Re-framing the dichotomies: Open/Closed vs. Integrated/Fragmented

Google likes to market itself as “Open” in contrast to “Closed” alternatives.

Apple likes to market itself as “Integrated” in contrast to “Fragmented” alternatives.

These dichotomies are judgmental and meant to portray “us=good” vs. “them=evil”. Neither gets to the point of how the two companies are structuring their businesses relative to the mobile computing value chains and clouds the judgement of observers.

I propose using a more informative and less judgmental distinction. It’s a division defined clearly by Clayton Christensen in his classic “The Innovator’s Solution” from 2003.

He introduced the concept of interdependent vs modular systems.

Interdependent systems have linkages between elements ranging from product components to the members of a product value chain. Essentially, the pieces communicate or interface via non-public or non-licensed protocols and have only specific partners (e.g. the iPod and iTunes).

Modular product architectures, on the other hand, are similar to the concept of plug-and-play and resemble building blocks, consisting of standardized interfaces, enabling parts to be easily swapped in and out and multiple suppliers or partners to compete.

Neither of these architectures is “good” or “bad”. They are natural ways to organize according to what business model or point in the evolution of a technology you are at.

Interdependent architectures are the only way to rapidly and competitively improve the performance of not-good-enough systems. Modular architectures are the only way to squeeze profitability from more-than-good-enough commoditized technology.

Interdependent systems often lead to higher performance than modular systems because the system is optimized. Engineers can iterate more rapidly to squeeze the maximum performance from sub-systems by connecting them in ways that makes the best complete solution. In contrast, modular systems often lead to lower costs because there are a greater number of suppliers, since interfaces are standardized and pieces can be swapped in and out.

Interdependency is often required to raise the performance of a new solution. In contrast, modularity is required to lower the cost of an overserving solution.

Continued interdependency often drives the performance of that solution beyond that which the target market is willing to pay for. The target market then frequently turns to modular solutions which, at that point, often offer good enough performance along with modularity-driven advantages such as lower cost, convenience, or other benefits.

So the question is not whether one is better than the other but when either one should be used.

Modular architectures are often the solution when interdependent architectures prove too costly. Interdependent architectures are the solution when modular architectures are not good enough or can’t solve new jobs that users discover they need to solve.

Apple’s value systems, priorities and processes are all tuned to interdependent architectures. Google’s value systems, business models and competencies are all tuned to modular architectures.

Apple solves the problems of new markets, Google solve the problems of over-served markets.

So the winning[1] strategy depends on detecting where a product lies in its march up the performance trajectory. Before it’s good enough, interdependent systems win. After it’s good enough, modular architecture wins.

So where is the smartphone on its evolution? If it’s still early days, iPhone will keep growing. If it’s saturating and improvements are neither valued nor used, then it’s time for a modular approach.

Place your bets now according to your vision.

By winning, I mean capturing the bulk of the profits in the industry.

  • dave

    Thanks for the lesson.

    I guess the question is how much form/function innovation is left in this space? Apple kills in design. Even though the iPod is old, they still revamp the design. This is to keep it 'fresh'. Can they/will they do the same with the phone? Apple is the leader, but they cannot lead too far ahead of the consumer.

    Apple is still the gold standard of phones and mp3 players. When they no longer are in the pole position they will be in trouble.

  • asymco

    Great description, and from 2008! This dichotomy has been at the root of Microsoft's entry into mobile phones back in 2001. When Microsoft entered into the smartphone OS market, their strategy was the same as Google's. They made an implicit bet that commoditization of mobile phones would occur well before the iPhone was introduced. It has not happened yet. Google then implicitly bet that commoditization of mobile computing would occur well before the iPad was introduced.

  • Mark Hernandez

    Excellent refinement and clarification of the distinctions, which is easily understandable by tech readers.

    Of course, it's a little too complicated for those who are more on the consumer end of the reader spectrum, so many of whom believe they're tech-savvy, but aren't really. I can see why the two giants chose the easier-to-grasp "sound bites" they did.

    As humans, our world is very complex, and it's a coping mechanism to simplify things, often reducing things a little too far down to a binary, black & white, "us vs. them" kinda thing.

    But it will be interesting to watch how Windows Phone 7 and possibly a resurgent Palm Pre will disrupt, complicate and partially dissolve the current "Google vs. Apple" war that was so simple to connect with for the masses.

  • Jon T

    But what if 'interdependent' captures a large slice of the limited components supply market? And, let's say, like the iPad, manages to make it available at a price well below what the modular market can achieve?

