There’s a question I’ve never been able to answer about Clayton Christensen’s theory of disruption and Apple’s iPhone, which seems to be an anomaly to which the theory doesn’t apply. But recently, I’ve had some ideas that may explain why Christensen’s brilliantly successful theory fails to explain the iPhone’s continued success.
In a recent podcast, Benedict Evans, a partner at Andreessen Horowitz, mentioned the iPhone’s integrated architecture in a reference to two new features in iPhone 6s – 1) new 3D touch feature announced for coming iPhone 6s, which allows the phone to respond differently depending on how hard the user presses on the touchscreen, and 2) the way the screen lights up to act as a flash if the user is taking a selfie with the front-facing camera in the dark.
After discussing the difficult engineering involved in making these capabilities possible, Evans made the following comments: “that’s much more difficult for an Android OEM to do because Apple controls the operating system and all the apps as well … that’s just a structural advantage Apple has” (3:00). “it is about that integration as well, it’s about the making the whole thing, and it’s really hard to do that when you’re not in that position. (5:44). “So Apple made a chip to give you a better flash to take better selfies. Think about the kind of hardware/software integration that’s involved to do that. (6:08).
The iPhone has an integrated architecture
It’s common knowledge that Apple doesn’t make its own parts for the iPhone (for example, see this teardown report on an earlier model). However, the product’s architecture is still what Christensen would refer to as “integrated”, as opposed to “modular”. The critical difference is in whether the specifications for the parts and the interfaces between parts (including the operating system) are proprietary or broadly standardized. Imagine a phone that is designed to utilize a standard 4-inch touchscreen that could come from any number of manufacturers who produce this stock part. The 4-inch screens from these manufacturers are built with a standardized interface that allows their screens to work with any phone that adheres to the interface specifications. This is called a modular architecture – the product is designed in a way that allows any screen that meets the standard specifications to be plugged into the phone.
The iPhone is not like this – although Apple doesn’t make the components for the iPhone itself, it requires many of the parts to be manufactured to a proprietary specification, and works with its manufacturing partners to design those specifications specifically to give the iPhone the best possible performance.
It’s worth noting that the distinction between “integrated/proprietary” and “modular” is not always clean-cut. Phones that would be considered modular, on the one hand, may contain some parts that are specifically designed for that phone, while its operating system is an open standard that any hardware maker can build phones around and the bulk of its parts are commodity parts designed to that standard. Conversely, I believe that there are a handful of commodity components included in the iPhone’s build as well. The iPhone has a proprietary architecture, though, inasmuch as its operating system is a proprietary standard, and the key components in the phone are designed and built specifically to meet the specifications in that standard.
Disruption theory says integrated architectures should fail
In their book The Innovator’s Solution, Clayton Christensen and Michael E. Raynor describe the way market forces can shape a market’s value chain, determining the types of firms that can prosper in the market. I’ll give a [sort of] quick summary of their argument. Christensen and Raynor assert that as markets mature, the basis of competition changes.  A firm that was successful in obtaining customers and earning attractive profits in the past may find that it can no longer do so, and wonder what changed. That firm may feel that it has somehow lost its competitive edge, it may talk about refocusing on its “core competency”, or pursue any number of other misguided strategies to right its course, but if it doesn’t understand the change in the basis of competition in the market(s) in which it operates, none of these things will fix the problems.
Christensen & Raynor argue that in a market’s early days, the basis of competition is almost always functionality and reliability (which they refer to as the product’s “performance” – an abbreviation we’ll continue to use here). They offer the personal computer as a simple example.  In the early days of the personal computer market, functionality and reliability were inferior to customer expectations. Sitting at a computer and trying to do something was frustrating, because the computer was slow and hard to use. Notwithstanding its superior ability to do the things that it could do, a typical customer found the performance of the product to be “not good enough”.  Therefore, in this phase of the market’s maturity, the dominant firms were those who could make the computers with the best performance in functionality and reliability. This is the pattern of any emerging market space – the products are not yet good enough to meet the basic expectations of the users, and so, assuming all other parts of their business are managed well, those who can get the closest to performance expectations can obtain market share and demand attractive profits in the market.
