What Can Vietnam’s excellent schools teach us about education quality and equality?

October 3, 2013

Unpacking India’s historic new Food Security law

October 3, 2013

What does your project plan most resemble – baking a cake, landing a rocket on the moon, or raising a child?

October 3, 2013
empty image
empty image

One of the main obstacles to having a decent conversation about the implications of complex systems for how we ‘do’ development


(donorship, programming, advocacy, campaigns etc) is the language itself. Complexity geeks may get a kick out of saying ‘it’s all complex/context specific etc etc’, but more normal/practical people tend to find such language offputting and disempowering. Often, they don’t want to revel in the complexity of the world, they want to know how to do their jobs better. Alan Hudson observed a while back that the most useful discussions on complexity are often those that completely avoid using the word ‘complex’.

So I’m always on the lookout for good metaphors or other ways to convey complexity in more practical ways – step forward Chris Roche in our recent double act in Australia. Chris, who’s been banging on about complexity for decades, divides up ways of ‘doing development’ into three broad categories: baking a cake, landing a rocket on the moon, and raising a child.

Baking a Cake: want to make a cake? Then find a recipe, buy the ingredients, mix, bake and voila! Some cakes are better than others (mine don’t look anything like this one), but the basic approach is fixed, replicable and reasonably reliable. In complexity thinking, this corresponds to a simple system and is the imaginary world of the typical NGO project plan. There are some situations where this is the

man on moon

right approach – distributing emergency aid, vaccinations, bed nets, cataract operations (although even these are never entirely simple), but in the increasingly important messier areas of development – governance, accountability, livelihoods – seeking the right recipes is often either a waste of time or counterproductive.

Landing a rocket on the moon: If you can assemble enough smart people, cash and computer power you can build, fire and land the rocket, and even bring it back again. You need to design the plan from scratch (no recipe here), but if you know the laws of physics, advanced engineering etc you can be pretty confident of success. I’m not sure NGOs do much of this kind of thing – sounds more like a giant infrastructure project funded by the World Bank or DFID. In complexity thinking, this corresponds to a complicated system.

Raising a child: This is what a complex system feels like. If you approach impending parenthood by designing an enormous project plan setting out your activities, assumptions, outputs and outcomes for the next twenty years, the chances are you will be a rubbish parent and your kid might quite possibly end up as some kind of psycho. Raising a child is reflexive, iterative, adapts to the evolving nature of the child and their relationships with you and others, and is most definitely devoid of any ‘right way’ of doing things (despite all those ‘magic bullet’ guides preying on the insecurity of new parents). (Come to think of it, maybe that’s why I have always instinctively

raising a child

phrases like ‘I am doing parenting/childcare this afternoon’, which suggest some kind of blueprint activity). What really helps parents is often advice and reassurance from people who’ve been through it themselves – ‘mentoring’ in aid world. Working in complex systems requires exactly the same kind of iterative, reflexive and agile approach (wonder what the aid equivalent of ‘I’m knackered and need a cup of tea, why not just stick a video on’ is?). And yes, I’m aware that the comparison with parent-child relationships has potential unfortunate colonial/paternalist echoes, but that is most definitely not the point.

The problem in the aid business is of course that while most project plans follow the cake-baking model, many real life situations are much closer to the complex unpredictability of raising a child. If true, we need to think, talk and act very differently.

Everywhere I’ve tried this out, I get nods of recognition – can you suggest any improvements? Also, any ideas for kinds of aid work that fit the fourth quadrant in the Cynefin framework – chaos?500px-Cynefin_framework_Feb_2011


  1. Children simplify reality in a way that is useful. You quickly discover the limits of logic and identify effective (or ineffective) tools of persuasion. These include bribery, withholding, guilt, force, etc. but what many overlook is the fundamental relationship based on trust and respect. It’s not a given, requires patience and takes time to develop, but when you’ve established it (and you know when you do or don’t have it) you don’t really need those other tools.

    Applying this to development, it’s about trust and relationships. It’s hard to develop, if not in part b/c we spend so much time trying to shortcut the process. Children, like development, are ready on their own timelines. Pushing often doesn’t help.

