Preempting the Praetorian Guard

The functional organizational structure may be causal to successful disruption. Although this hypothesis needs much more research and testing, there are correlations in the anecdotal evidence. One of the principles of this theory is that functional organizations are less prone to “political infighting” where a powerful division can disable a self-disruption through the wielding of the power of the Profit/Loss statement.

Under the divisional structure the modus operandi is that once an internal threat is detected, the incumbent division unleashes antibodies to destroy the smaller and less powerful opponent. This is possible because as soon as the threat is perceived, a threat response kicks in and the power to deflect the threat is abundant in the reward and incentive structure. This response mechanism is so common and well understood that it forms the basis of the Innovator’s Dilemma.

My hypothesis is that in the absence of product/divisional level power bases, the threat is not felt and the political power to respond is not available. In a functional organization there is no “business leadership”, where the P/L “belongs” to one person. Only the CEO has life-and-death power over products and their decision is “purchased” through discussion with functional heads who stand to benefit as much as to lose from disruptive change.

This functional organization decision process is typical if not universal in small companies and atypical if not unheard of in large companies. It’s one of the reasons small companies are inherently more disruptive. The challenge for a growing company is that functional organizations don’t scale well. There are very few and all depend on vigilant, almost maniacal defense of the structure. Divisional organizations are “natural” at scale and the preservation of functional structures is like fighting entropy.

The organizations which seem to preserve functionality at scale are militaries. This is why I’ve been thinking about it takes to manage such organizations and preserve their apolitical status.

As it turns out, militaries are also not immune from the corrosive effects of political power. In fact, the slide into corruption begins with the formation of elite divisions. The rationale is simple: elite soldiers whose loyalty and skill are most valuable are chosen for special roles and status. They begin by usually acting as bodyguards to the leadership or the state.

Historically the most prominent such politicized military force was the Roman Praetorian Guard. Having started as bodyguards to the field commander, they transitioned into protectors of the state and, eventually, into destroyers of the state.  Their leader, the Praetorian Prefect, became one of the most powerful administrators of the state and his Guard became responsible for numerous coups d’état, assassinations and corruption.

So-called “Republican Guards” became common in many militaries since then. Napoleon’s Imperial Guard, the SS in Nazi Germany, the Iraqi and today’s Iranian Republican Guards continue in a political role. In democracies, the politicized military is less common but the threat of their emergence is one states must always remain vigilant for.

And so it must be for the functional organization. The emergence of an elite which, by its skill and achievement, gains more responsibility and hence more power is the good intention which paves the road to hell. Leading an innovative organization means having a process of denying power to those who are elites. It means preempting the Praetorian Guards.

Doing so is a dilemma in itself as the elite are not likely to remain loyal if denied their opportunity to power. Managing the motivations of the bright and ambitious is where the art of disruption management lies.

For a more thorough discussion of this topic, listen to the podcast which accompanies the post:  The Critical Path #90: The Praetorian Guard


  • obarthelemy

    I broadly disagree with everything in that article:
    – anecdote is not the singular of data. I know your darling Apple is functional , but of late their innovation record is mostly blank, while divisional companies have kept on innovating.
    – correlation does not imply causation. Small companies are more disruptive. Small companies don’t have divisions (well, they mostly have 0 or 1 product/family). The jump to “small companies are functional and because of that are disruptive” is huge.
    – infighting is not inherently better handled by functional entities. That’s more a factor of CEO strength and company spirit than of org chart.One could argue that it’s easier to isolate internal skunk-works in a divisional company: IBM PC for example, or Xerox PARC…

    Essentially, you’re saying that monolithic states are more efficient than federal states. Which is anecdotically confirmed by: Not the US, not China, not Germany, a bit the UK, ah, my dear France, also Russia. Pick the successful countries in that lot.

    • The notions covered here are expanded on greatly on my latest podcast, not yet published. The crucial distinction is between disruptive innovation and sustaining innovation. The question of causality of disruption (including self-disruption) is discussed in detail on the podcast.

      I think your take on efficiency and the state is the opposite of what I’m suggesting. Efficiency is the antithesis of effective innovation.

    • I think that IBM and the PC coul also be examples of lack of innovation in a divizional organization. The PC is a one and done for IBM. All that the company learned from the PC has been apparently lost, because the PC experiment has never been replicated successfuly inside IBM.

    • I agree. I think leaders are more important than organization.

      • handleym

        Azure is irrelevant to anything here. Azure is not a threat to Office or Windows (or even Box).
        Substantially more interesting are

        – .Net, which never took off the way it was hoped by some throughout the 2000s

        – Singularity, a research OS described by MS in early 2007 which looks very much like iOS — and which remains a research OS.

