The iPhone 6 has set sales records since its first week of release and led to Apple's most profitable quarter ever. There have been some bumps in the road, but the market—and Apple's customers—continue to have that loving feeling. What could possibly go wrong?
In a word, software. A growing chorus of developers and Apple-watchers is raising the alarm that the buggy releases of iOS 8 and Yosemite are part of a systemic decline in the quality of Apple's software. The now-yearly release schedule for both iOS and OS X combined with the increasing complexity of the overall Apple ecosystem have put a strain on its engineers, these voices say. At the same time, Android continues to increase its overall market share while solving many of the fragmentation and quality issues that have made it second to iOS in terms of user experience. One of the leading critics of Apple's software quality, longtime iOS developer Russell Ivanovic, has even christened 2015 as "The Year of Android." Ivanovic has been building iOS apps for many years, but his studio, Shifty Jelly, now makes 80% of its income on the Android side of the fence. In a post titled "It Just Works," he writes:
Tim Cook keeps telling us that ‘Only Apple’ could do the amazing things it does. I just wish that Apple would slow down their breakneck pace and spend the time required to build stable software that their hardware so desperately needs. The yearly release cycles of OS X, iOS, iPhone & iPad are resulting in too many things seeing the light of day that aren’t finished yet.
Anyone who follows Ewan Spence's Apple Loop column or Gordon Kelly's posts is aware of the many serious bugs in iOS 8. Far fewer people use OS X, so users have been less audibly vocal about Yosemite's problems, but they are equivalent in severity. To quickly recap, iOS 8 has had problems with Photos, ringtones, data usage, Family Sharing, iCloud syncing, cellular network connectivity, WiFi performance, Messages, Accessibility and security. Yosemite meanwhile has had persistent WiFi issues along with iCloud and UI glitches and performance problems on not-so-old Macs. Yes, this is software, and there will be bugs and glitches. That's not news.
The only thing that could be news is if this is a trend that is getting worse. Macworld contributor Kirk McElhearn has written extensively on the Mac ecosystem. A series of his recent posts tells his side of the story well: "Why Has So Much Gone Wrong for Apple Recently?" "Is Apple Trying to Do too Much Too Quickly?" and "Why Are So Few People Updating to iOS 8?" McElhearn writes "Apple is finding it difficult to keep up with all these releases, and that quality is slipping.… Bugs abound; shoddy releases are followed by broken updates." Ivanovic and McElhearn are not Apple haters. To the contrary, their livelihoods depend on the health of Apple's software. Indeed, one of Apple's biggest fans, John Gruber of Daring Fireball, wrote back in October:
From the outside, it seems like Apple’s software teams can’t keep up with the pace of the hardware teams. Major new versions of iOS aren’t released 'when they’re ready', they’re released when the new iPhone hardware ships. On Twitter the other day, I suggested that perhaps Apple should decouple major iOS feature releases from the iPhone hardware schedule. That’s probably untenable from a marketing perspective, and it might just make things more complex from a QA perspective. But something has to give.
None of these critics seriously thinks that Apple will slow down its hardware release cycle or uncouple its system releases from its hardware. But all featured in Michael Tsai's roundup agree with Gruber that something has to change. Yoni Heisler, who brought this urgency into focus in a post on TUAW, The Unofficial Apple Weblog, points to the cultural factors at Apple that may be contributing to the problem. Up until now, it has worked for Apple to keep its software teams fairly small. When it comes to people, Apple has always valued quality over quantity. "Steve Jobs made no secret of the fact that hiring a select few A+ players is preferable to hiring a large number of B players," Heisler writes. "This hiring philosophy is still a part of Apple's DNA, which is to say that simply adding more bodies onto the iOS and OS X software teams isn't necessarily a clear-cut solution."
Based on the sales of both iPhones and Mac hardware, there is no problem at the moment. Whatever issues users are having with Apple's software, it is not stopping them from buying the hardware. But as is always the question with Apple, will this continue?