How does a firm create a product with superior performance? Christensen & Raynor argue that, for a number of reasons, an integrated architecture is essential for a product to perform better than its competitors in these early days of a new market. By managing the entire end to end production of the product and all of its components and subsystems, a company can ensure that its product achieves the best possible performance. In the early days of the personal computer market, this integrated player was Apple Computer, which made computers that “were easier to use and crashed much less often” than computers made by firms with less integrated architectures for their products. 
Disruption theory says products with modular architectures should eventually overtake those with integrated architectures
The other choice, instead of having an integrated architecture, is to have a modular architecture – where the design and manufacturing of a product’s subsystems are outsourced to other firms who specialize in producing that one component. The subsystems from these different firms are then built into the final product. In the early market days when no product in the market has performance that is good enough, these firms cannot deliver a product that is competitive with the integrated product. Interfaces between subsystems are not yet standardized, and more importantly, the non-integrated product doesn’t have the integrated product’s ability to milk performance out of every subsystem through its proprietary architecture.
However, there comes a time when the market matures so that the functionality and reliability of the products are more than good enough to meet the expectations of customers – or in other words, there is a “performance surplus”. In this stage of the market, “customers are happy to accept [performance improvements], but they are unwilling to pay a premium price to get them”.  At this point, the basis of competition changes. Since any firm in the market can make a product with performance that is good enough to meet the expectations of customers, performance is no longer the essential thing that a firm must do to obtain customers and earn attractive profits.
Christensen and Raynor argue that, at this stage, the new basis for competition becomes the ability to give the customer what they want, when they want it – in other words, speed to market and the ability to customize the products for customers. Microsoft, Intel, Dell, Gateway, Word Perfect, Lotus, and others were the victors in this stage of the personal computer market. By developing a product with a modular architecture, these specialized firms were able to offer a greater spread of product configuration choices for customers, and could even offer computers built according to customer specifications. Upgrades could happen more quickly, because subsystems could be upgraded without necessitating change in other subsystems. In this stage of the market, Apple’s integrated architecture meant slower upgrades and fewer choices. It “became a competitive liability”. 
This shift in the basis of competition can happen more than once. It can shift from functionality and reliability (i.e. “performance”) to speed to market and customization, and then back to functionality and reliability, or to something else. With these thoughts in mind, let’s look now at the iPhone.
The iPhone keeps winning with an integrated architecture
It would appear that the iPhone has exposed a gap in Christensen’s theory – something that the theory does not completely explain. Christensen himself felt that the iPhone’s integrated architecture would lead to its marginalization in the market.  Some commentators latch onto this statement as an error, considering the iPhones continued visibility and its wildly successful contribution to Apple’s bottom line.
However, these commentators seem to forget that in some ways, Christensen is right. As of Q2 2015, research from IDC concludes that the iOS operating system (which only runs in the iPhone) owns about 14% of the worldwide market, while Android owns about 83%. The market is strongly consolidated around these two players.  Interestingly, if you look at the numbers by phone manufacturer instead of operating system, Apple’s share floats around in the same range, between 10 and 20%. From this phone manufacturer share chart, the market appears more fragmented – probably because the Android operating system’s modular architecture enables many phone makers to enter the market. Looking at the data in this way, the conclusion of those commentators that the iPhone is untouched by the market forces around the shift from integrated to modular architectures appears to be inaccurate.
However, if we take the point of view that the iPhone’s incredible profitability means that it has not failed in the face of modularization from its competitors, then the question remains. Why hasn’t the iPhone been crippled by the modularized competitors who can move faster and offer a greater range of options? This is a very valid question given the iPhone’s long and continuing status as the premier smartphone on the market, and its ability to command huge margins. However, we should remember that we have also allowed a subtle compromise in our thinking – because it could be argued that the iPhone would be much more successful today had its competitors with their modular architectures not arisen to take so much market share for themselves.
