The Complexity of Simplicity

There have been a number of posts over the last few weeks that touted the importance of ‘creating simplicity’ for making BPM successful. To be honest, I see most of those posts as shallow in their understanding of the real-world business impact. Perceived simplicity is not created by making things simple, but requires substantial complexity. I even propose that it needs to be complex and not just ‘very complicated.’ Those who just enforce simplicity in rigid processes for cost reasons ignore accepted scientific knowledge.

Complicated versus Complex
The English Thesaurus unfortunately treats the terms complex and complicated as synonyms. A system is complicated if it is difficult to decompose in its constituent parts or if it is difficult to assemble. The complexity (complicatedness?) of a SINGLE object, named after Soviet Russian mathematician Andrey Kolmogorov, is a measure of the computational resources needed to specify the object. It is also referred to as descriptive complexity, Kolmogorov–Chaitin complexity, algorithmic entropy, or program-size complexity. In ‘Complex Adaptive Systems’ (Holland, Lansing, et.al) it means that the system cannot be fully understood or created by simple assembly. CAS describes an inner function of a system that is created by emergence, meaning the combined functionality cannot be predicted from its parts. Complex adaptive systems typically have a chaotic element, meaning that processes are susceptible to initial conditions. Without changing the inner function the outcomes vary when initial conditions change. Initial conditions cannot be used to predict or calculate outcomes. CAS are used to describe many natural phenomenon such as biological and social structures. CAS are defined as individually acting entities/agents whose interactions are only loosely coupled. They can be statistically observed but that does not produce data that predict behavior because of the chaotic component. Social Networks are CAS and thus one cannot slap social communication onto a rigid process management infrastructure and ‘hey, presto – its social.’ Flowcharts are complicated and social networks are complex! The science of social networks doesn’t enhance BPM, but should be heard as the bell that tolls for flowchart illusions.

The Economy as a Social Network
Social Networks aren’t really as new as they seem to be. All social species use them one way or the other. Already in 1908 Georg Simmel wrote about the aspect of social types and connections. Technology empowerment allows us see them and thus they can grow exponentially without person to person contact. But the technology does not change how social networks work, it just makes them broader, with different, mostly weaker relationships and generally more diverse. Mark Granovetter wrote in 1973 an article in the American Journal of Sociology that started it all: ‘The Strength of Weak Ties’. Duncan Watts, Stanley Milgram, and Albert-Laszlo Barabasi created over the next years what is understood today as Network Theory. Network Theory has improved our understanding about how revolutions, terrorism, ecology, economics, and organizations work. Watts found that networks go through a kind of phase transition as they grow outwards and form clusters until they reach a kind of equilibrium of local network communities that are widely linked by people’s weak ties. Barabasi found that the ‘power law‘ or Pareto principle applied, in that only a small percentage of people need to be active and strongly linked for a network to flourish. The one thing that does not improve a social network is control or enforcement beyond some principal rules. Social Networks can be analyzed once they have formed but they can’t be deterministically influenced because they are CAS. I propose that the same is true for all of economy and it is also true for human-interactive IT systems that are not just number crunching.

Influence is weak and indirect
Christakis and Fowler analyzed the impact of social networks in their book Connected. The results are counterintuitive. Surprisingly, we can influence such social clusters (usually smaller than the Dunbar Limit of 150) by reaching out to the friends of our friends rather than trying to get in there ourselves. And this works in reverse too. The friends of our friend’s friends have through their indirect influence a dramatic effect on our perceptions and perspectives. Now consider how this understanding of social networks should impact business management. This is not about using Facebook or Twitter as an advertising medium. Amazon is successful because their customers form a social network who share product information freely. Org charts, in difference, don’t show how influence takes place in a business. In reality businesses don’t function through the organizational hierarchy but through its hidden social networks. Modern management approaches already try to utilize that principle. Social networks are about understanding the human aspects of doing business  that are not taught in MBA programs. More on that in a minute.