Apple's power in the market originates in the huge store of positive feelings lodged deep in its customers brains. When people talk about the intimacy and love Apple users feel about the products they are actually referring to something quite physical and specific. Each touch, each swipe each positive experience encodes as rows and columns of neurons in our brains in areas categorized as "Apple." For many of us, those areas take up a fair amount of space. Seen as a whole, this quantity of positively associated brain cells serves as Apple's brand inertia.
The default for the vast majority of Apple's customers is that the next "x" will provide a better user experience than the last. That variable could be iPhone, iPad, iMac, MacBook, iOS or OS X version. The power of the default lies in our assumption that this assertion is plausible and therefore likely. This assumption will remain plausible in most people's minds even after it is no longer probable. What would it look like if user's expectations began to turn and how soon could we discern the pattern?
iOS 7 vs. iOS 8 in terms of user adoption. Note that the curve for iOS 8 has begun to flatten at the... [+]
The trend in the adoption rates of iOS 8 show that we may already be seeing this effect. Keep in mind that the critics consider that the current problems started back in the last release cycle. If we compare the adoption curve between iOS 7 and iOS 8 (see image above from Mixpanel), it is clear that iOS 8 adoption has flattened out at a much lower rate than iOS 7 had. Without doing a lot of math on the curves, we can observe that shift happening at around 70% after six weeks for iOS 7 vs. at around 40% after just two weeks for iOS 8.
OS X Mavericks vs. Yosemite, in terms of user adoption. Notice how the Yosemite curve is quite... [+]
For OS X, a different and potentially more troubling trend appears. Although the early adoption of Yosemite has been the largest ever for a Mac OS, that may not be the whole story. If you look at the Yosemite fever line itself since its release date (see image above from GoSquared), it is strikingly linear. The Mavericks line, in contrast, has several prominent spikes, indicating power law dynamics. Linear or normal distributions are common when instances are uncorrelated, whereas S-curve distributions suggest correlation. In layman terms, correlation could mean that, for instance, a person has a positive experience and their enthusiasm encourages other people to upgrade too. One action makes other actions more likely.
The premature flattening of the iOS 8 curve and the linear nature of the Yosemite curve are examples of the kinds of features one might look at in data flows to determine a trend. Machine Intelligence company Numenta has built many its algorithms on "anomaly detection" of this type. In both of these cases, the current data set is sparse compared to its predecessor. But this is always the case in real-time pattern recognition. Humans base most of their own judgments on such sparsity. For the majority of cases, these jumps to conclusion work out fine. Other times, though, there is a lag between a new stimulus and our recognition of it.
In Daniel Kahneman's book Thinking, Fast and Slow, he describes two aspects of our thinking process that he calls "system 1" and "system 2." System 1 is the part of us that instantly downloads a free app when see the "get" button. System 2 is the part of us that wonders about what kinds of data the app might access or what sorts of in-app purchases we might be compelled to make. The secret of most marketing, Apple's included, is to short-circuit or neutralize system 2. All of the good feelings we have about Apple come from system 1 and are validated by system 2.
On the other hand, system 2 is vigilant to threats. That is its evolutionary purpose. So if I detect a possible problem that I might encounter with the new iCloud Drive if I upgrade to Yosemite, it is system 2 that encourages me to wait. I think that the kinds of issues that the critics of Apple's software quality are pointing out are of the sort that will trigger the vigilance of system 2 for users.
If Apple makes a correction within the next cycle, it is likely that its customers will not place much weight on their buggy experience with the last iPhone or Mac OS X upgrade when they consider upgrading next. The real risk for the company is if its culture makes scaling its engineering staff difficult. It may not just be the shortage of the right kinds of talent (HackerRank can help with this!) but the need to restructure for its now-larger and more interconnected ecosystem. The promise of "continuity" in iOS and OS X creates a lot of engineering complexity. Now throw in Apple Watch, Apple Pay, HealthKit, HomeKit and AppleTV! You get the picture.
Will 2015 be the year of Android? Or will Apple tie it all up for us with a bow? Likely we won't find out until 2016.