Explaining the iPhone’s longevity
Remember, the reason an integrated architecture is successful in the beginning of a market is because the performance of the products in that market is not good enough. In like manner, the reason the integrated architecture loses market share and ceases to command attractive margins is because the performance of the products in the market all become good enough, at which point the buyer’s willingness to pay a premium for the unnecessary performance surplus of the integrated player goes away.
So why does the iPhone continue to command outsized margins and maintain a steady market share when, to a casual observer, it would appear that performance from many players in the market is good enough, or even equivalent with the iPhone? Commentators have offered different explanations. Ben Thompson, author of the Stratechery blog, for example, claims that disruption theory doesn’t apply to consumer markets.  However, despite the Thompson’s strong grasp of the theory, his conclusion is somewhat broad, and ignores the fact that certain of the cases upon which Christensen’s theory is formed come from the consumer market.
So what can explain the iPhone’s exceptional status to a theory that has proven to be correct so many other times? I will offer several potential explanations, which are not mutually exclusive. Reality may lie somewhere between two or more of these explanations. Each is a potential avenue for further thought and analysis. I may find time personally to explore some of these avenues in future articles – I invite others to do so as well.
I should note that, although I ultimately reject Ben Thompson’s conclusion, I do think he is right that the consumer market may require a slightly different approach. In the smartphone market, in any case, good enough is a concept that operates on a host of dimensions – not just technical performance. My thoughts on this will become evident shortly.
Performance of products in the market is not yet good enough
We should keep in mind the fact that good enough is a relative term. What does it mean, and who decides where that line should be drawn? Each of the following explanations is actually, to one degree or another, a variation on this theme – because the definition of good enough is at the crux of the theory that says a market turns toward modularization when some critical mass of products achieve good enough status. It’s possible that the smartphone market is not yet at the stage where performance is good enough. I offer this here more as an introduction to the following explanations than an explanation that can stand on its own. By itself, this explanation lacks the necessary nuance to allow one to draw any meaningful insights or conclusions.
There are market segments for whom only the iPhone is good enough
It’s very possible that our error lies in our attempt to treat the entire smartphone market as one unit of analysis. There are almost certainly a numerous host of distinct market segments that may deserve to be treated independently. Perhaps by analyzing them independently, it would become surprisingly clear as to how the theory of modularization is operating in each segment.
Eric K. Clemons and Rick Spitler, in an unpublished draft of a book I studied in business school, argue that there is a change in the basis of competition in the consumer market that will require market players to differentiate more specifically for targeted segments of customers, rather than attempting to address a larger, less specific market segment at scale: “…success today and in the future will increasingly be related to skill-based competition, knowing what to offer, and to whom, and knowing what to charge, and to whom. That is, hyperdifferentiation is part of the transition in strategy, away from seeking advantage and higher margins solely through scale and lower expenses.” 
Clemons and Spitler argue that, with this new basis of competition, companies must:
[move] from a focus on large market segments with huge numbers of customers to a focus on numerous market niches with satisfied and profitable customers. The move to sweet spots requires better understanding of the customer. It also requires better understanding of competitors’ offerings and carefully staying away from already cluttered product “regions” in which there are numerous near substitute products already available. By avoiding cluttered areas with numerous product offerings – the firm is able to avoid the paradox of extremely differentiated products that are similar to other extremely differentiated offerings, and thus to avoid the resulting price wars that this could create. 