    • asymco

      Any commodity can be "cornered" and advantages can be obtained from its manipulation. However, by its nature, in a commodity market new suppliers can easily enter to meet demand. That's the theory at least.

      • addicted

        If startup (or capital costs) are high, that need not be true. It gets even more complicated when there are multiple resources being cornered, that are both necessary.

        e.g. if both Flash memory, and displays are cornered, then it becomes even harder for an alternative to step in, because you have the classic problem of who goes first.

        But more importantly, in this market, the things being cornered are not commodities, but sophisticated technologies that cannot be replicated because of varying reasons (patents, lack of know-how, lack of skilled workers, etc.)

  • Russell

    Great read with some clarity for my day. This post brings up another question regarding possible competitive advantages in this category going forward. Can an interdependent system (smartphones as example) create a high enough cost to discourage its customers from switching to another vendor's products? Since the underlying architecture is under it's control, isn't this the next step?

    Your charts are great in how they show that the money is all the the hardware at this point for Apple. I wonder over time if the value will start to seep down into more of the software for these devices and perhaps that's where the dynamics could change.

    Any thoughts on this?



    • asymco

      The dream of capturing value from system software (vs. hardware) has been around as long as smartphones have been the twinkle in the eye of platform vendors. Palm dreamed the dream and spun off PalmSource. Nokia had the dream and maintained Symbian as a publicly traded consortium rather than as an internal division. Microsoft dreamt of Windows for mobiles. They all mis-timed the bet. Now Google thinks it has the right timing.

      I take a point of view that systems software itself is not where value will accrue. Its role is purely to maintain the operation of the hardware so it's not really a job that needs a lot of improvement. The fact that even Google relies on commodity software (Linux, by definition, as open source, is a commodity) means they are not really seeking to derive value from it.

      If not from hardware, value will come from services. The difference in opinion is that Google thinks those services will come about through a browser and Apple thinks they will come from apps and the app store and Nokia thinks they will come from operator hosted co-branded client/server apps.

      • dave

        The difference with Google is they derive revenue from ad/search and do not seem to be monetizing the Android platform.

      • I don’t see how Google can continue pouring billions into Android without also finding better monetization strategies than their already-saturated ad business, which will take decades to break even. Look for NFC transactions, Google Travel, Google Wallet and many more. Sooner or later, shareholders will ask to be shown the money.

      • marko

        a trip down memory lane 😉

        aside from timing i think a key aspect of this is… how do these companies actually define success of these strategies internally? outwardly, they are always about winning a new market, but internally, incumbents are often intent on just removing value from adjacent segments to protect the core profit center.

        i think the market is often unable to recognize offense vs defense and evaluate accordingly. and thats not even getting into how individual interests and politics affect decision making… 'companies' don't make decisions, 'people' make decisions.

      • Joe_Winfield_IL

        Ironically, Google's own apps are actually better than Apple's in some key categories. These are a source of increasing value on the Android platform, and are a cause of concern for Apple's brand. Google maps, voice search, navigation are all superior to the competition, or at least are perceived to be. Fortunately for Apple, Google sees the iPhone as a major revenue opportunity, and has a disincentive to pull the apps from iOS.

        The two companies are codependent, with Apple needing Google services on the iPhone (at least until the Poly9 acquisition is fully fleshed out), and Google needing to be the default search provider for iOS. This codependency throws a wrinkle into the dichotomy arguments. The same phenomenon is visible between Apple and Microsoft, with iTunes for PC creating huge iPod sales and Office for Mac a good revenue stream for Microsoft. In some ways, I think this drives the two companies toward a natural duopoly in the smartphone sector.

  • Iphoned

    How does theory hold wHen applied to Apple vs Microsoft approaches in the early day of PCs?

    • asymco

      The PC reached a state of standardization rather quickly. IBM built the PC from modular components because they felt that such a computer would be good enough. It turned out that they were right, however the design became easily cloneable. The cloning opened the door to system software to becoming the orchestrator of the ecosystem. You have to also realize that the IBM PC was meant to compete with minicomputers (DEC VAX) and with mainframes. In comparison it was a very modular alternative that solved computing problems in a good enough way. Apple's integrated approach did not work as its Mac computers were more than good enough for the typical tasks people used PCs for (Excel, Word processing). In many ways, it was also a problem of timing.

      • dave

        I disagree here.

        In my opinion the IBM PC was NOT meant to compete with minis and mainframes. Both businesses were big profit centers for IBM. IBM had a history of stunting products that impinged on other business units profits. The Unix offerings IBM RT & RS6000 were prime examples where they were not pushed in the ways they should have been so as not to cannibalize the mini computer offerings.

