Bring on the goats! Apple's cloud failure demands further sacrifice
Devices are Apple's thing. Distributed systems are Google's
Apple at 40 In four decades, Apple has done more than most companies to seamlessly blend software and hardware, creating beautifully functional products. However, software (and hardware) no longer live exclusively on our devices, and the one thing that continues to frustrate Apple is the cloud.
Way back in 2011, Apple chief executive Tim Cook stated an essential tenet of Apple’s philosophy: “We believe that we need to own and control the primary technologies behind the products we make.” That has led Apple to build its own chips, scratch-resistant sapphire screens, and more. While cloud services were more of a sideshow than mainstage, this was fine. But now that cloud services comprise so much of the Apple experience, the company needs to get smart about cloud.
This is why it must be so dispiriting for the company, despite 40 years and billions upon billions spent on data centres, to have to punt on its intention to own its cloud destiny, signing on to spend $400 to $600 million with Google for cloud capacity, adding to the (presumably) larger sums it already spends with Amazon Web Services and Microsoft. Apple, rabid about user experience, is outsourcing the infrastructure used to compose essential cloud services that increasingly feed user experiences.
Can this last?
It’s not for lack of trying
Apple has come a long way in the cloud. Though the company has for years depended on AWS and Microsoft Azure to power its cloud services, the company has committed to spending billions to build out its data center footprint, including a $2bn “command center” in Arizona for its network of data centers, as well as data centers in Ireland and Denmark.
As Amir Efrati and Steve Nellis report in The Information, Apple currently has at least six distinct cloud infrastructure projects underway, including Project McQueen, to handle everything from data storage to networks and servers to “help developers to power their apps.”
This is critical, because Apple increasingly relies on services to plump its revenues and profits. As I’ve written, though sales growth of Apple devices has slowed in recent quarters, its services revenue, covering everything from media purchases to Apple Pay transactions, has boomed. In tandem, Apple has spent heavily on the cloud infrastructure necessary to support all this services growth.
The problem, however, is that too much of that cloud infrastructure is being powered by Apple’s chief rivals, Google, Microsoft, and Amazon.
Underlying that problem appears to be a lack of expertise in stitching together that cloud infrastructure with software, as Efrati and Nellis detail: “Apple has more than enough cash to buy the servers, networking gear and storage boxes it would need to handle its infrastructure needs. But the software to operate that equipment and provide cloud services for hundreds of millions of iPhone owners has bedeviled the company.”
Connecting the dots
Apple’s problem, ultimately, derives from its success: Apple is at heart a company that makes incredible devices. In my home we have six iPhones, six MacBook Airs, a few iPads, and three AppleTVs. We have gone “all in” on Apple, and at the device level we couldn’t be happier.
Where this breaks down, unfortunately, is when we need to think beyond the device. As I’ve written, Apple Music lobotomized our collective music library (which had reached a happy stasis after years of fits and starts with iTunes Match), and it took years to get my calendars to sync properly. Despite super-smart engineers (I got to work with a few of them at a previous startup) and a big budget, Apple has struggled for years with cloud services like iCloud, Apple Music, and more.
Companies like Google were born in the cloud. As Benjy Weinberger details, “distributed systems are a core part of Google’s DNA,” given the company’s need to “run complex computations in parallel on large numbers of computers.” Apple is different. Its genesis and where it continues to make the vast majority of its revenue is as a seller of high-end hardware, with software perfectly complementing that hardware.
Google, in other words, knows how to connect the dots. Apple builds the dots.
Rising services revenue
It’s hard to fault Apple for this. For most of its 40-year history, there was no real need to think about cloud services because, well, they didn’t exist. Even the company’s fits-and-starts experience with iCloud wasn’t a huge black eye because the need for generic sync across devices still hadn’t hit home in a big way.
But then Apple helped to create the problem it struggles with today: more music on iPods and eventually iPhones had to stretch to MacBooks and even to (gasp!) Windows machines. Now multiply that problem across pictures, music, movies, and other content. Apple customers spend more on and do more with their devices. Apple suddenly had almost single-handedly invented a huge problem, but didn’t have the DNA to solve it.
And guess what? It’s getting worse. As noted above, Apple customers are buying big into Apple services. The population of “engaged” Apple customers - defined as those that buy Apple categorizes an "engaged" customer as one who has purchased a service (movie in iTunes, app from the App Store, and so on) within the last 90 days - has grown 25 per cent over the past year, according to Apple chief financial officer Luca Maestri. All that engagement adds up to $5.5bn on Apple services in the last quarter, up 15 per cent year over year.
Never before, in other words, has the problem of cloud been so acute for Apple. In the short term Apple has been forced to hedge its cloud bets further by engaging Google Cloud Platform. But that’s short term.
Longer term, the company must get as smart about running distributed systems as it is about building iPhones and MacBooks. Given that so much of its future now depends as much on the cloud as it does on devices, it’s a good bet that Apple finally has the incentive to really invest in developing the DNA to run massive, highly distributed services, rather than just rent them from rivals. ®