It’s complex, not complicated
I’ve been working in government for four years now. I’ve been lucky enough to work across several roles, getting to experience digital service design, policy design, and data strategy, working in both delivery and leadership teams, and engaging with brand new initiatives (‘greenfield’) as well as ageing and entrenched (‘legacy’) systems and technology. I’ve been in leadership roles, and I’ve observed other leaders working across small, medium, and massive teams. As a result, I’ve already gotten to experience quite a wide range of what central government has to deal with, and how we do it. It seems like a good time for me to reflect on all that.
I’m really hoping these posts will lead to some more discussion of how we can do better. So if any of this sparks any thoughts for you, please share them — in comments to this post, on Twitter (I’m @jallen300), or anywhere else that works for you.
I’ve gathered my thoughts into three posts:
- The first (this one) explains how we often apply the wrong solutions because we don’t understand the difference between issues that are ‘complex’ and those that are merely ‘complicated’.
- The second talks about how we can make better policy and services by linking policy and delivery work together in multi-disciplinary teams.
- And the third talks about what I think good leadership looks like. Because without good leadership, none of the rest matters.
So first, let’s talk about complexity. Three main points:
- We need to recognise the difference between problems that are ‘complex’ and those that are just ‘complicated’, and acknowledge that we can’t tackle complex problems with the approaches we use for complicated ones
- The complex problems government addresses will never be fully ‘solved’, and so government’s work on any given issue will never be ‘finished’
- ‘Agile’ and ‘systems thinking’ are not fads, they are responses to these fundamental facts about the world, which have been overlooked for too long
1. We need to recognise the difference between problems that are ‘complex’ and those that are just ‘complicated’, and acknowledge that we can’t tackle complex problems with the approaches we use for complicated ones
I think most of the insufficiencies of government can be traced back to a fundamental misunderstanding about how to ‘solve’ the problems society faces.
As humans, we seem to be hard-wired to see the world in terms of discrete problems we can solve. Our brains are always breaking things down to their component parts, looking for ways to fix them. We are constantly seeking greater and greater simplicity. The world around us is so complex, that’s only natural — we’re trying to get our tiny heads around infinitely large (and complex) things.
And that works for math problems, or engineering problems. You solve the component parts of the problem — no matter how many there are — until the whole fits together perfectly. We can do that to make a concrete block, or to build a wall from those blocks, or to build a prison from those walls.
But building a prison system that keeps people safe and creates a healthier society — that is another matter. It’s not simply a question of compiling all the right parts and assembling them in the right way. A prison system — one that ensures safety and promotes rehabilitation — has an infinite number of component elements, all of which are interrelated and continuously changing, often in unpredictable ways. That is what we call a ‘complex’ (or ‘wicked’) problem.
The vast majority of problems government addresses are these kinds of ‘complex’ problems. How to reduce crime. How to educate our children. How to increase prosperity and keep people healthy. There is no one right answer to these problems — there are just better or worse situations we find ourselves in.
Complexity theory tells us that the way to tackle complex problems is completely different from how we tackle discrete problems with a finite number of variables. For those kinds of problems — even the most complicated ones — we bring in the experts and let them analyse the situation and draw up a blueprint for how to best fix it. Think of a skyscraper — you hire an architect to map out exactly how everything will fit together, and then building engineers with expertise in each step of the process. But in that case, you are able to map out exactly what the end solution will look like from the beginning. The ground is not shifting beneath you.
In complex situations, however, there are no tried and tested answers and everything keeps changing. This is what architect Ann Pendleton-Jullian and scientist John Seely Brown call living “in a white water world”. So we must tackle those challenges differently, through a process of probing, sensing how things change, and reacting to those changes — in other words, a process of quick, continuous, and never-ending experimentation. You want to reduce crime and make communities safer in this particular society at this particular moment in its cultural, economic, and political history… you want to educate young people more effectively, improve people’s health, make work more rewarding… there is no blueprint for any of these, the ‘answers’ will only be found through continuous rapid experimentation.
And I put ‘answers’ in quotation marks because we must get comfortable with uncertainty. In fact there are no right or wrong answers — in complex situations we are constantly operating in uncertain waters. But what we can do is stubbornly work to improve things, incrementally and continuously, by experimenting with approaches that are safe to fail but designed to teach us what can work at that moment, in those circumstances.
Business consultants David Snowden and Mary Boone say that in situations of flux and unpredictability, where there are no right answers, unknown unknowns, and many competing ideas, the only useful approach is to experiment in ways that probe the system, then sense changes in the environment and respond proactively to those changes.
