Spotify Case Study

Users who use while driving

 

When it comes to my usage of Spotify, I'm pushing the threshold of a power user.

Seven years ago, I knew within the first two days of a 30-day trial that I would be marrying my bank account with a Spotify premium subscription.

My relationship with Spotify quickly grew from desktop/work usage to usage everywhere, and that included driving.

Still does.

 

It took about one or two miles of driving and navigating Spotify to realize my love affair with Spotify was being strained by small navigation and audio controls that would have the cuddliest of cuddlers asking for space. 

I wanted to put this annoyance to the test to see if I was being needy, or if Spotify had some growing to do. 

And what better way to test an app in a driving scenario than to build an epic driving simulator, right? 

 

Here's my thinking:

If navigation and audio controls are larger, then users who drive while using Spotify will be able to drive more safely. 
 

Continue reading to learn about my testing process.

OR

Fast-forward for my design decisions.

 
 

How I'll get there

This was an academic exercise, so existing user data was not available to me, so a provisional persona was used.

To help tell the story of who this ideal user would be, I created a couple "Jobs to be Done" stories and sketched out a use case scenario to provide some visual context of how an ideal outcome would look.

My pool of five testees would have benefited from having more power-type users who drove more. As it were, three of my test subjects use Spotify while driving, but limited their usage during those times. Their results added value nonetheless.

Bloopers

Using a driving simulator to test wasn't the first method I attempted. I contemplated a few others:

• Uber: I didn't want to be responsible for an accident
• Uber Pool: The idea of testing many users in a small space seemed nightmarish. And, no driving could be mimicked.
• My vehicle or yours: This was embarrassing.

Hey! My vehicle or yours?

Life is funny. I got caught up in wanting to make this test as real as possible and completely overlooked just how creepy it would be to ask strangers to sit in my vehicle or theirs and do a little testing.

Lesson learned, I took my hot embarrassment and headed home where I was hell-bent on making a driving simulator work. A driving video of sorts is all I really needed to make the test work to an acceptable level and I would be set.

Low-and-behold, there's a subculture of POV video recording junkies that enjoy taking drives and recording them. Which also means there are people out there that enjoy watching those drives.

Life is funny, indeed. Thank you, YouTube.

 

Windy road ahead

The simulator's effect on people surprised me. It was actually making them nervous and this was great; there should be a level of nerve at play.

During testing, I was surprised at how long some users would spend looking away from the road. On average, 1-2 seconds was by far the most common duration, with 3-4, and 5+ seconds falling off sharply. 3-4 seconds, much worse 5+ is an incredibly long and unsafe amount of time to have eyes off the road.

 
 
 

At a speed of 70mph, it takes a vehicle 4.6 seconds to travel the distance of a football field. 55mph doesn't improve much: 5 seconds.

Dude.

 
 

In a use case scenario, 1-3 seconds of looking away from the road were what I scribbled down as acceptable–a gut call–but after observing tests, it seemed under two seconds might be a better ceiling of time.

 
 

Studies show we don't know

An online search for "taking your eyes off the road for 2 seconds," "safe length of time to look away from the road," and "average time looking away from the road" brought up evidence that no one really knows what's considered safe.

There are too many variables.

What we do know:

• Most States and insurers suggest keeping 2-3 seconds between you and the car in front of you.
• Drivers take eyes off the road an average of five seconds 
• Looking away from the road for more than two seconds doubles your risk of an accident 
• It takes 4.6 seconds to drive the length of a football field going 70mph.

 

So, my gut assessment of about 2 seconds, when stacked against driving distance suggestions alone, means I'm not really solving much.

By my standard, when you bring your gaze back to the road you will be just in time to see your accident manifest.

I'm really solving things now! 🎉😐

This indicates a far more dramatic solution is needed, but until voice-automated is everything, and self-driving cars everywhere, I'll forge on. 

With recordings and interviews complete, I pulled out insights and began making sense of it all.

Affinity mapping proved to be an effective tool for distilling my new-found insights and two primary pain-points surfaced: Size. Readability.

Size: Overall feedback was in line with my hypothesis. Navigation and audio controls were complained to be too small and hard to tap.

Readability: Much of the typography in Spotify is rather small, especially for drivers. Complaints of not being able to see words, or words being hard to read were common.

 
 

Reassembly

Armed with insights, I began iterating solutions with quick sketches. As a gist, various changes in size, spacing, and layout were explored during this phase.

I didn't want to overhaul the app with a new design; that'd be far too assumptive of me given my lacking knowledge of Spotify's user or business goals.

My suggestions are meant to be seen as an update to the current interface that would feel like a natural next step in the apps' progression.

Would this be unsettling for non-driving users? To a small degree, I would hope. In my opinion, any worthwhile change is uncomfortable. And then it's not again.

Feeling pretty good after having toured my prototypes around to some office mates, I jumped into building out a prototype based on the UI sketch that was most well received.

 

Hover for design decisions 

 
 

Tap for design decisions 

 
 

BEFORE: Things to be done.

AFTER: Things have been done.

Ohhhh, that's pretty ...

Distracted driving continues to be the leading cause of traffic accidents. 

In a study performed by PNAS, 68% of severe accidents were caused by a driver not focused.

And in 2015 deaths related to distracted driving grew from 3,179 in 2014 to 3,477. Respectively, injuries decreased from 431,000 to 391,000; a large number, still.

It's not just phones. Food counts. Talking counts. But, people these days are focusing more on phones, not fries.

Citation: Vietnam casualties

Citation: Trade Center death toll

 
 

WTF just happened

With a workable prototype in place, I retested the most common task flows that caused friction during the initial test that pertained to navigation and audio controls.

The improvements from a black and white perspective were a success. Increasing the size of navigation icons and the tap zone around them, along with improvements to the space surrounding audio controls, and the overall increase to typography size made for a safer driving experience while using Spotify. Seemingly anyway.

Rarely is something ever black and white.

One user commented, "It's better, but still kind of small." I believe that's a fair assessment, too.

And, let's not forget my acceptable 2-second window of time was basically a wash according to safety data.

The aim of this test was to see if minor improvements to Spotify's interface would make driving safer, so on that gray scale of success or failure, I landed successfully, and discovered there is much yet to be learned and conquered in the realm of the driving user interface.

 
 

It's worth mentioning that during the course of this project, an article surfaced about a driver interface Spotify is developing. No details were given, but this news was timely, being on the heels of Audible’s new driving interface announcement. A step in the right direction.