        – The constant infighting between the old-guard and the Win8 folks within MS, evidence of which we can see, for example, in the surly and unenthusiastic “support” for Win8 within the newest MS Office.

    • Kizedek

      – infighting is not inherently better handled by functional entities. That’s more a factor of CEO strength and company spirit than of org chart.One could argue that it’s easier to isolate internal skunk-works in a divisional company: IBM PC for example, or Xerox PARC…

      I think the in-fighting is a result of duplicating the functions, with each division/product group having their own instance of similar functions (design, software, hardware). If these are not co-ordinated under one Function Head (ie, SVP of Design, SVP of Software, etc.), then there is competition. Competition equals “infighting”. Then you get the Kin Team competing with the Windows Phone Team or the Zune Team. The Kin Team got killed off despite its evident potential and talent from outside (Danger); the MS anti-bodies were too potent.

      If there is Functional focus at a company, rather than divisional or product, then the company can worry about making the best phone product, period; because a victory for one is a victory for all.

      One could argue that the skunkworks at big divisional companies get too isolated and lost. They become ivory towers. R&D spending becomes a huge black-hole with duplicated efforts (Apple spends relatively little on R&D, managing it far more efficiently). The Skunkworks at Xerox PARC got isolated and lost, because Xerox didn’t know they had a good thing when they had it… Apple asked for access, took one look and negotiated a deal to build on what they saw there because they saw the potential for a consumer product.

      In contrast, Apple skunkworks is about all the senior executives sitting down with each other every week and asking each other what they are doing. Apple acts like a start-up. So, yes, there is causation; it is not the leap you propose: Apple can focus on exactly what will disrupt the status quo, precisely because they are NOT focused on maintaining the status quo as a divisional- or product-oriented company would be. That’s the whole point!

    • DesDizzy

      OB as usual you miss the point completely. Try to forget your bias anti-Apple for a second and think logically (this is a French tradition?). Countries have nothing to do with this argument and have not been proposed as a proxy by anyone.

      And suggesting that Apple’s innovation record is mostly blank might be acceptable for mindless analyst and reporters, but is not acceptable within a serious discussion for supposedly more thoughtful analytical minds.

      The evidence of other tech co.s struggling to catch-up with Apple is out there. Amazon, Google, MSF, Intel all having been disrupted and forced into changing business models into hybrid structures (which put them into competition with their clients), to try to compete. Never mind those organisations that have been disrupted into a death spiral, such as Nokia and Blackberry.

      • obarthelemy

        Yes, countries have been proposed as a comparison by me. “divisionnal” companies can be likened to “federalist” states, hence my comparison.

        As for innovation, again, Apple innovated in making electronics gizmos easy to use and socially desirable. That culminated with the iPhone 1 years ago. What innovations do you credit them with since then ?

      • alcatholic

        Disruptive or sustaining innovations?

        Plenty of the former.

        Assuming you don’t accept any post 2007 Apple work as disruptive, I ask you what other company has produced disruptive products post 2007?

      • privacy

        Apple has brought us personal computers, iPod, iPhone, and iPad.

        In comparison, Google hasn’t innovated since 1996 when their search project started at Stanford. Unless you count robbing Overture as innovation. Google’s innovation has stagnated so badly, that Bing is now beating them in blind comparison tests. Bing’s marketshare has reached 17% and is growing fast.

      • obarthelemy

        Exactly: not much since the iPad, and arguable that was just combining a preexisting form factor with a preexisting OS.

      • Kizedek

        LOL, that’s like saying McDonald’s “just” combined pre-existing bread-rolls with pre-existing minced beef.

        Of course, what McDonalds actually did was create a business that no-one in the history of the world dreamed of; the first of its kind, and one that has been emulated ever since by every other company wishing to grow into an international “chain”.

      • GiveUsABreak

        “Exactly: not much since the iPad, and arguable that was just combining a preexisting form factor with a preexisting OS.”

        “since the iPad”…people who make those comments are like little kids, for whom a year seems like such a ginormous length of time. Get real.

        And what has anyone else invented in the past two decades that was so amazing? (Please don’t talk about Samsung making 100 different size and shape of phone an innovation, when you don’t even see the iPad as a separate innovation than the iPhone).

      • OpenYourEyes

        “As for innovation, again, Apple innovated in making electronics gizmos easy to use and socially desirable.”

        I’d say there’s a bit more to Apple’s special recipe than simply 1) “making gizmos easy to use” and 2) “socially desirable”….although that would already put them ahead of the competition. How about the most robust yet also most elegant physical design? How about the App Store? How about iTunes; I think that qualifies as an innovation. How about iCloud? Sure, it has a long way to go in improvements, but the fact is there is no one else that can seamlessly connect a personal ecosystem of a smartphone, tablet, desktop, and TV-accessory.