Unfortunately, I think we’re currently taking the wrong approach to solving most of the issues we’re faced with. We rely on experts even when there’s no blueprint they could possibly create for us — they do their best, but the domain of uncertainty is not the domain of experts.
We’re using a hammer to bang in a screw, and we don’t understand why it won’t stay in. We need to acknowledge that first, so we can start to change the way government institutions function, enabling them to more effectively tackle the complex problems society faces.
And we also need to recognise that…
2. The complex problems government addresses will never be fully ‘solved’, and so government’s work on any given issue will never be ‘finished’
Since there are no right or wrong answers to complex problems, we must acknowledge that what we are trying to achieve is not solutions, it’s continuous improvement.
Working in the prison system, our goals are to improve the safety and security of prisons, protect the public from harm, reduce re-offending, and deliver swift access to justice. None of these problems will ever be ‘solved’ — they are not items to tick off a to-do list, and then move on. All the projects our teams work on, the services we design and deliver, the policy decisions we make — they’re all trying to move us as a society just a little bit closer to those goals.
A quick example…
During the first wave of COVID in early-2020, our teams were asked to implement on a wide scale the prison video-call service we had been testing in a few prisons. After understanding how the service could function, identifying what technologies existed, going through procurement exercises, and building out and testing the online and offline aspects of the service, we had to then tackle complicated technological and policy questions around cost and security. Now that we’ve fully rolled out the service to all 100+ prisons and ironed out many of the kinks, other questions will surely arise.
- Have we got the timings right for the service — who’s inadvertently been excluded by the rules we’ve set for the service? Are those rules keeping the service from having the impact we’d hoped?
- How might we use the service to achieve wider goals around safety, building people’s skills, and helping secure accommodation before people are released from prison?
- And as technology changes in the future, will our cost and security models need to change too?
Policy and service design questions will continue to arise, as long as the service exists and the policy objective (to connect people in prison with people in the community) persists. The service will never be ‘finished’. The policy objective will never be completely ‘achieved’.
And remember, all of this work is simply improving how people in prison connect with their loved ones, or access training or other support. It doesn’t ‘solve’ the larger problems of how to create safe and secure prisons, or reduce re-offending — it’s just a contributing factor. That wider work will also continue, forever. We need teams that are in it for the long haul — managing (or ‘stewarding’) systems to continuously perform more efficiently and achieve society’s goals. (More on what those teams could look like in the next post.)
So with all that in mind, ultimately, I firmly believe that…
3. ‘Agile’ and ‘systems thinking’ are not fads, they are responses to these fundamental facts about the world, which have been overlooked for too long
The only way to achieve the goals we set for teams tackling complex problems is to develop a deep understanding of the issues, test improvement ideas, learn from those tests, and adapt our approaches — continuously.
And those processes must be done quickly, because if we wait months or years to see the impacts of our changes, it will be impossible to disentangle the impact of our changes from everything else that’s happened around them.
That is why policy and service design teams that see the whole system and work in agile ways to improve it — and organisations that adopt this agile mindset of continuously learning and adapting and improving, and design their org structure to enable that process — will be more successful than those that tend to see the world in terms of discrete problems to be solved.
Organisations that acknowledge the uncertainty they’re operating within and humbly work to make things better day by day, year by year, and decade by decade are the ones that will, over the long run, help us achieve the results we all want to see in society — safety, prosperity, and opportunity for everyone.
That’s not to say that agile ways of working and systemic design approaches are definitely going to take root across government — I’m very well aware that they might not get widely adopted anytime soon. If that’s the case, though, it won’t be because they were just a fad. I think it will most likely be because senior leaders will (1) believe it’s too risky to publicly acknowledge the uncertainty and the insolvability of the complex issues their institutions are aiming to ‘solve’, and (2) find it too difficult to transform the way their institutions function so that their teams can pursue objectives through experiments that are safe to fail but designed to teach. (I’ll talk more about how that can work in the next post.)
Let’s hope the work of the many policy design exemplar teams and projects across government helps senior leaders understand what kind of structural and cultural transformation is required, and provides the evidence they need to rise to the moment and make that transformation a reality.
—
Apologies for the extra-long post, but, unfortunately, changing the way government functions is not the most straightforward thing to do. Please do share any thoughts about any of these points, or additions of your own, in the comments below. Or tweet to me @jallen300.
And if, after all that, you’re still interested in hearing my thoughts about how we can make better policy and design and deliver better services by linking policy and delivery work together in multi-disciplinary teams who spend more time in the real world, that post is here. Or jump straight into the final post about what good leadership looks like.
