Chapter 4. Ashby’s Law of Requisite Variety
Pal’chinskii’s first principle[1] urged us to experiment with a variety of ideas to increase the chances of success and this was echoed years later by the British scientist, W. Ross Ashby in: “An Introduction to Cybernetics[2]” — the science of communication and control in living and mechanical systems. Ashby coined what became known as the ‘first law of cybernetics’ — Ashley’s “Law of Requisite Variety” — the amount of variety in a system must match the amount of variety in its environment to achieve control. In other words, if your environment is changing faster than your ability to respond, you’re in trouble.[3]
To understand the implications of ‘Ashby’s Law’, picture a thermostat controlling the temperature in a room and switching the heating system on or off whenever it reaches a pre-determined level. Now imagine that, instead of constantly monitoring the room’s temperature, the thermostat only checks it once a day, switching the heating system on or off accordingly. This means that some days the room will overheat as the heating was left on, while other days it’ll be cold as the heating stayed off. Here, we can say that the thermostat — the room’s ‘control mechanism’ — lacks the ‘requisite variety’ of responses to adapt to changing conditions.
In organisations management is the ‘control mechanism’ as it chooses the direction, hires teams, sets incentives, and decides which technologies to invest in to deliver results. However, unlike the simple mechanical system the thermostat controls, management has to deal with an unpredictable human system. Faced with the uncertainty caused by a system where the parts literally have minds of their own, managers often try to turn the organisation into a predictable, ‘well-oiled machine’ in order to make it easier to control: favouring “best practices[4]” over disruptive new ideas; measuring success against fixed KPIs rather than how well the systems is adapting to external changes; and pushing out the mavericks — contrarian individuals who challenge the status quo — in favour of those considered a better ‘cultural fit’ and less likely to rock the boat.
The pursuit of predictability comforts managers, but also creates a dangerous disconnecting that can threaten the organisation’s survival. By filtering out all disruptive ideas the organisation can fail to develop a ‘variety’ of alternatives responses that will be needed if the external environment suddenly changes, for example, following the appearance of a disruptive new technology or the emergence of a powerful new rival. The rigid organisation, built on ‘the one right way of doing things’, has developed no repertoire of approaches it can draw on and has no mavericks left who can provide alternatives fast. As a result these managers, when facing a crisis, are left hoping everything will ‘go back to normal soon’ — but hope is rarely a viable strategy.
Adaptivity Intelligence (AQ)
An organisation that struggles to respond effectively to change can be said to have low Adaptivity Intelligence (AQ). Low-AQ organisations lack the range (or requisite variety) of options needed to adapt successfully. They are “dead players” — capable only of following outdated scripts and lacking the ability to innovate. In contrast, high-AQ organisations are “live players[5]” — continuously experimenting with new ideas and technologies to develop a range (or requisite variety) of options they can quickly draw on and deploy if the old ways of working suddenly stop working (e.g. physical spaces close, customer behaviour shifts, supply chains are cut off). This is how live players adapt and thrive — while dead players fade away.
When told something is “best practice” dead players ask: Who did this? What was the result? Can we copy it? While live players ask: Why is this best practice? When was it best practice? How does this apply to us today? Live players are open to learning from others, but avoid relying on backward looking ‘past practices’ because, in a fast-changing world, this is akin to driving on a motorway by navigating through the rear-view mirror. Live players look forward, they seek out what works now and next in their context, embracing Pal’chinskii’s first principle of experimenting with a variety of ideas to increase their chances of success. When a sudden change hits, live platers aren’t left scrambling to respond — they have options and they deploy them.
Live players rarely go all in on single ‘big bang’ initiatives that promise to solve all their problems in one go. They understand that large projects often fail to get completed before conditions change again and seek to avoid the dilemma of having to throw good resources after bad or admit failure. They know that, even if big projects succeed, they risk locking the organisation into a rigid way of operating that will prevent it from adapting in future. Live players therefore also follow Pal’chinskii’s second principle, by accepting that some failure is inevitable, they keep projects small enough that any failure is survivable. They favour multiple, small (sometimes even contradictory) safe-to-fail[6] experiments that quickly deliver incremental results that can be built on later, giving them the edge over low-AQ rivals who struggle to deliver slow-moving ‘big wins’.
Fig.9: Ashby’s Law of Requisite Variety and Adaptivity Intelligence (AQ)
To thrive in conditions of uncertainty, management — the organisation’s ‘control mechanism’ — needs to operate like a thermostat continually monitoring information flowing in and making course corrections in real-time. But some managers struggle to deal with the amount of information flowing in, information that’s often distorted by complex, ambiguous, unreliable or incomplete data[7]. And, like all humans, managers have a natural aversion to uncertainty[8] — they find it far more comfortable to deal with the world they way they would like it to be, (predicable and stable) rather than the way it really is (uncertain and changing). To limit the disruption to their equilibrium they filter out ‘noise’ by insisting information coming in is presented in short reports, quick debriefs or case studies they can follow; whilst bigger, strategic questions tend to get pushed to the end of year budgeting and planning cycle. Yet, filtering out ‘noise’ also means also filtering out ‘signals’ about critical changes they may need to respond to now if the organisation is to survive, let alone thrive.
What Can Managers Do?
The answer lies in Pal’chinskii’s third principle: managers need to develop effective feedback loops between decision-makers and those closest to the action, to quickly identify and select what works in the local context. Managers need to support frontline staff — those closest to the action and in direct contact with customers — to experiment with new ideas and feed back information about what’s working in practice and what isn’t. This simplifies decision-making for the organisation, they need to do more of what works and less of what doesn’t. Management’s role shifts from vainly struggling to predict the future and control how people work, to one of enabling a sufficient variety of experiments to be launched (Pal’chinskii’s first principle) by hiring diverse talent (people from varied backgrounds, with diverse experiences and ways of thinking) and ensuring their experiments are safe-to-fail, (Pal’chinskii’s second principle) to ensure failure doesn’t bring the organisation down, whilst successes can quickly be recognise and amplified. This is how organisations develop the requisite variety of responses they need to adapt intelligently in a changing world. Let’s explore how this looks in practice in the next chapter — Innovating Out of a Crisis.
1 See chapter three — Pal’chinskii’s Principles.
2 An Introduction to Cybernetics. W. Ross Ashby (1956) http://pespmc1.vub.ac.be/books/IntroCyb.pdf
3 Echoed also by Jack Welch, see chapter two — Adapt or Die!
4 On the futility of copying “best practices” see the introduction: Why Best Practices Are Holding You Back.
5 https://samoburja.com/live-versus-dead-players/
6 As opposed to ‘fail-safe’ (building something that won’t fail) ‘safe-fail’ encourages the launching of multiple experiments to learn what works or doesn’t in a particular context. The portfolio of experiments launched can even be contradictory in order to test competing hypotheses — far better to take action and learn what really works (and what doesn’t) in your context than waste time debating theoretical ideas in meetings, often based on data from the past extrapolated into the future and topped with much wishful thinking. Experiments that have positive outcomes can then amplified with further action, whilst those with negative outcomes are dampened, with lessons learned as to why they failed fed into new experiments. This is an approach popularised by the Cynefin Framework.
7 See chapter one — Stop Making Strategic Plans.
8 See Introduction: Why Best Practices Are Holding You Back.
