The collapse of complicated business models

Clay Shirky, in the collapse of complex business models, notes:

Bureaucracies temporarily reverse the Second Law of Thermodynamics. In a bureaucracy, it’s easier to make a process more complex than to make it simpler, and easier to create a new burden than kill an old one.

The premise of his article is that successful organizations and industries become more complex over time and are unable to embrace new ways of doing things, which at the onset are much simpler. He discusses the complex television industry and how it cannot produce simple, and low cost, fare for the web.

I’m not sure if complexity is the issue. I see it more as complication. Companies and industries start out as relatively simple operations and then become more complicated. Complicated systems can be analyzed, and we can tell how things work. Modern organizations are not complex, they are merely complicated. A complex organization could not be managed.

The real problem is on the outside, not the inside, of the typical complicated organization. The outside environment has become complex and the complicated organization lacks the ability to deal with it. Systems like the Neilson ratings don’t give us the kind of information we need to make decisions on programming. The media landscape is too fragmented to completely analyze.

The Cynefin framework describes the complicated & complex domains as:

  • Complicated, in which the relationship between cause and effect requires analysis or some other form of investigation and/or the application of expert knowledge, the approach is to Sense – Analyze – Respond and we can apply good practice.
  • Complex, in which the relationship between cause and effect can only be perceived in retrospect, but not in advance, the approach is to Probe – Sense – Respond and we can sense emergent practice.

The typical large modern organization tends to thoroughly assess a situation before acting, assuming it can be analyzed. This does not work in complex environments where we need to first do something and then see what happens. We see this with Beta releases of web services, which adapt as they are used by more people. many web companies understand this.

I don’t see simplicity as the solution to dealing with complex environments. A new organizational structure is required that is 1) based on simple units but is 2) connected as a network that is much more complex than any hierarchical organization could ever be. This type of organization will be too complex to manage directly. It will self-manage and adapt. The best work structures to deal with complexity will be complex networks, and likely some mix of wirearchical, chaordic, democratic, etc.

17 Responses to “The collapse of complicated business models”

  1. Howard

    I see 2 problematic issues at play.
    1. Most large bureaucratic organizations come to depend too much on command and control style standardized practice. Would it not be that most large intrenched orgs are the opposite of a wirearchical org. It leads them to sense – respond with little or no room for analysis or probing. In difference to critical thinking, they would not know what to do with disconfirming evidence.
    2. Infrastructure plays a role where there is disruptive change. IBM is a master at developing new infrastructure. From one that supports a main frame company to one that supports a micro computer company, to a mini computer Co., and eventually to a software and service organization. (This idea came from a Harvard Bus. podcast I’d have to look it up Let me know if you need it). It is very difficult to make such drastic change. I believe the podcast suggested that IBM was successful because it did not change so much as set out to create a new organization at each stage of it’s business.

  2. Stephen Downes

    The Cynefin framework isn’t the only thing to define it like this, of course, and I’m not sure we halp our case by citing it as definitive of ‘complex’. The term can stand on its own without a framework, surely.

    And even if not, I’m inclined not to cite the Cynefin framework until it gets a proper name. ‘Cynefin’? Really, now.

  3. Martijn Linssen


    I absolutely agree with you on complex vs complicated. I’ve used these two words for years to describe and analyse problems, organisations, challenges, etc.

    In I describe certain phase of organisations, from entrepreur to extinction (where extinction is definitely the last phase)

    I’ve (just now) realised that Intimacy plays a decisive role in all this. What stands in between complex and complicated are humans: at a certain point, humans become a factor in the business model. Customers and suppliers aren’t meant with this, but employees. Uncontrollable employees. They have become uncontrollable because they’ve lost interest, and care; they’ve become anonymous to the organisation, like the organisation has become anonymous to them

    Your solution is simply brilliant, and touches the hiveminds of Social business Design. It will eradicate anonymity, and that’s a good part of the reason why it’ll work – I think

  4. Jo Jordan

    Oh great! I was rabbiting on about complex and complicated yesterday as well! It does seem to be a key idea in work psychology that unlocks a lot of well, complication!

    I like your definitions and the point toward a solution. Shirky is a sociologist though. So he isn’t looking at what an organization could do (we are drawn to complex because we deal with emergent actions).

    He is talking about what society is likely to do. He should make suggestions at a policy level but he is making the point that some societies go into grid lock. At the point we have a catastrophic change (in phase state). It is emergent all right, but with consequences that destroy what we once took to be a good system.

    I have seen this happen at close quarters and wish I had read the book he quotes 20 years ago as I watched a collapse scenario unfold. I knew exactly what was happening and lived through the frustration of grid-lock. In the end the collapse happens because we can’t make the small changes that would unlock our ability to act.

