Emergancy. (Ants math)
The video below shows a fun and famous example from the study of complex systems. It is a computer simulation of an ant colony looking for food.
The behaviour of each individual ant is very simple: they walk around randomly, without purpose. If they come across food, they take a piece of it back to the nest and leave a scent trail for other ants along the way. Other ants that are walking around randomly and happen to come across this scent trail will follow it.
You can see the anthill in the middle and three food sources at different distances. We see how the colony first takes in the closest food source, then the middle one, and finally the furthest one.
The behavior of individual ants is extremely simple, and no ant oversees the situation or even the problem. Let alone one that has a plan, or an analysis. But the colony as a whole is able to solve the geometrical problem of which food source is closest, even though no individual ant knows anything about mathematics! In addition, the colony is goal-oriented: as soon as food is found, more and more ants go after it. From the combination of the simple behaviour of the individual individuals, intelligent behaviour emerges at the (higher) level of the colony. This is also called emergence. Emergence is not as exceptional as it may seem. A traffic jam on the highway is such an emergent phenomenon: it is not found in any individual car, it is a property of (the behaviour of) a group of cars. But then again, emergence is not always the case. The group panic in response to the Damshueller, for example, did consist of a great deal of individual panic. Some phenomena are emergent, others are not.
Rules in use
In lean, we also see emergence. Not in the caricature that is often made of it, in those workshops and large programs that consultants like to advise and sell, but in the real lean - the lean that works, the lean that Toyota developed and that Womack and Jones discovered.
Over twenty years ago Steven Spear In his PhD dissertation for Harvard Business School, he distilled a few simple rules of conduct or patterns from Toyota's routines that lead to a learning organization. Individuals follow simple rules, but the organization as a whole learns as if doping has been used! The rules have a number of important characteristics. I highlight three of them here:
- They embody systems thinking;
- They give the organization self-correcting power;
- They give the organisation a sense of purpose - here: the ability to develop itself further.
1. Systems thinking
A system is something that consists of parts that influence each other. Examples of systems are atoms, the universe, body cells, cities, but also the organization where you work.
If we want to improve an organization, it is not enough to just stick Post-Its® on individual process steps. Systems theory teaches us that systems are more than just their parts (also known as "the whole is more than the sum of the parts"). This used to baffle me: how can 1 + 1 be more than 2? Of course it can't, because 1 + 1 is not a system. A system is something else: besides all parts in every system, those parts are somehow connected to each other. 1 + 1 is not a system, because the ones and twos have no relation to each other. But a family, for example, is indeed more than just a collection of individuals, while 'all people whose first name is the second letter A ' is nothing more than a loose collection of people without any influence. In the same way, an ant colony is more than many ants, you are more than just some body cells, and an organization is more than a collection of loose teams or process steps.
Consequence of systems thinking
Drawing more process steps on a longer piece of brown paper does not make for systems thinking! Because those parts have relationships with each other: they deliver products, services, documents etc. to each other. So we must not only optimise individual teams and process steps, but also design, manage and improve the connections between the teams/processes. A service-level agreement (SLA) emphatically does not do this - it specifies an intended outcome (the service level), but not how exactly the connection should work. The rules for a learning organisation do: they describe how activities, connections and paths over which customer value flows (think 'traceability'!) should be designed, run, and improved.
2. Homeostasis (self-correcting ability)
We are used to thinking of problem solving as an improvement technique, as a transition between two relatively stable, static periods - a transition between the old, problematic way and the new, improved way:
- We analyse the current way in which everything goes wrong;
- With problem solving, we design and test a change;
- If successful, we will introduce the better way of working.
For the record, there's nothing wrong with that, it's a great way to improve.
A different role for problem solving
But the rules use problem solving in a completely different way. A recurring magic word in the rules is 'self-diagnostic'. All processes (teams), connections and flow-paths must be 'self-diagnostic'. This means that literally everything we do is imbued with jidoka and becomes an experiment. Takt, standardized work, continuous flow, supermarkets, and heijunka leveling schemes are not really improvements at all but problem tellers, alarm bells, forms of jidoka, a continuous ongoing experiment. And every missed takt, every exceeded stock level, every unrealized leveling scheme, in short every failed experiment, every refuted hypothesis leads to learning. The key word here is 'each': by seeing 'work' not as a stable period until the next improvement project, but as a literally continuous chain of experiments, every disruption is immediately noticed and remedied.