  2. Thanks Duncan. Chris Roche deserves lots of praise for the work he has done in encouraging the development sector to think more about how to incorporate real-world complexity in our thinking. The cake, rockets, child-rearing model however should be credited to Sholom Glouberman and Brenda Zimmerman who used it in a helpful 2002 discussion paper published by the Commision on the Future of Health Care in Canada entitled “Complicated and Complex Systems: What Would Successful Reform of Medicare Look Like?”
    Best wishes
    Keith Johnston

  3. Chaotic systems are like nature itself and one needs a diversity of approaches, and select for the ‘positive deviants’ knowing that most initiatives will fail.

    It is all about how fast ideas and approaches can evolve and adapt, and a diversity of approaches is key.

    This is something we are seeking to do – with our bottom up approach to encouraging innovation and enterprise within the education domain.

    One of the best examples of where this occurs is innovation in the digital technology – no one knows where it is going, but in the clusters and coffee shops of Silicon Valley, new initiatives and start ups keep pouring out with many, many failures and a few spectacular successes, which are well nurtured, amplified and institutionalised.

  4. Chaos = humanitarian aid, fragile states, working in political crises?

    A good metaphor for chaos is the one we have been using in the ongoing DFID project on complex systems tools – fighting fire.

  5. Hi Duncan – There is a framework I have found useful for project management and uncertainty. It was created by Eddie Obeng and he emphasises the difference between knowing/not knowing what you want to do and knowing/not knowing how you are going to do it. In his framework the ‘cake’ approach (when one knows what the problem is and how to respond) is called ‘painting by numbers’. here’s the hyperlink
    He also does a great Ted talk ‘Smart failure for a fast-changing world’

  6. Oh, I like this. First as a parent with the vague idea that listening and responding to my child is a pretty important component of muddling along together. Secondly, as an development worker trying to promote organizational change towards a more joined up, iterative integration of evidence generation and programming. However, the stark reality is that my (our) colleagues in the field face multiple pressures for demonstrating results fast(from donors, employers, government partners) in ways that push them towards either baking the cake or attmepting to squeeze their messy reality into some horrendous moon shot logframe (and leaving out all the bits that can’t be easily meausred as they do so). OK, I admit, in my parenting (at least)recourse to short cuts and authoritarianism somehow also seems to happen, despite best intentions. But back to development work: how can the reward system be changed to encourage more iterative approaches? Showcasing those making the effort? Accepting well-documented and analyzed failure as a positive result (would that international development had the equivalent of “Failure Fairs”)? Genuine partnership, involving longer term relationships and more predictable funding, with some privilege given to practical analysis — this being conceived as “doing” as much as “doing” is?

  7. Duncan,

    Discussions like this are always challenging, stimulating and valuable. However… “distributing humanitarian aid” is “fixed, reliable and reasonably replicable” – surely not. The caricature that humanitarianism is as easy as handing out sweets to children was outdated a long time ago.

    Humanitarian crises take place in the messiest, most brutal, life and right-denying places in the world and those who seek to offer assistance are faced with appalling physical, political and moral dilemmas -and a dilemma will always be a choice of two bad options. The instability, and chaos of crises in Syria, DRC or Baluchistan stretch our understanding of complexity to the limit.

  8. You mean how aid agencies implement programming or projects, right? Not that you think some other part of the world population is your ‘child’ to be taught and mentored by their parents, right?

    Western aid agency staff using a child metaphor to describe them ‘doing development’ might require some more careful caveats and nuance if it’s not to be offensive. I’ll be surprised if this doesn’t appear to some people to be a bit paternalistic and arrogant.

    1. Kevin, which bit of ‘And yes, I’m aware that the comparison with parent-child relationships has potential unfortunate colonial/paternalist echoes, but that is most definitely not the point.’ is not clear enough on this?

  9. Nice metaphors, and you could have fun with the child-rearing one by introducing nannies/au-pairs/domestic help. Now who in the Aid Chain corresponds to them?

    I’ve used driving as a metaphor, especially in talking about learning vs training, in response to complexity. Being trained in the simple set of tasks, learnable patterns, involved in reversing into a parking space, or getting into fourth gear. Driving when there is nobody else around is simple, once those skills are learned. Driving when there are other people, 100s of them, on the same roads, each with agency and intention, requires extraordinary levels of awareness, learning about evaluating risk and opportunity, flexibility and speed in response.,, creativity (not too much, perhaps) …and so on… and all in the blinking of an eye.

