interviewees.
Reframing the challenge:
Podcast listeners need a way to locate and be informed of podcasts they are interested in without being overwhelmed with recommendations or notifications.
Design
We began sketching and ideating the user flow including the designs for our users, taking into account what our users informed us during the user interview.
The user flow and the design studio organized our ideas of how our solution was going to be organized in our prototype. Using the user flow, we discussed where each feature would live in the prototype.
Design Studio
The most important part of the product is the onboarding process. That's where our users will select how podcasts will be recommended for them. So, I began sketching and wireframing the onboarding process with 2 main things in mind:
1. Fast and easy.
2. Easy selection of podcast interest.
Sketches for web prototype:
Mobile responsive sketches:
Mid-Fidelity Prototype
and Usability Testing
We conducted a usability test with 4 podcast listeners to the user flow for our users.
Usability testing takeaways:
1. "Would be great if the become a member button is easier to locate/spot".
3. Users need the option to go back.
High-Fidelity Prototype
and Next Step
Our next step is to:
-
Test our MVP
-
Work with the data team to find more complete datasets on podcasts and individual episodes.
-
Building a profile system to allow Podmyne to stay updated with individual user interests.