Innovation and Change are Natural
The most profound element of change is innovation. I would even say that innovation is the only true human-driven change taking place, beyond catastrophic events. Many seem to believe that they can control or manage change/innovation or feel that they have to promote it. The opposite is the case. Change can not be promoted but just be prohibited, mostly by trying to control it. People do not resist change or innovation, but they resist the insecurity created by change beyond their influence. Artificially enforced organizations resist change, because people are insecure in their role in the structure. Social networks grow and change without anyone promoting that in particular. Change happens naturally and always as long as the environment is friendly. If there is enough pressure change happens in unfriendly environment such as revolutions. Product innovation like from Apple is driven by highly motivated individuals who are NOT interested in money. Their innovation is just gradual improvement too. Steve Jobs did not invent a single thing. He just took existing products (PCs, MP3 players, laptops, phones, tablets, Napster and now even the Cloud) and made them incredibly user-friendly! So this is not about Darwinian evolution in economy and business or being a technology pundit. Let me simply turn your thinking upside down because innovation in economy (as business) and technology are intertwined whether we like it or not. This understanding makes BPM as a management methodology ONLY a waste of time!

The Nature of Technology
Economist W. Brian Arthur discusses in his book the core principles of technology innovation from an economy perspective. Technological progress is linked to discovering the workings of natural phenomenon and putting them to use in ways that are desirable to people. Beyond the invention of the wheel, virtually all technology is complicated, meaning the functionality is progressively building on previous discoveries. See my previous comment on Apple. Once we reach a certain level of dependency we are no longer able to predict how the numerous discoveries will together form the basis for further innovation. The combinations of technology are not driven by additive functionality but by applying human creativity to the possible recombinations of technology. The technological advance as it impacts the social landscape becomes complex and goes through a phase transition – rigid crystal structure turns into weakly cohesive molecules each one exhibiting Brownian motion. It starts to resemble a CAS! Apple could not predict which Apps the now 300.000 developers for the AppStore would develop. They just created the infrastructure for a huge, fairly open social network of developers and customers, only providing the security for money and information. They also made it simple on the surface, while the underlying technology is immensely complicated. But the Appstore ecosystem does not prohibit social complexity. It is now obvious as a hindsight that 300.000 motivated programmers would easily outsmart those rigidly organized Microsoft software labs who don’t really listen to consumers.

Everything is Obvious (Once You Know the Answer)
Duncan Watts also wrote a book by the above title in which he vividly shows that most of the explanations that we invent to explain something that has happened in social networks are utterly useless. He provides a grandstand historic view of why most tries to manipulate human interaction fail. Watts argues that we do not have usable models of human motivation for individual or collective behavior to predict outcomes of such manipulation. Like Dan Pink, Watts also presents the case that most people aren’t motivated by incentives (ie. to hit KPIs) but by autonomy and recognition (AppStore developers will take the money too).

So what am I leading up to? Let me assure you that this is perfectly applicable to the subject of process management. The analysis of business processes is no more than an ‘Everything is Obvious’ illusion of WHY something happened. Trying to control that process in its execution produces more rule beating behavior rather than well-defined outcomes. It is the law that makes the criminal and it is the rule that produces the transgression. Each rule requires an enforcement mechanism that carries a substantial cost and it needs a verification and innovation bureaucracy that is even more expensive. Clearly rigid rules and processes do not save money! They only produce expense and thus we need the least possible number to ensure compliance. No more. Using technology empowerment to motivate people to pursue well-defined goals autonomously, while receiving recognition for a job well done is a lot cheaper than process enforcement. Because businesses are social networks too, it means that technology empowerment rather than rigid processes will be effective and efficient. Let me remind you that empowerment is about authority, goals and means.

Planning creates Risk
Considering the above there are few who realize the true connection between planning and risk. Most people will say without thinking that more planning reduces risk. That is however not correct at all. It is the planning that creates the risk. Risk assessment is not about reducing an arbitrary risk of the unexpected, but about creating awareness what risks are being created by the plan. The more assumptions a plan makes the more risky it is. The deeper the planning hierarchy is, meaning the level of dependencies of planned outcomes, the riskier a plan is. The more time passes between planning and completion, the more risk the plan creates, mostly because the target context will change whether one plans for it or not. BPM defined processes are no more than ‘repeatable plan fragments’ that are assumed to guarantee outcomes. The process analysis uses many ‘terms’ to describe the plan, carrying over the risk of term confusions. Business rules create the risk of event/context mismatch. Process monitoring uses a measurement assumption the increases the risk of self-fulfilling prophecies. Each rigidly defined process increases the risk of business execution not producing intended outcomes through term confusion, assumptions, late monitoring and contextual change over time.