  10. Well, to bake the kind of cake you show here, will require some “parenting”.
    I don’t think the parenting analogy should over-philosophised, if you listen to developing states you will hear the voice of governments and organisations wanting mentorship, for complex and complicated contexts, not really for the simple.
    I believe previous blogs have painted a fairy good picture of the 4th context, most recently the example of Barclays and Somalia.

  11. I agree that telling people their problem is complex (if they don’t know it already) can feel disempowering to them without providing tools. I’m a fan of Patton’s Developmental Approach and have personally relied a lot on Eoyang’s conditions for helping systems self-organize around complex issues. Eoyang’s systems terminology is a bit clunky (containers, differences, exchanges). But I’ve found community workers, after fairly brief training on the concepts with some concrete examples, find it very useful for critiquing their own processes and working toward complex problems.

    In brief (sort of), community workers use the framework to identify Containers: boundaries where self-organization happens, whether physical (e.g., school), ideological (e.g., shared goals), or social (e.g., family) and learn when they need new ones or change current ones; identify Differences (in perspectives, roles, power, knowledge) to know which stakeholders are missing, getting heard, or could provide access to new resources; and identify Exchanges (relationships and communication patterns that control how information/resources are transmitted or communicated) to learn where they need to form new relationships or communication.

    Even non systems thinkers quickly learn to critique their own systems, identify missing voices or resources and understand that a shared goal is necessary to establish. It helps them to ensure their system continues self-organizing as the complex problem shifts.

  12. I am getting a bit tired of this oft-recycled metaphor.
    Lets boil it down to its essentials and see what is missing:
    1. Cake: 1 person working with material objects (which don’t have any agency)
    2. Rocket: Many people working with material objects (which don’t have any agency)
    3. Raising a child: One (or two) people) working with a child (that does have agency)
    4. Most development projects: Many people working with many people (who do have agency). This is where we get serious complexity, and the extent of which is heavily influence by their network structure i.e. how they are connected to each other. As they might say, “Its the connections, stupid” (especially when looking at groups of people, rather than material objects).The structure of connections makes a difference between the sort of changes that take place, between stable states, complex or chaotic behavior.

  13. Hi Duncan, I like the analysis. I believe mentorship can be devoid of paternalism, and produce good products. Many see iterative work as risky, and only risky donorship subscribe to it, but we are learning our lessons, and while the ‘safe play’ yields ‘chaotic’, the risk takers and mentors are busy celebrating the ‘simple’.

  14. I agree that the cake-making and moon-landing models are both inadequate and yet are what development imagines itself doing (particularly cake-making), but I’m with Rick Davies on the limitations of the metaphors to accurately describe the situation. Pete is closer with his driving analogy, which also puts everyone involved on an equal footing. But another suggestion is that of a meadow – what Kevin Kelly (author of Out of Control) would call a “self-sustaining system” with many variables. In all the 3 models in the blog, the development wallah is ‘doing unto others’. In a meadow analogy, the development wallah is part of the system, not as a farmer with hose or scythe but as an element of the life system, doing a bit of pollinating here, putting out roots there and hoping the other grasses, insects and gasses will respond positively.

  15. Oops. Well, every bit of that’s very clear Duncan. Obviously I need to be read more careful before leaving public comments on blogs!

  16. Most seem to understand that chaotic situations (riots, spontaneous uprisings, loss of stable political/social structures) are by definition unpredictable situations to which organizations mostly can only react. However, from a development perspective, the important take away from chaotic situations is that order starts setting in quickly as the diverse elements self-organize. Fractions and alliances form, people unite around certain groups or goals etc. So chaotic situations have enormous potential to promote new goals or alliances around which people can unite and to quickly become stakeholders in the naturally forming structures/ alliances.

  17. Bit late coming to this post , but NGO’s have themselves partly to blame by simplfying their message in their fundraising appeals and that’s all most people know of development. We want to make it immediate and baking like. Not much mention of complexity or risk.

  18. I’ve found it more useful to use these metaphors in terms of classifying ASPECTS of interventions rather than whole interventions. It’s helpful to identify which aspects of a program or policy can be treated as simple (even if they’re not exactly so), and which as merely complicated. Otherwise it’s too easy to classify everything as complex, which can be quite daunting.

    (I also embellish the metaphors – simple is not about baking a cake but baking a packet mix cake – just add water. Complex is not just raising a child, but raising a second child – where the rules about the first child don’t apply…)

Leave a comment

Translate »