The Evolution of CarPlay

on February 19, 2015

Given the reports over the past week or so about Apple’s ambitions in the car space, I’ll share some thoughts related to a report on the connected car I’ll be publishing in the next week for my clients. This is a huge and broad topic, so I’m only going to focus on Apple’s potential role in all this and, specifically, the evolution of what is today called CarPlay.

A natural evolution

There’s a natural evolution in the relationship smartphone vendors such as Apple will have with the car and it’s mostly a question of how far this evolution will go. The diagram below shows four possible steps in this, with the red box in each case defining the contribution of the smartphone vendor to the overall solution:

Evolution of CarPlay

Phase 1: Generic UIs controlled by car OEMs

In this evolution, Apple’s CarPlay is currently in the second phase. The first phase was everything that came before, with fairly generic sync capabilities to allow smartphones to extend some of their functionality onto the in-car infotainment systems. The UI for these solutions was entirely controlled by the car OEMs, through systems designed by the various Tier 1 vendors that built these systems. Ford’s Sync has been one of the most sophisticated and they have just debuted a new version based, for the first time, on BlackBerry’s QNX rather than Microsoft’s in-car technology stack. But in such a model, a smartphone vendor such as Apple gets zero control over the in-car UI and almost no opportunity to extend its proprietary features such as navigation or voice control onto the car’s infotainment system.

Phase 2: OS-specific UIs for the infotainment system

2014 saw the introduction of phase 2 in this evolution, with Apple announcing CarPlay and Google announcing Android Auto. Both represent solutions for extending OS-specific UIs and apps onto the in-car infotainment system, focused on the main infotainment display. The smartphone vendors now define this experience to a far greater degree, including the look and feel, the specific apps and interaction models, and so on. Their sphere of influence has extended from the smartphone itself and an element of syncing technology to the display within the car. But this model is challenged in several ways: first, it threatens the car OEMs’ vision of what the in-car infotainment solution can be, namely something they can differentiate on and monetize. Having the smartphone and OS-specific UIs at the center removes their ability to differentiate and to monetize, putting smartphone vendors and car OEMs in conflict. The second challenge is that such solutions are in some ways a band-aid, sticking an OS-specific UI on top of the overall UI and in-car experience, which includes many other elements totally disassociated from them. Many modern cars have at least two information displays. Quite a few have as many as five, including the main infotainment display, the traditional dashboard dials, a small information display in the dash, a clock and climate control display, not to mention a slew of dials, knobs, paddles and so on to interact with all these. CarPlay and Android Auto represent progress beyond phase 1, but they suffer from conflict with the car OEMs’ objectives as well as a very limited role in the overall technology systems in the car.

Phase 3: Broader participation in in-car systems

I mentioned QNX and Microsoft earlier in the context of phase 1 systems such as Ford Sync. These companies have traditionally been among the major suppliers of in-car operating systems and software. Yet these companies are chosen, not for their ability to provide great user interfaces, but for the underlying technical capabilities they have, which are well-suited to the core functions of the car and the driving experience rather than the user interfaces and displays in the car. Microsoft lost the Ford business to QNX precisely because its system was poorly suited to the way consumers expect touch displays to operate today. One possible role for the Apples and Googles of the world in phase 3 is to extend beyond the band-aid-like approach taken with CarPlay and Android Auto into the in-car OS itself. This requires a much deeper integration with the car itself, beyond the infotainment system and that, in turn, requires going beyond each company’s traditional competencies. Apple and Google have always provided consumer grade operating systems and software, whereas both Microsoft and QNX have a long history in industrial grade operating systems, with both security and reliability characteristics suited to the highly performance sensitive environments in which their software has to operate.

I think it’s entirely possible Apple is exploring this kind of extension of CarPlay into a fully-fledged in-car OS that would go well beyond the current capabilities of CarPlay but, if so, this is a challenging task. There’s no doubt Apple’s expertise could make a huge difference to the in-car user interfaces we’re used to seeing but getting its operating systems and software ready to run core functions in a car would be a big leap. There is already talk about using Android as a potential operating system in cars and there are specific rumors about the next version of Android coming with a flavor designed explicitly for cars, but this is a leap for Android too, as a platform not known for robustness and flawless performance. On the other hand, Google is working in parallel on self-driving cars, giving it a great deal of insight in to and experience with what it takes to design software for more than just a car’s various displays. Given Google’s investment in this area, Apple may be tempted to make a move analogous to Google’s acquisition of Android, which was designed to prevent a Microsoft hegemony in mobile operating systems.

Apple may be better suited to work in close partnership with car OEMs to provide the design and user interface elements they’re so poor at, while leaving the underlying systems to those with a longer history. I’m not convinced Apple has the capabilities or the strategic reasons to participate at this lower level in the car technology stack, when participating at the higher, user-centric layers, makes a ton of sense. This is, to my mind, the most plausible explanation for the recent rumors about Apple’s broader ambitions in the car.

Phase 4: Building a car

Of course, as long as Apple works with the carmakers, it has several problems:

  • The inherent tension with carmakers who want to differentiate their experience, not adopt an Apple-flavored experience which might be shared with other carmakers and who want to monetize their own infotainment services rather than simply see Apple extend its services into the car
  • The inability to completely control the experience from end to end, as Apple tends to want to do. Steve Jobs was always fond of Alan Kay’s aphorism that “People who are really serious about software should make their own hardware,” but Apple wouldn’t be making the hardware, or even controlling it, in the phase 3 scenario. CarPlay shows, in a small way, what happens when Apple can’t have complete control over the UI for something it designs: a wide variability in the quality of the implementation within the infotainment system
  • The fact competing smartphone vendors may well not want to integrate with a system designed by a competitor.

Tesla has demonstrated that it’s possible to totally overhaul in-car systems in such a way that both provides an enormously better UI, while adding a technology layer to core car and driving functions, but it’s in the position of having built the whole thing from top to bottom. Apple seems like one of the few companies out there that could match this level of reinvention, but it simply doesn’t have the heritage in actually designing and building cars. It’s been reported it would use a manufacturer to build the cars but it would still have to design the car itself, something that’s entirely outside of its core competencies – designing a car, after all, isn’t just about the outside and the technology functions but, most importantly, the mechanical functions Apple has no history of. Again, Tesla has demonstrated that it’s possible for a new entrant into the market to do these things enormously well, but it’s not clear Apple would benefit strategically from designing a whole car top to bottom when the technology functions are the main area where it would add value.

Phase 5: Self-driving cars

Of course, the whole concept of the connected car may just be considered a step along the way to autonomous, or self-driving, vehicles. This is obviously an area where Google has already invested significantly, and Uber, Tesla, Audi, and other car manufacturers are also far down this path. I continue to be skeptical we’ll see mass-market self-driving cars capable of driving nationally within the US, let alone globally, within the next five years, but it’s fairly clear this is an attainable goal within a longer timeframe. If this is what Apple is working on, which I consider a pretty long shot at this point, it’s a project that likely won’t see the light of day any time soon. But I continue to think the most likely scenario is Apple is working on what I’ve described here as phase 3: an evolution of CarPlay that goes deeper into in-car systems.