Not every system has this self-correcting capacity - the tendency to be resilient in the face of disruptions. But in systems that do, systems theory calls it homeostasis. In lean, this takes shape as ubiquitous jidoka including problem solving - but in all work, not in projects.
3. Adaptation (Goal orientation)
Homeostasis prevents us from slipping, but it does not yet ensure progress. There is another element in the rules that ensures this. They prescribe that everything (activities, connections and flow-paths) should be improved. We all try to do that, of course, so it doesn't seem like a big deal. However, the rules also provide a clear definition of 'improvement' by specifying in which direction we must improve. In doing so, they define what is an improvement (in the right direction) and what is not. You could also say: they tell which changes should be held as improvements by specifying what the ultimate goal, 'perfection', means for each activity:
- 100% perfect quality;
- on demand (i.e., only when the next step requires it);
- in batches of one (i.e. piece processing without any changeover time);
- with instant response times (i.e. no waiting for the next step);
- without any waste;
- with physical, emotional and professional safety.
We will never reach this ideal. But we do need it to improve in a clear direction (how many contradictory projects are there in your organization?). Such a step-by-step adaptation towards an ultimate goal of a system is called adaptation in systems theory. In lean, this takes shape as kaizen - but then the real kaizen, not those brainstorms, workshops and suggestion boxes that are called 'kaizen'.
So what are those rules? Obviously you're not holding it in by now 🙂 but I'm going to disappoint you. They are somewhat academic, formal and abstractly formulated. Many terms are used in a very specific, technical meaning and therefore need explanation and clarification. Then I will lose all my readers 🙂 and this column is already way too long! By the way, they are not secret, they have been public for decades and can easily be found online. It's just impossible to understand 'just like that'.
So I won't go into the details here. But when an organisation manages to get the rules into the daily routine of team leaders and middle managers, emergent learning occurs. The organisation as a whole will learn to become much more effective, beyond our pathetic little departmental ants.
In conclusion: lean as a learning organization
Learning organisations (and systems thinking too, for that matter) have been talked about for a long time. Everyone wants to be a learning organisation, but it always remains unclear what exactly we mean by it. Under the guise of "learning organisation" we free up some training budget and buy beautiful scaled agile frameworks, and culture programmes, and big data, and apps, and analytics, and "smart technology", and machine learning, and other promises made by a slightly too confident consultant.
If it seems too good to be true, it is. The toys I mention aren't necessarily evil, but you don't get a learning organization by spending some money.
Toyota's insane, decades-long growth and improvement are rooted in simple patterns of behavior for individuals: the rules. They have led Toyota to become the first and still the best learning organization. "Toyota has learned to learn," as John Shook said in Managing to Learn. I don't care if Toyotas are the best cars or not. I don't care about cars at all. Yawn. Toyota is the organisation that learns best through emergence from simple behaviour, and thus offers its employees a lifetime of inspiring, meaningful work.
Lean, the real lean, is the only working operationalization of 'the learning organization'. Lean is not a process for staff members or feel-good Post-It® workshop consultants. It is an unprecedentedly ingenious way of creating emergent organizational learning from the simple behavior of individuals, via homeostasis (immediate, non-project self-correcting ability) and adaptation (the actual non-workshop cooccurring to perfection).
But then there is work to be done. More is needed than obligatory one-day starts, a kilometre of brown paper and 37 facilitating workshops . Even this column is not enough; the rules are still too abstract to implement 'just like that' (because how do you know if you have implemented them correctly?) and of course this is only a summary. But we have all the pieces of the puzzle. We know how to put systems thinking and homeostasis and adaptation into practice and it is a pity that we are so desperately trying to convince ourselves that we really do it already, "really". We do our best, that's not in question, but what we do is not automatically lean because of our sincere intentions or because we do our best and we label the result as lean. If there's anything worth giving your organization, it's learning to put these rules into practice.
Become an ant colony. And organize the behavior in your organization to find your closest food pile. I'd love to help!