        As far as devices, there is far more than iPhone. There’s the outrageously sweet iMac, the stunning new Mac Pro, the game-changing laptop of MacBook Air, there’s the Apple TV (which has been crazy successful for a hobby, and will explode when they open up an App Store for it)…and oh, yeah, the iPad and iPad Mini, which are fundamentally changing both the way people interact with their computers, and radically altering many endeavors, such as education.

    • NoNeedForSnarky

      “I broadly disagree with everything in that article”


      “-anecdote is not the singular of data.”

      No need to be rude to Horace. He is one of a rare breed of sophisticated thinkers on the dynamics of Apple’s industry. He has made it clear this theory is a work in progress, and he is investigating history and other firms to try to get more data.

      ” I know your darling Apple is functional , but of late their innovation record is mostly blank, while divisional companies have kept on innovating.”

      Again, no need to be snarky. There are a hundred and one Apple discussion boards for that.

      But as to innovation, what are the “innovations” of the divisional companies making smartphones and tablets and–especially–personal computers?

  • John R. Moran

    I think you’re on to something here, Horace.

    Still a bit confused about the analogy between an elite, politicized military guard and the formation of business divisions in general.

    But disruption is based on motivation and incentives, which in turn are based on organizational structure. So it follows that different structures will respond differently to disruptive threats and opportunities.

    See this for a few more thoughts on how this applies to Apple: “One Strategy, One P&L”:

  • Horace, military organizations switch to divisional structure at war time. During WWII there was a head of the Pacific and a head of Europe, or a head of the North front and South front, etc. This reinforces your thesis that you deploy a divisional structure when you need efficiency.

    Interestingly enough, both Eisenhower and McArthur, two of the “divisional” heads during WWII entered politics after the war, although with different degrees of success.

    • Just got to the point in the Critical Path when you give Eisenhower and McArthur as examples 🙂

    • There is one more argument that military switches to divisional structure during war. You mentioned that in divisional organizations rotate jobs and this happened in all the armies during WWII: Generals like Montgomery, Rommel, Patton, Timoshenko were promoted or punished through rotation between various front lines. And in more recent times:Gen. Petraeus rotated jobs between Irak, Afghanistan and CIA.

      • Yes, but I refer specifically to rotation into new functions. The head of the Air Force won’t get rotated to head the Army (either as a promotion or demotion.) The degree to which armed forces begin to “compete” for missions is a crucial inter-service rivalry problem and is symptomatic of a functional organization. It’s one of the sources of inefficiency. But, I believe, it’s a price worth paying.

      • nt

        A COCOM commander is a 4 star general or admiral tasked with overall command of a region so in a sense they do get rotated into heading another branch. For example Adm Fallon was USCENTCOM commander in 2007 relieving Gen Abizaid (US Army). CENTCOM functionally is dominated by land operations, not naval even if the Navy plays a significant role.

        The intent of a unified command structure, at least at the top, is to try to minimize inter-service rivalries.

        That said, CENTCOM generally goes to the Army or Marines, and PACOM to the Navy because that makes the most sense.

        That isn’t just the wartime structure. It’s just the way it’s been since WWII in peacetime or wartime.

    • Good points. You might operate this way during a crisis but it’s important to dismantle the power structure after the crisis ends. Too often that does not happen.

  • Sina Motamedi

    This is a great post.

  • James King

    I have to disagree with this post vehemently.

    Based on my experience, the reason why larger organizations become less innovative is because of what I will frame as the Incentive Paradox. Simply put:

    As the small organization experiences more success and becomes more relevant, financial incentives become a more powerful force;

    The organization becomes polarized around the areas of the business that expand its power, influence and relevance. In other words, the aspects of the company that bring in the most money or generate the most influence become the most important;

    The people who run these areas of the organization become the most powerful;

    When these people face challenges from other innovators in the organization, they protect their turf vigorously;

    These same leaders choose lieutenants with passive personalities. Once again, the motivation is to eliminate competition;

    The dynamic creates an environment in which politics and in-fighting become more important than overall company success.

    The problem is that the process of smaller successful companies achieving great success seems to inherently creates stratification. The management layer closest to the founders and early investors ends up wielding an inordinate amount of power and prestige. These people may not truly be the “elite” of an organization when it comes to skills and talent but they are generally the FIRST. A single idea generated by one of them may contribute billions to the company’s bottom line. How does a newer person in an organization bring innovative ideas up in a company against the very people who are responsible for the company’s initial success?