But isn’t good planning the epitome of good management? Isn’t using well established methodologies the reason for teaching them in MBA programs? I propose that inept management tries to create predictability where there is none (see my post: Design: Art applied to Productivity!). Henry Mintzberg is a former professor at McGill University who was teaching MBAs at MIT in the early Ninetees. In his book ‘Managers, Not MBAs‘ he describes a profound disconnect of MBA graduates between the practice of management and MBA teaching by words such as ‘mindless’ and ‘arrogant.’ Mintzberg focuses his sight on high-level processes such as strategic planning and sees the same problem that I see in BPM: “the assumption underlying strategic planning is that analysis will produce synthesis: decomposition of the process of strategy-making into a series of articulated steps, each to be carried out as specified in sequence, will produce integrated strategies.” Mintzberg proposes adaptive strategy-making that is capable of incorporating emergent opportunities. His exact words, not mine!

The Complexity of Simplicity
The foregoing elaboration (I apologize for its necessary length) shows that we are entering a world of visible and fast changing, social complexity in which simplistic methodologies of strategic planning, business and process management fail to take into account the economic phase transition of complexity that we are going through. While simplification for the user is the key to technology adoption it is the underlying social complexity (not just complicated technology) that enables it. Technology is not making our world more complex as many claim, but it simply shows to us the true complexity that exists. Technology does however cause a shortening of time scales and therefore challenges today’s slow strategic planning cycles. It clearly challenges the proposition that process analysis will enable synthesis in realistic timescales. Rather than relying on analysis, rely on technology empowerment for social business interaction. Do not forget: Economy and technology are Siamese Twins.

Conclusion

This shouldn’t be surprising to anyone interested in more than just flogging their product. Information Theory was created by Claude Shannon in his 1948 paper ‘A Mathematical Theory of Complexity’ that also defined the ‘bit’. So it all started out by understanding complexity. Benoit Mandelbrot took the next important step when he focused in the early 1960’s on the randomness of data channels and identified what he called persistent trends (Joseph effects => structured => bell curve) that describe incrementally dependent information, and sudden events (Noah effects => unstructured => power law) that create discontinuity.

A Mandelbrot Set

Mandelbrot sets of simple rules can be utilized to reduce the Kolmogorov complexity of life-like models, which is vital in computer graphics. Mandelbrot sets are unpredictable, but clearly exhibit recognizable patterns in all levels of hierarchy. Therefore, I see pattern matching as the future of process mining and not algorithmic processing (aka rules) of analysed past causes that will hopefully repeat themselves in another context in the future. The phase transitions that are inherent in Complex Adaptive Systems also mean that a change in scale (company or market size) dramatically reduces predictability.

IT is however mostly about algorithmic data processing and even predictive analytics can only identify the trends of Joseph effects. But correlation is not causation and even if we can create a simple rule set it cannot model the emergent phase transitions of CAS. Rather than simplicity and predictability, we ought to focus on manageability as we can’t avoid the inherent complexity in social structures by defining rigid processes. Enforcing structure for predictability is expensive and stops innovation. Technology empowerment for adaptive processes reduces term confusion, planning depth, late monitoring of outcomes, and time scales. Use a Business Architecture with value streams to define strategic objectives, operational targets, and process goals, but do not nail down rigid processes that assume outcomes. While unpredictable Noah effects are unavoidable in business processes they tend to follow recognizable patterns that we can identify and highlight.

About Max J. Pucher

I am the founder and Chief Technology Officer of Papyrus Software, a medium size software company offering solutions in communications and process management around the globe. I am also the owner and CEO of MJP Racing, a motorsports company focused on Rallycross or RX, a form of circuit racing on mixed surfaces that has been around for 40 years. I hold 8 national and international championship titles in RX. My team participates in the World Championship along Petter Solberg, Sebastian Loeb and Ken Block.

‹ The Risk in the Beauty of Value

Agile and Scrum versus the God Complex ›

Tagged with: Barabasi, Dunbar, Duncan Watts, Granovetter, Kolmogorov, Pattern matching, Stanley Milgram
Posted in Adaptive Case Management, Adaptive Process, Business Architecture, Complexity, Economy, Social Media