As a non-millennial Old Person™️, I’ve come to regard the rise of dashboard touch screens in cars with increasing derision. (In my day, we had knobs and buttons!) But the trend is clear: 82% of vehicles sold this year came with touch screens, according to market data cited by Consumer Reports. Tesla has led the charge on this front: The center console of its Model S sedan is one enormous touch screen. And customers, apparently, love it: Tesla’s “infotainment system” recently topped a Consumer Reports owner-satisfaction survey.
But now, Nielsen Norman Group—one of the most prestigious design consulting firms in the world, and whose cofounder, Don Norman, literally coined the phrase “user experience”—thinks Tesla is on the wrong track with these mega touch screens. Ha—I knew it! But . . . why?
Before we get into that, a disclaimer: Raluca Budiu, Nielsen Norman Group’s director of research, doesn’t think that touch screens in cars are a priori awful. Buried in her lengthy, technical article are a few compliments for Tesla: The huge screen makes it easy to see multiple information sources at once; it’s really good at pointing out charging stations on a map; and “the autopilot and self-navigation systems acknowledge the possibility of failure.” (Damn, that’s faint praise.) “Many of these features should make driving a safer and more comfortable activity,” she writes.
But that’s the key word: should. In reality, she argues, small but fundamental design flaws can make car touch screens overly fussy to use in cars. And when you’re traveling at 60 mph, that fussiness has a higher cost—particularly in a Tesla, which puts so much dashboard functionality in its touch screens that The Verge called the Model S a “tablet on wheels.” As Budiu puts it: “In a car, time spent with the UI is time spent ignoring the road.”
Giant touch screens are nice, but offer no haptic feedback
The Tesla Model S’s entire center console—the space between the two front seats that’s traditionally studded with physical knobs, buttons, and dials—is one enormous 17-inch touch screen. It looks eye-poppingly futuristic, and goes a long way toward making owners feel like they’re driving a “magical space car,” not just an automobile. But like any “pictures under glass” UI technology, Tesla’s controls require you to look directly at them in order to operate them. In the terminology of interaction designers, they lack “haptic feedback”: Your fingertips can’t tell what they’re touching (other than a sheer, featureless glass surface).
Unfortunately, fluently controlling an interface while focusing your visual cognition on something more important requires haptic feedback. As Budiu explains, this is how people can play the piano while reading sheet music (or, in my case, touch-type while composing this sentence). “With a physical button, we can learn its location and acquire it without directing much, if any, attention to it,” she writes. When driving a car, your visual cognition should be almost wholly occupied on the road, not hunting and pecking through submenus on a touch screen.
But even if Tesla’s touch screen skeuomorphically recreated an old-fashioned center console with large controls that stayed in fixed, learnable locations, this haptic problem would still persist. “Locating a soft button requires us to visually confirm its position,” Budiu explains—even if you’ve used that button a million times. (Try it for yourself: Put an iPad on a tabletop and try to type in your passcode while staring straight ahead.)
Budiu doesn’t say it, but this lack of haptic feedback implies that the difference between using Tesla’s built-in touch screen and doing something objectively dangerous, like texting while driving, is merely a difference of degree. In both cases, you must focus your visual attention on the interface (and ignore the road), or it simply won’t work.
Touch screen targets are poorly positioned
Budiu really lets Tesla have it here, saying that the main row of soft buttons “are placed at the very bottom of the 17-inch screen—an area that is next to worst possible.” The actual worst possible placement would be on the right edge of the screen, according to an interaction design principle called Fitts’s Law, which states that the smaller and farther away a target is from your hand, the longer it will take you to successfully point to it.
In other words: Not only does the touch screen force you to divert your eyes from the road, the placement of buttons on that screen perversely prolongs the amount of time your attention will stay diverted.
It’s interesting that Tesla’s Fitts’s-Law-breaking becomes a problem only because of how it entangles with the touch screen’s lack of haptic feedback. Stretching your arm all the way across the dashboard to fiddle with a vent may not be ideal from a Fitts-ian standpoint, but you can accomplish it without ever taking your eyes off the road. The Tesla’s touch screen, though, makes no-look operation impossible—which raises the stakes for any non-ideal button positioning.
Offering incomplete information too easily
If there’s one context where taking your eyes off the road to do something is actually the lesser evil, it’s when you’re checking your blind spot to change lanes. Why? Because swerving in front of other cars is an inherently risky maneuver, and to pull it off safely, you need to know with as much certainty as possible whether anyone’s in your way (or about to be). As we all learned in driver’s ed, the foolproof way to maximize this certainty is to . . . wait for it . . . turn your head and look. Shocking, but true!
So how does Tesla’s touch screen mess this simple habit up? According to Budiu, it does so by working too well—by offering a “lower interaction cost,” to use the technical term. Every time you activate the turn signal, Tesla’s center screen displays a bird’s-eye view of the road area around your vehicle to aid you. It even shows other cars, and marks the lane red if it’s blocked. And since flicking your eyes down to the touch screen is actually easier than twisting your head all the way around to check your blind spot (i.e., the interaction “costs” less, time- and effort-wise), that’s what most drivers do. Magical space car, amirite?
Well, not quite. Just like Autopilot (another “magical”-seeming feature whose low interaction costs may subtly train you to use it unsafely), Tesla’s lane-changing visualization isn’t 100% accurate. “In fact, Tesla warns against relying solely on lane assist for lane changes,” Budiu writes. You can’t fault a corporation for its CYA disclaimers, but a feature whose very ease-of-use encourages drivers not to seek more reliable information (i.e., that whole look-with-your-damn-eyes thing I mentioned earlier), and potentially risk their safety in the process, may have a fundamental design flaw.
And so concludes Norman Nielsen Group’s case against in-dash touch screens, especially Tesla’s. Are they a menace to society, a 21st-century version of exploding Ford Pintos? No. But from a design perspective, the difference may be in degree, not in kind. The design choices behind Tesla’s touch screens solve certain problems while creating others. Not to be a cynic, but the “problems” that car touch screens solve seem more on the automakers’ side than on a driver’s. A giant touch screen is great for marketing; it looks undeniably futuristic, and people love jabbing them. It also lets the automaker update and expand a car’s features continuously (and at lower cost), just like an iOS app. These “solutions” certainly overlap into customer benefits, too. But when touch screens arguably degrade an automobile driver’s most important job to be done—operating the vehicle safely and comfortably—then they may need a serious rethink, regardless of how “magical” they seem.
Tesla did not respond to a request for comment by press time.