    Another way to look at it is that companies in that situation suffer from indulging in an advanced form of logical fallacy: argumentum ad crumenam. In other words, the entrenched leadership is never wrong. Power and wealth end up protecting the leadership at the highest levels from their own mistakes and arrogance. Their initial success makes them believe in their own intrinsic “rightness” which actually shrouds a deep-seated desire to protect their own status and relevance. They respond negatively to any idea that could challenge their standing.

    I highly doubt that it is the “elite” of an organization who end up stifling innovation. To the contrary, it is a power structure that only moves in ONE direction that leads to an organization’s stagnation. Success is rarely rewared and failure is rarely punished. Leaders are rarely DEMOTED, especially at the highest levels of an organization. When a person reaches a particular level, especially the executive level, there are almost no dire consequences for failure. Blame is generally shunted down to the person with the least political cover. People at the highest levels of a large organization are generally compensated so well that even penalties in compensation are token and rarely are severe enough to create a change in lifestyle.

    It is the fact that the cream DOESN’T rise to the top that leads organizations to stagnation and failure. I think that if companies created power structures that went in BOTH directions, it would allow the most capable leaders to rise within the organizations. Company power structures should allow for DEMOTION is the face of failure, all the way up to the executive level. Greater fluidity in company leadership would properly incentive those in power to champion the best ideas for fear of being knocked down the chain. The dynamic would change from “leaders” who are only interested in receiving credit to those who will place the well-being of the organization first.

    It’s not a perfect solution. But keeping the best from rising in a company seems like an easy way to demoralize a company to its deepest levels.

    • Bruce_Mc

      “I think that if companies created power structures that went in BOTH directions, it would allow the most capable leaders to rise within the organizations.”

      Any examples of organizations that practice this to some degree?

      • James King

        I wish. In any case, this is an idea that needs to be flushed out. My main point is that failures of an organization are generally failures in leadership. Not only do you want your best leaders in place at all times, you need a mechanism in place that allows the next generation of leaders to step up when the previous one starts to become too conservative. I don’t think there is enough research and effort being done regarding these types of organzational dynamics.

  • Horace, for me this has been a though provoking post and Critical Path episode (a 4th comment can prove it).

    The evidence does not support your hypothesis, I am afraid. The rush of innovation around iOS happened mostly at a time when iOS was practically a division inside iOS. The anecdotal evidence suggests that Forstall was the czar of iOS answering only to Jobs and annoying everybody else with his stubbornness. iOS was not part of the “Software” function until recently. That allowed it to design the UI in ways that were radically different from the Mac UI. It looks like even at Apple the way to protect from incumbent antibodies is by quarantine. Now, when the iOS design paradigm has actually taken over the whole of Apple, maintaining the iOS division would have done more harm than good, so there is no iOS division anymore.

    I must say that, although I disagree with your conclusion, I enjoyed a lot reading and listening to your thoughts. I also think that studying the success and failures of military structures can teach a lot about how to manage large organizations:

    • I believe Forstall was fired precisely because he violated this core principle. He was Apple’s MacArthur.

      • Indeed, but in the same way as MacArthur he was useful for a while as division head. And the innovation did not stop while Apple switched temporarily from functional to a divisional structure. The key is not only in a certain organization structure, but also in the transition between one operational mode and another: you need to hit the right time window and have the right people in place. Eisenhower proved to be able to adapt to both war and peace, while MacArthur could not adapt to the post-war time.

    • handleym

      “The rush of innovation around iOS happened mostly at a time when iOS was practically a division inside the company.”

      Like obarthelemy, you’re not going to see anything interesting if you insist on being wilfully blind, for example claiming that the only innovations in iOS occurred in 2007.

      iOS today is still the only real OS thinking seriously about the consequences of Moore’s law for the next ten years or so. I’ve said many times that compute is becoming so cheap (and not just compute, but connect and now display) that it makes no sense to obsess over how to create a single compute device to fulfill all your needs — that’s like spending the 1980s thinking about how to create the optimal mainframe.

      With compute so cheap, the question is how best to use these new skills. iOS is thinking about this seriously, along the dimensions that actually matter — security (and how to balance it against usability), power management, UI on ever smaller devices. The competition still see the issue as one of replicating desktop ideas (about the security/usability tradeoff, about power management, about UI, about specialization vs generalization) on more and more devices.

      You don’t see innovation in iOS because you think innovation means some sort of variant on what was innovative 10 years ago.

      Look at iOS B2B market, there’s a whole world there that most of us are completely unaware of — a world which is trying (and to some extent succeeding) to do an end-run around MS’ control of enterprise software.