Clemons and Spitler suggest that a strong brand will be less critical in the transformed market they describe, and also argue that modular product design will be critical to simultaneously addressing the niche markets successful companies will target. While these aspects of their theory do not contribute to understanding the longevity of the iPhone’s success, the idea that companies should focus on “numerous market niches with satisfied and profitable customers” does seem to describe the iPhone’s history rather well. Think of the customers who would never use any other smartphone, and who line up outside stores on launch day. This is not every customer in the market – but it’s a steady segment of the market. As the IDC report shows, the iPhone market share viewed by both OS and phone manufacturer has hovered between 10 and 20% for the last 3 years, with some of that fluctuation appearing to result from variations in quarterly buying trends. This group of iPhone buyers has proven to be unswerving in their loyalty.
I just posed the question “what does good enough mean, and who decides where that line should be drawn?” Realistically, every buyer makes the decision of what is good enough for himself/herself, and it seems obvious that for a very stubborn market segment, only the iPhone is good enough. While the theory from Clemons and Spitler might suggest that Christensen’s disruption theory could benefit from more nuance in the definition of what is good enough based on customer segmentation, it does not fundamentally challenge disruption theory.
With the rapid pace of change and creation, the line for good enough performance doesn’t stay put
While each of my other suggested explanations operate with a very broad definition of “good enough performance”, including the phone’s performance on dimensions such as display quality, processor speed, network speed, physical attractiveness, social prestige, etc, this explanation focuses more on the hardware’s computing performance.
One could argue that the definition of good enough performance is almost impossible to nail down in a market where change is so constant. The requirements for good enough may increase with a new app that’s heavy on resource requirements (and it’s interesting to consider whether this is what Apple is intentionally doing with the 3D touch and selfie flash features mentioned at the beginning of this article). In other words, it may be more difficult to make a good enough phone when this happens, resulting in a smaller percentage of phones on the market that are good enough. The requirements may be relaxed when some platform improvement changes the way things get done or lowers the resource requirements for apps to run or for certain operations to be performed.
With a line for good enough that moves around so frequently, the tendency for the market to shift toward modular architectures may be hindered. And with moves such as the introduction of 3D touch and selfie flash, Apple may actually be turning the sword in the opposite direction – if it can introduce features such as these that are successful enough that the bar for good enough is effectively raised for some significant portion of the market, modular providers will actually struggle to surpass that bar, as they have to work together at a slower pace to introduce the features in their own phones.
The complexity of the product defies any good enough analysis
What is a smartphone? It’s simultaneously a phone, a computer, an app store, a personal assistant, a GPS device, a fashion statement, and more – it’s each of these things, and it’s all of these things. Can any product ever be good enough on all of these dimensions individually, and good enough on all of these dimensions holistically? It’s possible that no working definition of good enough is achievable with our current tools and understanding.
In a market where “performance” is an open-ended set of possibilities, perhaps the description good enough may not apply
The smartphone market is characterized by frequent product launches, each pushing the envelope as far as possible. While the improvements in recent time have become less radical and more incremental, the attitude of consumers toward phone launches (who ultimately are the only ones who can determine where the bar for good enough is located) is still one of anticipation. When expectations for infinite improvement underlie the market’s relationship with the product, including expectations for innovation never before seen, perhaps the usage of the measure good enough should be re-evaluated for use in this market.
My attempt in this article has been to explain why Clayton Christensen’s remarkable theory of disruptive innovation has, to this point, failed to completely describe the performance of the iPhone in the smartphone market. I hope this analysis is helpful. Comments are welcome.
1. Basis of Competition: the thing that a company must do better than competitors to obtain and retain customers and earn attractive profits in a given market. Put another way, it is the thing for which customers will pay a premium and offer their loyalty.
2. Please note that this simplistic example of the personal computer market is not offered as proof that the concepts put forward here are valid, but as an example to aid the reader’s understanding of the concepts. The research done by Christensen and Raynor applies broadly and stems from broad research.
7. Critical Path podcast, episode 36, May 2, 2012. 42:55.
8. see IDC report
9. see blog post September 22, 2013
Update 9/29 – minor corrections to grammar and clarifications made to arguments.