    Hmm, looking at your definitions again, I would have though that complicated usually implies not having the right feedback. It is when there is feedforward (arbitrary procedures usually), little or distracting concurrent feedback, and feedback that might suggest feed forward was arbitrary.

    Complex means bringing responding to several concurrent feedback streams where action on one stream can affect another. We enjoy that because we are getting feedback immediately and are dancing with the situation. (Too many jobs aren’t sufficiently complex and we get bored.)

    The first disrupts the tracking task and the second aids it.

    Isn’t your complicated a diagnostic task which can be highly paid because it is difficult (e.g., investigating an air traffic accident.) Landing a plane is complex. Dealing with disgruntled customers at a restaurant is disruptive for the waiter but diagnostic for the owner. Matching the task to the rhythm changes its status.

    A creative task such as good negotiation is different again and is like your complex task. We have at least one feedback stream that is aesthetic/holistic. As we see a pattern emerge, we try to support it (we could over egg it too.) When this is a social task, involving many people is the goal.

    A thought.

    I like where you are going with Shirky’s idea – at the company level. I am still curious about the possibility of a global collapse.

  5. Daan Assen

    Interesting post! The distinction between complex and complicated is insightful. The external environment is complex in many industries, but this also counts for internal organizations (often with the downsides of this complexity as a result). I agree that the complexity offers opportunity. I am currently reading about Absorptive Capacity of Organizations (Cohen & Levinthal, 1990; Zahra & George, 2002) and this is an interesting concept in relation to your plea for new organization structures. Organizations need structures, systems and culture that enhance the absorptive capacity of the organization. This enables them, via a network of skilled professionals, to reduce complexity through the absorption of knowledge and generating competitive advantage from this knowledge.

  6. Esko Kilpi

    Dear Harold,

    I may be problematic to say that “the relationship between cause and effect can only be perceived in retrospect”. In complex systems, such as the weather, it is normally impossible to find the relationship between cause and effect – even in retrospect. You can’t trace the butterfly effect back to one butterfly 🙂

    • Harold Jarche

      Yes, I think it would more accurate to say that the relationship between cause and effect can only be seen in retrospect, if at all. I believe that is the intent of the wikipedia article as well as the original HBR paper.

  7. Chris Rodgers

    Hi Harold,

    I agree with you when you say that what Shirky is speaking about in mal-functioning bureaucracies is the tendency to overcomplicate things. The more formal systems, processes and procedures that managers introduce, the more complicated that these tend to become. So, reducing the clutter of unnecessary or overly complicated structures, systems and procedures would certainly make organizations simpler in design terms.

    However, I cannot agree with your view that organizations are not complex. On the contrary, I would argue that they are inherently so. As soon as you have a couple of people interacting, you move from complicated to complex. Indeed just being a human is complex! So managing human beings will never be complicated; it will always be complex.

    The complex social dynamics of self-organization, emergence, non-linearity and paradox are ever present. They are integral to the ongoing conversations and interactions that make up everyday organizational life. These dynamics are affected by – but do not depend on – how simple or how complicated an organization might be in terms of its ‘infrastructure’, inter-connections, issues or whatever.

    Viewing organizations as dynamic networks of self-organizing conversations also means that there is no “inside” and “outside” in the conventional sense. We might think of an ‘organizational identity’ as providing a boundary between “us” (inside) and “them” (outside); and we might also define a boundary for legal purposes. But the conversations that determine organizational outcomes are unbounded.

    Based on the above, therefore, I have to disagree with you when you say “A complex organization could not be managed”. Managers are necessarily dealing with these complexities all of the time – even if they are not aware of the fact or believe that these can be ‘designed out’ or otherwise dealt with through rational planning and control. The challenge instead is for them to actively engage with these complex social dynamics; seeking to shift the patterns of interaction, conversation and resulting behaviours in organizationally beneficial ways. Even then, they can act with intention but with no certainty of outcome.

    Cheers, Chris

    • Harold Jarche

      Very good points, Chris. Given your statement: “Viewing organizations as dynamic networks of self-organizing conversations also means that there is no “inside” and “outside” in the conventional sense.” I would think that no conventional organization fits that description. Almost all organizations have hierarchies and boundaries between inside & outside. These are the organizations that Shirky says are getting too complex, and that is the point I disagree with. The world is getting more complex, but we are still encumbered with simplistic to complicated organizational structures. We need new organizational structures (something like value networks or chaords) but these cannot be managed in a traditional command & control way. I guess I should have said it that way initially. Thanks very much for your comment.

  8. Chris Rodgers

    Many thanks, Harold, for your response to my earlier comment.

    However, to clarify, I’m not advocating a new organizational design. I’m simply describing what is already happening – today – in all ‘organizations’. This is the case whether these are governed by command and control, led by enthusiasts of empowered self-management, or if they happen to be formally structured in any of other ways that you advocate.