      Look at another unglamorous subject, Apple’s handling of subscriptions for iOS (and OSX) apps. This started off in the natural way as part of Newsstand, but it’s technology that could turn ultimately into a widespread micro-payments network. All the underlying tech part, most of the API, and much of the UI are already in place.

      Look at the support for low power bluetooth, something which is actually useful as opposed to the gimmick of NFC. This started to unleash health-type devices (like the Striiv Play), and we’re seeing the next step in the game controllers discussed at WWDC — like I said, a network of specialized devices all of which work well together.

      Look at the new semantic text sizing stuff. While MS and Android still struggle to get their retina support working well in actually existing software, Apple has moved on to the next stage of improving visibility for everyone whose eyes are getting worse.
      This is an example of real innovation. Simple-minded innovation is giving your browser a minimum font size spec — it’s a solution that works in only a few places, and which works badly (the “visual” size of a sans-serif font is substantially larger than for a serif font). REAL innovation is looking at the big picture, not “how do I make the fonts bigger” but “how do I give programmers a UNIFORM way to describe text which will allow the user to scale it (VISUALLY CORRECTLY) to the size that best meets their eyes?”

      And so it goes.
      Innovation is not adding a lousy checkbox feature to a device, which is barely supported for the life of that device. It’s not S-beam, or adding a lousy thermometer to the phone. Any monkey can do that.
      Real innovation is the hard stuff, the thinking about how best to add a collection of features so that they will work together well not just today but also over the next ten years.

      • How does your diatribe answer my point that Apple innovated both under a functional structure, as well as under a quasi-divisional structure?

      • InnovateOrDie

        “How does your diatribe answer my point that Apple innovated both under a functional structure, as well as under a quasi-divisional structure?”

        His wasn’t a diatribe; it was quite insightful.

        As to your own argument, it is simply false.

        First of all, in the initial stages, there were two completely different operating systems competing to be the OS for iPhone. Forstall’s just happened to win, on the merits. Not because he ran the “iPhone division”.

        Secondly, Forstall wasn’t in charge of the hardware, the marketing, the pricing, the supply chain surrounding iPhone. Various divisions all collaborated (under the command of only the CEO).

        Your idea that just because iOS was developed as a separate OS than OS X, the whole notion of a functional organization is false, simply is wrong. At most, you could argue that during the development of the iPhone and iPad, “Software” was split in two. Yet even there, we see functional division, not product division. There wasn’t an iPhone division, and a separate (and competing for victory inside the company) iPad division.

  • I have always sustained that business companies have all to learn from military organization and most military leader pass to private companies as a demonstration of the principle.

    Aviation passed from central maintenance to distributed local maintenance with local budgets before the gulf war and as result during the war the planes downtime was 100 hundred times smaller than before.

    You could say that they passed from functional to product organization because they divided a central work between multiple offices to gain more efficiency.

    The weakness of military organization in WWII was the separation between armies, for instance in Japan aviation and navy were in war between them and coordinating an attack between the armies was an hard work because armies had separate command chains, often without a will to collaborate.

    Functional divisions, planes with planes, ships with ships, privates with privates was not the more effective organization to win the war, a central command of all armies was the right answer.

    It is not that divisional organization is better, it is that functional organization must be divided between function in a way that allows cooperation not in a way that block it.

    Organic work is the key for functional organizations (I mean cases in which every functional division has its own command chain).

    Hardware and software functional division works because they are interdependent and must cooperate, database and graphic interface functional division don’t work since cooperation is not needed and must be forced between the command chains.

    Every strategic goal (not products, goals) must have its own unique command chain with a unique cost center to measure effectiveness of success.

    Division inside strategic goals command chain must be made with a zen vision, what is already divided must be divided what isn’t must be unite.
    You have to divide in functions that have to cooperate with each others not are ordered to cooperate. That can be done at any level where a division can be found.

    A by product division segment the company in groups that does not have to cooperate and creates cost centers that are not bound to strategic goals.
    A function division, if done right, dividing what is divided (zen pause), under the same strategic goal and cost center can achieve great effectiveness.

  • Cook Knows

    Might this be a significant part of the reason that Scott Forstall was fired? Both by being head of the rainmaker operating system, and by reports I’ve seen being kind of prideful of that, might have been exactly the kind of threat you’re posing. Maybe Tim Cook saw that and understood that he had to nip that in the bud.

    In any case, whether right or wrong, firing Forstall was a very bold move for a new CEO.

    • I believe that indeed it is.

      Sent from my iPad

  • Bruce_Mc

    Praetorian Guards are created because the leaders of the country fear it’s inhabitants. One job of the Guards it to make sure that fear is mutual: to make the inhabitants fear the leaders.