        The PC was a quick response to their customers surreptitiously bringing in PC's because the centralized computing systems were not responsive enough. This was IBM's quick response to get in on a growing threat to their computing hegemony.

        In my opinion the problem with the MAC was more complicated. The first and biggest problem was cost. It cost as much as a car. The second problem is they were too sensitive to perception. They killed the gaming potential of the machine because it was derided as a toy and only just for games. I agree that timing may also have been an issue, but cost was by far the biggest issue.

        Cost is not an issue with Apple's current offerings.

  • OpenMind

    Great article. However, the dichotomy described between interdependent and modular are too clean cut. Apple's integrated approach has many modular components in it: processor (Intel and ARM), memory (flash), screen (capacity sensor), etc. Google's modular approach has interdependent components in it too: Android market place, google search, etc. I think either approach works fine regardless of market maturity stages. The important question is value/price perception to user. If a user think he is getting better value/price ratio, he will pay more for an item. So far Apple provides a very attractive value/price perception to customers. For Apple, question is how long can they maintain such value/price perception?

    • OpenMind

      Continue to rant. For PC industry, value of PC is pretty much fixed in last 5 to 10 years with minor twist such as processor speed, memory size, etc. Most of PC competition is to lower price to improve value/price ratio, ie. race to the bottom syndrome. It is a epic failure of Microsoft to deliver better value for PC system. Apple can compete with Android with both front: better value, equal price. Look at the street price of iPhone and Android, iPhone is not much pricer than Android phone, especially after their purchase price is amortized over 2 years contract time. And look at the iPad and other tablets, Apple even has better price point in some case (vs. Sam Galaxy). Apple failure in early 80s and 90s is MAC failed to provide double value than PC, yet command almost double of price.

    • asymco

      There are modular pieces in Apple's architecture and integration in Google's but the basic difference is whether major components like the Operating System is licensed or not. Google does not make hardware (Nexus One was an HTC device that Google marketed–very badly). Apple is not a licensor and rarely a licensee of technology. These are very different approaches to the market. Not only that but it's very hard (I would almost say impossible) for a company that depends on inter-dependence to switch to modular and vice versa. The resources, processes and profit algorithms are completely different.

  • kipek

    So Nokia started with interdependent mobile phones years ago with snake games and Navi key and all, and then started gaining market share and getting big and started cut costs going for modular with it's open Symbian and that lead to not-good-enough smartphones that users got frustrated with?

    At least in Finland a lot of people still think their old Nokia smartphones are good enough, and they even try till the very end argue they are actually the most advanced smartphones on the market. Even when official Nokia admitted at summer that they had not the best smartphones in their hands these Finnish Nokia engineer-kind of people insisted something like N900 was and still is the greatest achievement in smartphones.

    • asymco

      Nokia started with integrated (and still is that way with its Series 40 because they need to drive costs down and being modular in the low end will not work). They got the fever for licensing from watching Microsoft and also from thinking the device world will follow the PC world which led them to modularize smartphones early. As a result the product development process slowed to a crawl as engineers spent most of the time coordinating and not pushing what could be done. That opened the door to entrants to sweep them aside. What Nokia touts as achievement is simply individual module specs not the whole product.

    • If what you want is a portable Linux computer, not just a smartphone, the N900 is easily the coolest thing going. It has no peer. It tramples all over the iPhone and Android and has some great UI ideas and great architectural ideas.

      That makes it such a pity for N900 users that Nokia have moved on to MeeGo and Qt and largely left it in limbo.

      I had one for a few days and I was both very impressed and very frustrated with it. For me, having pretty much a full Firefox browser, shell access and repos made it the perfect portable computer. Perfect VoIP integration to every aspect of the OS – also perfect. I'm a Linux sysadmin and work from anywhere so ssh and VoIP are my main things in a 'phone' I need.

      The phone interface, size, poor keyboard, poor battery life, resistive screen and lack of software support from Nokia (I use Ovi Maps a lot) though was frustrating as an 'only one device' device.

      I'm still tempted with one on eBay every now and again as they're heading to about £250 now.

  • You said place your bets, so here's a prediction to get me off the fence.

    I believe that Apple's interdependent system will continue winning for the next few years.

    Why? I believe Apple's strategy (and you can bet they already have it planned out) is to constantly create new markets by using the success of the current interdependent products to educate users. They are then more able to accept their latest products in new markets when they announce them.

    Apple are all about innovating towards a completely interconnected system of devices that help make users lives better and more enjoyable. You can see this as far back as you'd care to look, but I'll cite examples such as the iMac as a 'digital hub' and iLife the software to enable it.

