In April 2017 to May 2018 I led the design of BBC Sounds Discover feed across mobile web and desktop. My responsibilities included user research, concept ideation, aligning key stakeholders on product/UX goals, designing user flows, visual design, modular design system, prototyping, user testing, incorporating user feedback into design iterations and pushing design into high-level strategic decisions.
So why reimagine?
Audio at the BBC was originally split into two products, BBC iPlayer Radio & BBC Music. We had two apps, two websites, two separate experiences for BBC audio content. Combine this with declining linear radio, and the rise of music streaming services, we needed to do something to showcase the breadth and depth of the BBC's audio output.
Confusing. Ok, so what were the project goals?
The main goal was to modernise the audio output of the BBC, bringing together speech and music audio content into one combined audio product. While at the same time also targeting our under served audience the under 35s. Sounds simple 🙃
So who will be using it?
BBC iPlayer Radio had an ageing audience profile due to its mainly linear radio output, whereas BBC music captured a small but important youth audience with its music mixes, playlists and on the pulse news. So combining the two, thats our audience, which is obviously very broad.
Knowing me, knowing you
First things first we needed to understand our new target audience. What are their listening habits? Why do they choose to listen to different types of audio content at different times of the day?
So how do you map this out in a useable format?
A MENTAL MODEL!
A mental model? Right, explain please. The mental model captures our users needs and behaviours split into sections that map out the journey of finding and selecting something to listen to.
To get this, we did sessions with our audience segments and discussed as a group their thought process when selecting something to listen to. This gave us a solid base for us to understand our users mental thought process when choosing a podcast over a music playlist for example.
Nice. So how'd you use it?
We used this heavily in the start of the project, even just having this on the wall helped massively to understand our audience and always keep them in mind. You can see on the mental model image above we had a swim-lane for features under the needs/behaviours, we started using this in our design sprints to map our ideas to makes sure that idea was definitely fulfilling user needs or behaviours.
Design sprints you say
Yep. At the start of the project we split the focus of the project into 3 key sections. We did this so we could focus on our KPI's for the project, and get all of our key stakeholders in a room to get input into those areas
New Users – How might we familiarise users with what we can offer them, and find out their personal taste and interests?
TESTING TESTING
Each week we followed Google's design sprint structure, meaning that we tested our selected ideas with our users on the final day. This was absolutely vital at this stage to get feedback but was also a point to embed UX practises into the editorial and business mindset, making sure we test our hypothesis early and often.
Prototypes tested during the design sprints
Feedback from users
Personalised feed
Following the design sprints, the ideas that tested well and fulfilled our user needs/behaviours were around a personalised content offering. This match our research into the wider audio market, where products like Spotify and Apple Music use data on what you have listened to to recommend users content.
Personalised feed idea
Seems obvious, but with the way the BBC is structured the data we have on our users is limited, we had no way to track what they were consuming so we could connect to a single user.
You're the real MVP
To really test our hypothesis around personalised content we needed to try it on the current iPlayer Radio site. This means real users, with real content, listening to real recommendations as an MVP. But how do you do that without building the full BBC Sounds cross platform signed in experience?
Implicit vs Explicit
To do this we split personalisation into two categories, Implicit & Explicit personalisation. When users are recommended content based on play data, thats Implicit. And Explicit is based on an action from the user, for example following, or liking. This helped us build an MVP for personalisation as we already had the Explicit actions like Follow & Like, so all we needed to do was design ways to showcase that content we get from explicit actions, then we can get data on how well a personalised experience could work with BBC content.
Bring on the brand
To reimagine the audio at the BBC meant combining to existing products. To make sure this new combined audio product was understood to be a totally new product, we had to rebrand, relaunch and refresh the whole experience.
To make sure the Sounds brand had a cohesive experience cross platform, we had to design each component to work for all platforms. Also making sure each platform could run of the same API, so that we had no platform differences. This was also important as we knew the responsive website and apps were only the start, in future we would look other platforms like voice and TV, so getting the data structure right first time was crucial.
The big challenge
Combining speech and music content into one feed. It wasn't easy. As we have gathered earlier, the mental model for speech based content is very different to music. How do you combine one feed with a mixture of content types and make sure user understand what type of content they will get.
How?
We tested many ways, like splitting the IA by speech and music. This was difficult because of the cross over of radio, and sometimes users don't actually know if they want to listen to music or speech until they see the content in front of them.
We then tested simple things like sectioning content within the discover feed, and then introducing shapes and patterns to define content types. The result below showcases the different shapes that defined different content.
If BBC sounds went to a bar, what would they order?
Weird question but this helps us understand BBC sounds' tone of voice (TOV), another part of the project that I lead on. This and many other questions helped in a workshop that I ran to understand what TOV sounds would need to reach out to our new target demographic, but also not alienating our heritage radio listeners.
This helped not only helped our editorial team to understand and document the way they should communicate to the audience, it helped with UX copy and examples like sign in up sell.
Engagement workshop with editorial
Sounds personality photo fit
Our TOV
OK, right lets see the final thing then
I've written this case study to cover over a years worth of work. It gives you a good overview of the project, but I'd love to give you an even more detailed look at this when we chat (confidence 😬). So below are some fancy visuals, because we all love a fancy visual right?
Did it work?
So BBC sounds launched in May 2018 (and I left the project soon after) but it had a slow uptake at first, with our heritage audience remaining on iPlayer radio and our youth audience still lacking some of the content that would really attract them. But after BBC radio stations started to upsell and a huge marketing campaign kicked off, Sounds picked up pace. And now a year on from launch, with iPlayer Radio gone, sounds has accelerated to become to home for all audio at the BBC as well as the the No1 podcast app in the UK.
Below shows you week Q4 of 2020, our main KPI's around total accounts, 16-35 accounts and time spent listening all smashing new records for the BBC.
Reflections
So many things to learn from this project. It was my first dive into fully leading a team, and I jumped straight into the deep end with this. I totally under estimated the amount of communication needed at the start. As a lead you need to drive the vision for the UX wider than just the team, you need to be bold and push it into all areas, from marketing to editorial, so all areas have ownership of the UX not just your team.
Also some strategic/business decisions ended up causing us issues, my goals for the future is to get UX into all conversations so that these decisions arn't made without thought for the end user.