    I think the failure of a company to change in response to changing conditions is often due to fear, of losing what they have and of failing at something new.

    What you are suggesting in functional vs. divisional structure is that it turns out to be an indicator of a specific kind of fear (or lack of it) in an organization.

  • Bill Esbenshade

    Horace – great post and show! I’m speculating a little bit, but I think Forstall was fired because even though he was a functional head (of iOS), he wanted divisional like control over the future direction of the iPhone. He wasn’t good at collaborating with the other functional heads, like Ive and and Mansfield, because he wanted divisional like control over the entire product and all the functional aspects of the product — just like MacArthur wanted absolute/divisional likeGood design emphasizes the usefulness of a product whilst disregarding anything that could detract from it. control, across functions, over his particular “product,” the Korean War. MacArthur and Forstall were

    • James King

      I really don’t think these military analogies apply to the Forstall circumstance. It makes for great mind exercises but has very little real-world applicability.

      Think of Forstall’s relationship with Steve Jobs. It’s obvious that Forstall had a ton of political cover because of that relationship. However, when Jobs died, Forstall’s greatest champion was gone. Any commonalities regarding the direction of iOS also went with Jobs. In the end, Forstall became vulnerable politically. This is more a matter of Praetorian Guards turning on one of their own.

      As a result of his position, Forstall was rich and powerful. More importantly, he saw himself as the champion of a vision shared by he and Jobs’, one which was crushingly successful INITIALLY and was still overwhelming successful by P&L standards; the old “if it ain’t broke, don’t fix it” maxim that ends up running companies into the ground.

      Tim Cook is an innovator, just in a field that is not very glamorous. That makes him very unique among CEOs, who are normally hired for their bean counting or charisma. Cook validated Jobs’ faith in him when he realized that iOS was stagnating, despite the numbers to the contrary. He knew he had to make room for new ideas but that Forstall had accumulating far too much power and influence due to his relationship with Jobs and the initial success of iOS. It was a bold move and a rare one. Most CEOs in that same position would have deferred to Forstall and backed him because he already had a proven formula. This was a case of the “elite” turning on one of their own to protect the big picture.

      Comparisons of companies to militaries isn’t accurate simply because the psychologies do not really parallel. While companies attempt strongly to control costs, most militaries, at least Western ones, are given a blank check. While most companies make a point of controlling labor costs, most militaries are constantly seeking new bodies to fill their ranks. Militaries have the incentive to push the envelope technologically because their potential enemies are constantly investing in counter-measures which are generally much cheaper. The U.S. particularly fields the most technologically-advanced military in history and continues to invest heavily in R&D. In many ways, militaries are the antithesis of private sector companies. When people say that it takes a CEO to run the U.S., my first response is that s/he wouldn’t be able to terminate a huge portion of the popoulation to improve the country’s bottom line. Running a company is very little like running a military… or a country.

      The analogies simply do not apply.

      • InnovateOrDie

        “Militaries have the incentive to push the envelope technologically because their potential enemies are constantly investing in counter-measures which are generally much cheaper. ”

        That sounds precisely like Apple!

  • James King

    I’d like to address the “functional vs. divisional” argument created by Horace:

    I could be wrong but Horace seems to think that structuring a company around the “verbs” (what is done or what action is performed) instead of the “nouns” (a particular product or service) allows Apple to constantly allow for a level of disruption that is managed internally to its own benefit.

    There is the illusion of causality to Apple’s hierachical structure. However, I’d like to propose an alternate reason Apple seems to be able to innovate and manage disruption so effectively.

    Can anyone name ANY Apple executive outside of its highest chain of command? Someone at the Rahul Sood, Charlie Kindel (when he worked at MSFT) level. I can’t and I doubt that anyone who doesn’t actually know someone who works at that level in Apple can.

    Apple under Jobs was the epitome of a benevolent dictatorship. Now it is an oligarchy with Tim Cook being first among equals.

    The Peter Principle states that people will rise to their highest level of incompetence in an organization. People will also HIRE at that level as well. Jobs was a brilliant man who was very well-rounded intellectually. He was literally a one-in-a-billion leader who hired up to his level. Every member of Apple’s leadership team is brilliant in his particular field. In other words, because he couldn’t manage roughly 70K employees directly, Steve found a bunch of brilliant guys and placed them at the most valuable strategic areas of the company and then managed the small handful of them. The expectation was that thinking would permeate throughout the rest of the company. I’m sure that every one of Apple’s SVPs has lieutenants that are very similar in personality and capability.

