Bringing the AI DJ to the steering wheel
With radio dominating the huge in-car listening market, we saw an opportunity to make Spotify even more accessible on the road. Data showed users already loved using car hardware controls to engage with Spotify, so we launched an experiment to expand our reach—unlocking more Spotify features through seamless, multimodal controls to make it easier than ever to make Spotify your go-to in-car listening experience.
Scope included
Identifying Opportunity
Building a narrative to anchor the strategy with senior leadership (VP of Prod)
Getting it on the product roadmap
Setting the overall UX strategy & experience
Driving experimentation with engineering & UXR teams
Results
90% of users expressed a need to change the original context.
Data showed a significant increase in users pivoting contexts.
Users with compatible hardware became heavy users
Main Responsability
Experience design
Opportunity assessment
Product strategy
Experimentation
Insight
When looking at data, we saw that a significant user base managed Spotify using secondary controls like the lock screen, steering wheel, and head unit.
We believed there was untapped potential in harnessing these channels to enhance Spotify's overall value proposition.
So, we set out to determine whether adding modalities would make a difference.
Frequently used modalities in the car
Our data showed that majority of users interact with Spotify from multiple modalities.
Opportunity
The opportunity to enhance user engagement during driving remains immense, but increasing consumption in this context requires breakthrough approaches. Google’s introduction of Google Assistant Driving Mode demonstrated the challenges of influencing driver behaviour, highlighting the need for innovative solutions to unlock this potential.
Opportunity
As Spotify launched Car Mode, Google also attempted to capture market share from radio but failed to significantly shift consumption trends, reinforcing the idea that the mobile screen alone isn’t the key to unlocking this potential.
Hypothesis"If users are offered a seamless way to switch content when their session loses engagement, listening time will be extended."
When we had previously gathered learnings by observing user behaviour in test settings, we found that patterns often changed in real-world scenarios, leading to significant discrepancies between qualitative and quantitative data.
Research in the car
Scope
We knew from experience how challenging it was to gather accurate insights from controlled testing. User behaviour in test settings often differed from real-world scenarios, creating discrepancies between qualitative and quantitative data.
To address this, we conducted a diary study with Spotify users identified as drivers. We tracked their interactions and supplemented the data with follow-up interviews for deeper insights.
Success metric:
Number of times users change playlists in a session
Metrics
Determining when a session ends naturally is challenging, as the duration of a drive is unpredictable. To address this, we’ve chosen the number of pivots within a session as our key metric.
While an increase in pivots might signal we’re serving the wrong content, it also indicates that users can effortlessly switch contexts. In that case, we can be confident we've created a low-friction experience.
Goal:
Our goal was to identify a suitable button combination that could be activated from the head unit, steering wheel, or mobile phone lock screen.
Pressing the buttons would prompt Spotify to switch to another context, with a text-to-speech (TTS) system announcing the new context.