    Formal organizational designs (structures, strategies, systems, processes etc) inevitably both enable and constrain everyday local interactions; but the latter are affected much more powerfully by the informal, ‘shadow-side’ dynamics of organizational life. It’s here, in the give-and-take of everyday conversations and interactions, that people make sense of these formal ‘designs’; of managers’ stated intentions; and of emerging events. And it’s also here, in the light of their specific, local circumstances, that they decide how they will act.

    Organizational outcomes then emerge from the population-wide interplay of these ‘local’ (i.e. one-to-one and small-group) conversational interactions. This is inherently a complex social process, in which overall outcomes are unpredictable and uncontrollable by any individual (or group of individuals) – irrespective of how thoroughly they might have assessed the situation before acting.

    So complexity isn’t a function of design. It’s a natural dynamic of people interacting together (whether they formally reside at Board level; have jobs at the ‘work face’; or, from a conventional standpoint, have nothing to do with the organization at all, but interact informally with those who do).

    Too often we talk about ‘organizations’ as if they are ‘things’ that exist – and act – independently of people’s day-to-day interactions. In reality, as Ralph Stacey would say, nothing exists outside these local conversational interactions, other than more interactions. And these conversations show no respect for the formally defined ‘organizational boundaries’.

    Cheers, Chris

  9. John Tropea

    I really like the clarity of your comments Chris.

    Agree that interactions like conversations are complex, therefore an org is, but that doesn’t make it a system.

    Even though an org may be complicated, I don’t think that makes it a system either.

    Just the same even though an org is not a system, as removing a part does not kill it., and also people have their own motivations…the missing or injured part still has impact on the rest of the org…but we can have more ambient awareness with new social tools…all we can do is cope.

    A photocopier is complicated like an org, but is also a system, But what I’m stuck on is that a complicated system like a photocopier, can still operate (inefficiently at best) with a missing part…but if the missing part is something essential like a carburetor in car, well then it can’t operate.

    Can this happen in an org? Can you take away a part of an org, leaving it inoperable…or will it always just be wounded?

    I recall on one of Bas’s posts that I commented on…complexity environments being subject to “unknowns”

    A photocopier is complicated as all parts are “known”.

    The interactions that happen in organisations contain “unknowns”

    So a photocopier in of itself is “complicated”, but it’s used in a complex environment. Which has got me thinking about the “inside” “outside” aspect of this discussion

    Patrick Lambe puts this nicely:

    “Let’s look at a photocopier in an office with a service
    agreement that allows parts to be changed whenever there’s a problem. Let’s say that the
    photocopier suddenly starts a strange pattern of behaviour. It jams repeatedly every
    Monday morning. Technicians come and replace the rollers, but it still happens, and it
    happens every week. It’s clear that we don’t solve this problem simply by replacing parts.
    It’s clear that something beyond the sum of parts is creating the problem.

    The smart technician will start asking questions about habits of use. He might discover,
    for example, that because of cost cutting measures, the airconditioning is now switched
    off in the building over the weekend, the paper already in the machine expands from the
    higher humidity, and it jams easily because the rollers are calibrated for thinner, dryer
    sheets. He might also discover that the sales presentation meetings have been moved
    from Fridays to Monday afternoons, placing added stress on the machine on Monday

    Understanding the science, and being able to label and replace components does little to
    resolve problems that arise from social and human initiatives and changes. And most
    knowledge and learning issues arise precisely from social and human initiatives and

  10. Jon Husband

    The complex social dynamics of self-organization, emergence, non-linearity and paradox are ever present. They are integral to the ongoing conversations and interactions that make up everyday organizational life. These dynamics are affected by – but do not depend on – how simple or how complicated an organization might be in terms of its ‘infrastructure’, inter-connections, issues or whatever.
    Viewing organizations as dynamic networks of self-organizing conversations also means that there is no “inside” and “outside” in the conventional sense. We might think of an ‘organizational identity’ as providing a boundary between “us” (inside) and “them” (outside); and we might also define a boundary for legal purposes. But the conversations that determine organizational outcomes are unbounded.

    I’m glad Harold write this post, and I like the train (ot web) of thought it has attracted.

    I think that I think that today’s organizations are complex, as Chris has astutely pointed out, but that they are managed as if they were complicated. That’s what management science has brought us, and sold us.

    What the networked era has offered us is a more complete picture of the complexity that has been there – which has, at least to most appearances, been tamped down to fit into less-than-natural industrial era structures. Yes, people in organizations deal with the complexity engendered by human interaction all the time .. much of that dealing-with is accomplished as work-arounds to the formal structure and protocols that accompany that structure, Dave Pollard has been eloquent on this in the past,


Leave a Reply to Esko Kilpi

  • (will not be published)