    But Steve did something that is not particularly appreciated by proponents of “flat” or “matrix”-style management: he enforced the hierachy BRUTALLY. In Apple, real power ONLY exists at the SVP level, everything else is transient or expendable. The top of Apple’s leadership chain is also the most fanatically devoted. Jobs expected the absolute most from his leadership and anyone who failed or rocked the boat was ruthlessly culled. By the end, he had a leadership that reflected his vision made up of some of the most brilliant minds in the tech world.

    A ruthless, brilliant, fanatical elite? Sounds a lot like Praetorians to me.

    By the time he died, Jobs knew that Tim Cook was the man to run the NEW Apple, the company that had gone from scrappy upstart to the most powerful tech company in the world. That company already had Jony Ive to design, it needed a leader who could PRODUCE at the scale to which Apple had grown, a logistics genius. He also needed someone who was both even-tempered yet could execute ruthlessly to maintain the vision laid out by Jobs. The vision that Jobs had for Apple became its legacy with his death. Tim Cook understands that Apple will never have a new vision because Jobs vision is embedded into the company’s DNA. He only has to use good judgement to keep that vision alive in a world that is changing rapidly.

    With Forstall, the Praetorians culled on of their own because his ego began to interfere with the mission. Forstall likely made the mistake of thinking he had inherited Jobs role as the soul of the company because they shared a common vision and were good friends. However, Jobs had already made it clear who was to LEAD the company after his death and Forstall likely made the mistake of not reading the tea leaves.

    Brilliance begets brilliance. Apple is an extraordinary company led by extraordinary leaders. There are no CRIPPLING politics in Apple because its leadership is secure in their roles and its corporate culture has given them an unchallenged and unchallengeable mandate. They are only beholden to each other, none has to worry about challenges from hungry, talented subordinates. They all know that just having Apple on a resume is pure gold, any talented subordinate who is not patient enough for the old guard to step down has more than enough credentials to start their own venture or stand out at another company. It’s a power structure that could only be built by a once-in-a-lifetime leader. That’s why it is so hard to duplicate.

  • Michael

    I haven’t listened to the entire podcast yet, but I find this to be an interesting thought about Apple’s success and its organization. I am not sure there is enough info to decide which was the causality runs, and in any case, the most likely answer is, in both directions, as this is a living organization that responds to its environment.

    It is a high level discussion and I am pretty surprised to have not heard or read any comments about Apple University. Apple is probably unique in having within a huge corporate structure, a unit devoted to studying the company and its environment in an academic sense and ostensibly teaching the lessons learned to its employees. In fact, I am extremely surprised that Horace hasn’t been hired to contribute to that effort, though it does allow us to benefit from his creative analysis.

    The point being that a company with a ‘university’ within it is bound to have a unique corporate structure and the titles of its senior management clearly shows that it is particular functions are the core organizing principle.

    • Michael

      Course Apple U was discussed in the podcast, just 2 min after the point at which I paused to write the post above!

  • KirkBurgess

    The lack of divisional structure I think was clear during the WWDC iOS 7 introduction with the complete lack of any demonstration of iOS 7 running on the iPad, and any mention of iPad specific improvements in iOS 7.

    This is possibly a negative example of the policy of functional structure, unless of course Apple is saving some iPad specific iOS improvements for the next iPad launch event.

    Despite the massive success of the iPad, I think most here would agree that it has a slightly different job to be done than the iPhone, and perhaps the lack of focus on the iPad implementation of iOS is perhaps holding back the iPad platform slightly, even though it must be said that having a shared OS with the iPhone has obviously given the iPad massive benefits.

    • neutrino23

      My understanding is that iOS 7 wasn’t shown on an iPad as it wasn’t ready, not for any political reason.

    • It’s ready now and wow!

  • A fascinating topic, post and podcast, Horace!

    Your are right that the rarity of functional organisations in big companies makes this a difficult topic. However, there is the Nokia experiment in 2007-2010 that gives some insight to your hypothesis.

    Announced in late 2007 and operational in January 2008, the re-org dismantled the old divisional business groups (Multimedia, Enterprise Solutions, Mobile Phones) and organized the company around functional teams and interdependency. The chosen model wasn’t 100% functional all the way to the top, but I doubt that Apple’s is either if we would know all the details how the financial delivery responsibility is split. For Nokia, the blueprint for the model was taken from the academic work of Mikko Kosonen (the Nokia strategy SVP in the late 90s and early 2000’s and the current head of Finnish Innovation Fund Sitra) and is well explained in the June 2007 Harvard Business Review article: The New Deal at the Top (

    We all know what the results were. And as things started to fell apart, the divisional P/L returned, first to Mobile Phones in 2009 and in Feb 2011 to the whole company.