    Google's approach is different in that they want to be everywhere so they can display Ads to you.

    We can see this again more clearly with the iPad where users have been educated to understand touch and now they are taking to the iPad quickly where tablets earlier failed.

    Apple's vision relies on innovating into new markets, and it's clear that with the above vision they have plenty of room to move.

    As my final prediction I believe we are about to see (starting next month) the beginnings of Apple's expansion into the TV space. The terrible modular attempts like Sony Google TV, is, as Horace said 'bad for creating new markets'. Instead Apple are beginning to educate users with AirPlay and the interconnected approach with Apple TV, before moving to create the next new market.

    For my money, I believe the iPhone/iPad/iPod Touch will become the device that displays the interface for choosing content, while every monitor or TV or display that Apple enable with AirPlay will become the device to watch things on if there is one to hand.

    • r00fus

      Apple's margin relies on being at the forefront of the wave of new innovative solutions. The interdependent model, executed properly is hard to beat in this space.

      Apple2 -> Mac -> Laserwriter -> Powerbook -> iMac -> iLife -> iTunes -> OSX -> iPod -> iTMS -> iPhone -> iPod Touch -> iOS -> iPad -> Apple TV -> iPod Nano -> FaceTime -> Macbook Air -> Mac App Store ….

      All of these are innovative in their own right, but depend on previous innovative (or work done by previous efforts) to drive the brand, company and solutions forward.

      Apple's interdependent wins for cutting edge tech and that's where the money is, and will continue to be.
      The only reason other similar systems didn't work is that they didn't innovate/iterate fast enough or didn't execute well enough (almost all the items I mentioned above were successful products/solutions).

      Apple, unlike Microsoft, didn't get too greedy – they leave some value on the table as it engenders goodwill and provides opportunity for future efforts. Examples include adherence to web standards and participation with open-source projects (which are critical to the underpinnings of Apple products).

  • berult

    Innovation; patent protection and enforcement; financial, human resources fine tune procurement; consumer expectations subtle management; and rigorous consistency across the board. Never cease to nurture the "Feel good" effect, for there lies a successful integration of a philosophy into a moody market place.

    Apple neither sells causes nor effects; Jobs sells Causality, …to the highest bidder!
    Google sells either causes or effects; the highest bids get the causes and the lowest get the after effects!

    One would be tempted to draw a parallel with the development of actually. Which one of these two models would it fit well into I might ask? Excellence and authenticity just cannot bear straying away from originally laid out intents and purposes, lest it banked on financing due time irrelevancy.    

  • The author's figure 3.3 reflects a somewhat fantasy world in which all participants in a modular commodity device make lots of money. While conceding that the interdependent approach to device development yields a superior product for the consumer, he loses that focus, and turns to his fantasy. Only Google and the carriers will bank from Android.

  • I'm curious to know which system you think Apple's Mac line falls into? Is it modular because it's built using commodity components, or interdependent since the value of a Mac is the integration of the hardware and software into a seamless device? The Mac is sold in a modular market, competing against modular products, but still looks and behaves like a interdependent solution. Since the PC market is an established market, and most deem the current performance of PCs "good enough", wouldn't an integrated approach fail because of cost? Apple's trend with the Mac indicates otherwise. Does this mean the PC market is not really modular, or that it is, but markets can reverse them selves and swing back and forth between the two systems?

    • asymco

      The mac is a more integrated Windows PC. It's not a fully integrated device as the original Mac or Apple IIe was, but it's far more than a Windows PC. And a Windows PC is far more integrated than an Linux PC.

  • Joe_Winfield_IL

    In a completely unrelated note, here is the Steve Ballmer quote of the week:

    "I get all kinds of questions about 'what if you don't do this or that,' or blah, blah, blah. BOOM, baby, that's what we're going to do!"

    This guy is a terrible sound-bite machine. He was addressing developers' concerns about the viability of WP7 and its lack of market share and current low app count.

  • Another perfect article. Does the iPod product break this model? It would seem that iPods are interdependent too, with no loss of ground to the modular model despite attempts by Amazon and others. Yet it's an older product line and business.

  • Steven

    While this makes perfect sense and great reading for those of us who care to engage in a more in-depth discussion, it's a bit much for the dumbed-down debates which the general public is subjected to by the hit trolls. Unfortunately, we will continue to be bombarded by the simplistic dichotomies you seek to "re-frame."

  • Steve Weller

    There’s an asymmetry at work as well. It is possible, but difficult, to go from integrated to modular and remain profitable, but nigh impossible the other way. Once the choice is made, certain avenues are forever denied.