In the first article in this series, I looked at feedback loops as a way of both understanding and influencing how systems resist or adopt change. But that’s only one of the things shaping how complex systems behave.
Donella Meadows is one of the biggest names in Systems Thinking. She set out a hierarchy of places you can intervene in a system. When it comes to effectiveness, feedback loops sit somewhere in the middle1. Let’s take a look at the other opportunities, or levers, to make an impact.
Beware of what seems obvious
Before we get to the most effective levers, it’s worth saying that the interventions we instinctively think will have the biggest influence are usually the weakest. Worse, they can backfire when we don’t fully understand how the system behaves, which is how plenty of well-meaning initiatives end up achieving the opposite of what we wanted.
Some of the most striking evidence for this comes from studies of physical and structural environments.
Take Bernstein and Turban’s 2018 study of open-plan offices, which followed two large corporate headquarters before and after their switch from more traditional office layouts. In both, the stated aim was more collaboration.
Using wearable sensors and electronic communication data, they could measure face-to-face and digital interaction directly. The results were surprising: face-to-face interaction dropped by around 70%, with email and instant messaging rising to match – the opposite of what was intended2.
Their explanation was that when colleagues can see you all day, you protect your privacy by withdrawing socially: headphones on, eye contact avoided, conversations moved to chat. A change meant to boost collaboration measurably reduced it.
It sounds counter-intuitive, but this kind of thing happens all the time in complex systems. It’s why social problems like homelessness, illegal immigration and drug addiction are so stubborn (and why we should be wary of anyone who promises easy answers).
Which levers work?
From Meadows’ excellent hierarchy of levers3, Let’s look at some of the most impactful for change programmes. I’ve picked out the four that best support new behaviours that last once the programme ends. In order of effectiveness:
- Mindset/paradigm – the answer to the question, “why do we exist?” at the organisational and team level, as well as our implicit and explicit assumptions and values.
- Goals – “what do we need to achieve?”
- Rules (incentives, punishments, constraints)
- Access to information on how well the system is achieving its goals (the KPIs, OKRs, dashboards, reports and status updates people can actually see)
Working at these levels beats things like adding roles or making physical adjustments to the system such as rearranging the office or spending on new tools and comms channels communications channels and so on). It’s why the open-plan example went so badly: they changed the furniture without ever showing what ‘greater collaboration’ was meant to achieve, or why. None of which means ignoring the other levers – just that the ones above will make more difference, particularly if they’re currently working against you.
| Practical example Information Lifecycle Management (ILM) Data and information sprawl is a classic system-level dynamic. Before cloud storage, most companies were limited by on-premises capacity and the cost of maintaining it, so all the new data employees created was balanced out by deletion, usually done by hand. Email inboxes filled up, staff got an alert, and they deleted what they didn’t need. As organisations moved to the cloud, often with allowances that felt limitless at the time, people stopped bothering – if nothing pushes you to delete anything, why not keep it ‘just in case’? That ‘just in case’ mindset is now thoroughly embedded, especially in the public sector (who were often given cheaper or even free storage by the likes of Microsoft and Google). I’ve watched this play out everywhere from large government departments to universities. Tooling helps, but only alongside the other items on the list above: information about the hidden costs (money, carbon, wasted time, risk and reputational damage), rules (either ILM rules built into the software, or incentives/rewards built into the performance management process), goals (linking the risks to what the organisation actually needs to achieve) and a mindset shift (data isn’t ‘free’, and it still has to live somewhere). |
Acknowledging that systems are not simple
We’re all drawn to simplifying things. We want tidy stories that explain why things happen, and predictions that make sense to us. But systems are complicated, and they can respond to change unpredictably. I’ve worked in this field for longer than I’d like to say, and I still get caught out. But one thing stays the same: people setting off on a change programme underestimate the challenge and oversimplify the process. Just as people share broad traits and tendencies while remaining endlessly complicated, every organisation has its own quirks. No model or framework is going to capture all of that and map an action neatly onto an outcome.
The best we can do is turn up with curiosity, humility and empathy. That last one – treating people kindly and with respect – might be the most reliable lever we’ve got.
Checklist
- Systems Thinking suggests there is a hierarchy of leverage points within systems.
- The highest-impact leverage points are rarely the ones we think.
- The most effective ones work at the level of mindset, goals, rules and how easily people can see how the system is performing against its goals.
Footnotes
- Meadows, D. H. (1999) ‘Leverage points: Places to intervene in a system’, The Sustainability Institute (working paper); Meadows, D. H. (2008) Thinking in Systems: A Primer. White River Junction, VT: Chelsea Green. ↩︎
- Bernstein, E. S. & Turban, S. (2018) ‘The impact of the ‘open’ workspace on human collaboration’, Philosophical Transactions of the Royal Society B, 373(1753). ↩︎
- Meadows, D. H. (2008) Thinking in Systems: A Primer. White River Junction, VT: Chelsea Green. ↩︎
