Thinking in systems

This is probably THE book to read this year. It invites you to step back from problems and their solutions (or pseudo-solutions), and it reminds us that the stakes are the same at the scale of an individual, a team, a company or something far more global — technical, political, economic or otherwise.

The book in (too) few words

Donella Meadows invites us to see systems as triplets: elements, interactions between them, and a systemic purpose. Such a system behaves according to stocks (an accumulation of...) and flows between them (a movement of...), and it is governed by feedback loops — balancing ones, which stabilise certain behaviours or stocks, and reinforcing ones, which amplify behaviours, or make stocks grow or shrink, sometimes exponentially.

From the thermostat in our homes to natural ecosystems, by way of economics or the regulation of drug addiction, the book teems with examples that are easy to grasp and relevant to the model it proposes.

After a first exposition of the way of thinking (without a single line of mathematics, don't worry) and a guided tour of fairly simple systems, the author explains why our systems can prove so stable or so stubborn, whether the stability is wanted (temperature regulation, say) or is sheer obstinacy (our inability to break an addiction, say). She points out the traps and opportunities we face with systems, and the more or less effective leverage points for making them evolve towards the behaviours we want.

What we made of it

You will have gathered that I am rather enthusiastic. From my engineering studies to my PhD thesis (see below), I think I can say I had already been thoroughly exposed to systems thinking. That did not stop me from enjoying the exposition and learning a fair amount from it.

The systems thinking Meadows proposes is both simple and broadly applicable. In my view it would do a great deal of good in school curricula, in the final year of secondary education for instance. Failing that, it should be a compulsory course at university, and certainly a mandatory step for every decision maker, economic and political among others. Nothing less.

I cannot help looking for the similarities and differences between Meadows' approach to systems and KAOS (Keep All Objectives Satisfied), a systemic method for analysing software systems that was taught for a while at master's level at UCLouvain, and to which I contributed in my PhD thesis.

Although they share a common basis — a system is made of agents interacting to reach objectives — the two approaches are quite different. Flows and stocks, for instance, are not explicit at all in KAOS. Conversely, systemic refinement through obstacle analysis, one of KAOS' great strengths, which consists in improving a system on the basis of the behaviours you do not want it to exhibit, is not as well guided in Meadows. The comparison is not easy, though, and arguably not very relevant, since KAOS belongs to the realm of discrete and rather linear analysis, whereas Meadows' systems thinking applies above all to continuous, non-linear systems.

Among the other fundamental differences I appreciated: Meadows proposes to see a system's high-level purpose in the de facto resultant of its behaviours. A descriptive approach, somewhere at the opposite end from KAOS, where one seeks instead, prescriptively, to decompose the desired high-level objectives of a system into expected behaviours (of humans) or required ones (of hardware and software).

This proposal of the author's is firmly anchored in reality, and it speaks to me personally on the societal, entrepreneurial and ecological level. I paraphrase her:

The best way to know the purpose of a system is to watch it long enough. The bad way is to listen to what its actors say about it.

Finally, we know the author is behind "The Limits to Growth" (Club of Rome), much discussed these days in the face of the climate and ecological emergency. I personally appreciated seeing the question addressed outside the economic sphere, where it is all too open to debate. I hope reading the book will let a good number of people discover that this line of thinking is less a highly contestable political opinion than a scientific observation about complex systems. (Contestable too, of course, since science is only the state of our knowledge at a given moment. But it is only seriously contestable insofar as a reliable counter-model is then put forward.)

And where does Klaro Cards fit in?

Klaro Cards may let me bring these two complementary views of systems closer together in the future. The software is inspired by:

• relational databases, and in particular by the uniformity and neutrality they use in representing information: a piece of digital information is a fact about the world that digitalisation seeks to represent.

• KAOS, which formalises the conditions necessary for agents to collaborate effectively — conditions on the availability and control of information, precisely.

• Lean, which seeks among other things to optimise work by decomposing and managing flows and stocks (material and immaterial).

It is Lean thinking that is a priori closest to Meadows' systems thinking. In Klaro terms:

• Each Klaro board permanently displays a subset of cards, hence of facts about the (digitalised) world. For example, the features still to be developed, the tasks to prioritise, the client files to discuss at the next management meeting. That is a stock, no more and no less.

• Beyond the refinement of the information itself (prioritising a list of ideas, for instance), moving cards in Klaro automatically creates flows: cards move magically from board to board, exposing the necessary information to whoever needs it.

• The very existence of boards, and the ability to create them on the fly to expose the right information to the right person at the right time, is nothing other than a regulation mechanism: it aims to guarantee the collective intelligence of the whole system, through a high degree of transparency and an objectification of the (digitalised) real.

So it remains for me to see which traps, which opportunities and which leverage points proposed by Donella Meadows could lead to new features in Klaro Cards, to improve our projects and teams. Plenty of work ahead :-)

Note: this post was originally written in French about the French translation published by Éditions Rue de l'échiquier in 2023. English readers will want the original: Donella H. Meadows, "Thinking in Systems: A Primer", Chelsea Green Publishing, 2008.