    Now, in all fairness, between 2007 and 2010 the other things around Nokia org were hardly ceteris paribus. The were so massive changes in the competitive landscape (including Apple, the world’s most capable computing firm entering to the business) and some such a big strategic missteps – both in the company and the functional level – that it is guesswork to say conclusively what is the role of the operative model in the downward spiral.

    I did experience, however, the model close enough (2007-2009 as Product Director and 2009-2010 as Product Mgmt VP) to witness many of the things you talked about in your podcast. For example:

    * Like you point out, most senior executives are fundamentally wired to become general managers and CEO. That desire to one day “run one’s own show” keeps them working those ludicrous hours, even when they are already wealthy enough. They build patiently CEO-readiness by accumulating as broad functional and “sub P/L” experience as possible. Even if they’d like to get to the top differently, the lack of dual ladder thinking in the CEO succession playbook of public company Boards (and their headhunters) makes it unrealistic. In Nokia’s case, I’ll just say that the transition of many these type of leaders to succumb to “just” to their functional roles didn’t go smooth, creating all kind of issues that plagued unity, focus, speed and decision-making.

    * Consistent with your “there’s no head of iPhone” example, the lack of smartphone segment P/L leaders did help to turn the focus of the organization from (over)optimizing the current products towards building new business. Much of 2008, the company was emotionally really wired about moving to Services business, with a lot of people trying to feverishly position themselves for the right bandwagons. And, surprisingly, at that point the lack of divisional leaders for the existing business turned into a major liability. While the features in Nokia’s products may have bordered overoptimization (or overserving), operationally the crazy growth years had left their scars. The Symbian OS had accumulated massive technical debt and the only reason for the consumers accepting the quality was that the competitors were even worse. The product overlap after the business groups was much bigger than admitted yet people were felt locked to their previous growth targets. Chipset implementation were in the middle of major transition from telecom architectures to computing, moving radically the goalposts of device performance. In most markets, the ability to sales folks to sell anything else than a sealed box was a suspect. The software distribution architectures, quality and and customer care of legacy products couldn’t handle the flood of new “cloudified” features. And so and so forth. All these problems were very difficult to solve within each function alone, but also presented a massive conflicting agenda for the CEO-level, trying to focus on moving the company to the Services era. And how it all turned out can be googled.

    * And of course, the functionality strengthened information silos as people really know only their piece of the product and the interface towards other teams. This is great for secrecy but detrimental to speed and quality if peope can’t come to resolution how the pieces fit together. In Nokia’s case the interface between Services and OS group was not mature enough. How could it have been as Services was brand new and mainly consisting of people with no device engineering experience to whom the front-line device guys felt threatened by the losing the control of delivery schedules. As a result, way too many issues ending up escalated to be solve in cross-functional so called Solution Board.

    As a summary, I wouldn’t intepret any of my learnings to disagree with your hypothesis. Indeed, the functional organization may well best suited to constantly drive innovation and move product offering forward within one easily understandable trajectory. However, the Nokia experiment gives supporting evidence how difficult it is to transform into such a model, especially if one tries to evolve business model similarily (from device monetization to services monetization) and the “old” business has still big cross-functional optimisation to do.

    So what really makes functional model work so wonderfully to Apple may indeed be that it’s the only way they know how. And that people stay in the same functions for forever, meaning that any unaddressed problem they leave behind they will always always find in front of them.

    • DesDizzy

      JPS Thanks for your real world insight. One of the things that makes this forum so valuable.

  • neutrino23

    In addition to the structure we should consider Apple’s mission statement. Steve Jobs removed computer from Apple’s title. They are not a computer company. Their purpose is to bring technology to consumers. As such, they are not tied to any particular product. Other companies were tied to specific products and suffered when those products lost popularity. Dell made cheap PCs, Microsoft is tied to Windows, and BlackBerry was identified with secure email.

    Put another way, if you grow roses you want everyone to buy roses. But if you sell flower arrangements, you don’t care what kind of flowers are currently popular. If you grow roses your company has job titles specific to that task. If you produce flower arrangements, your company does not have any job titles tied to specific types of flowers.

  • Watcher

    One must also not forget that militaries tend to be divided between officers and enlisted personnel. Officers are usually trained in a single academy, and will rotate much more freely during their careers (leading units of increasing sizes, basically), while enlisted people will receive a more specialized training and tend to remain on that role for their careers. In a move akin to your ‘preemption’, senior enlisted personnel will still be subordinate to junior officers, notwithstanding their long experience.

  • Pingback: The Leadership Dilemma | Ego Sum()

  • Pingback: Microsoft’s future | Bayjinger()

  • Pingback: Does a functional organization do a better job of… | Todd Simmons()