Full stack decision making
A few years ago now I did a lot of work with a number of police forces and I was always struck by the fact that the best of them were very comfortable with the fact that while community policing needs to be fluid and networked any kind of crisis policing needs to strictly hierarchical with a strong chain of command. I am very ok with that and so was everyone involved. However passionate an officer was about ‘digital’ (we were still calling it that then….) they all agreed that once you have people running around with guns its best to have someone actually in charge.
What this highlights is that how we take decisions needs to not only reflect the content of the decision but also the context in which is made. In that policing world you might expect a Community Police Officer to be tweeting with a high degree of autonomy but once Gold command is in place then they would expect to be part of a much more structured communication process. I think this is a principle we can usefully apply within organisations. Think about the context as well as the content of the decision when you work out how to make it.
Outside of a crisis or incident context I find the democracy stack a useful tool for figuring out how to structure decision making. The stack was something I came up with to explain how you can think about democracy in a system context and also in response to the literature which is so often making a ‘either/or’ of different democratic forms. At the time I didn’t really think about the effect that context makes on the form of decision making but we are in the middle of an ongoing piece of work at CRUK which is thinking deeply about how you apply the principle of agile first to our governance and the example about policing came to the front of my mind.
The stack is fairly simple and knits together three main forms of democracy:
- participative democracy which is what you see either in participatory budgeting (good!) or referenda (bad!) but is essentially one person one vote. It is most effective within highly connected groups which have high levels of knowledge about each others wants and needs (hence the ‘bad’ when applied to national referenda)
- deliberative democracy is less well known but has started to be used more in the form of things like the citizen’s assembly in Eire which shaped both the abortion and gay marriage referenda. In deliberative democracy a random but representative selection of people are asked to become informed, debate and then make decisions on something which effects the wider system. Its hugely powerful and its exciting to see it applied to complex issues like social care here in the UK
- representative democracy has different forms but its what we tend to think of as democracy — a single representative is selected and then represents a group of people on a whole range of issues within the system. The range is defined by their sphere or democracy — ie local or national
Centuries of work has gone into thinking about these different democratic forms and how you balance the rights and agency of the individual with the need to actually take effective decisions — and this is key — there is no point in having a decision making mechanism that doesn’t actually help you reach a decision and for most people its better to have a well made decision you don’t agree with than no decision at all. Techno-evangelists are still looking for ways in which technology ‘fixes’ democracy and enables participatory decision making at scale but this ignores the hugely important factor of trust and the fact that these decisions are taken within living, breathing, human systems. Decisions taken by humans for humans.
I find democratic theory more helpful than organisational theory when thinking about decision making in the context of a networked organisational because at its heart is this concept of an individual with agency and self-efficacy — the human and not the interchangeable Industrial Age ‘resource’ that underlies so much of our management theory. I know that many organisations are still designed around that hierarchy but if your goal is to end up with an organisation that is less silo’d at the same time as being more collaborative, adaptive and flexible it seems sensible to look to the thinking which is designed to support a more sophisticated view of decision making then that of a hierarchy where things get rolled up and then down the hill to get an decision.
The point of the democracy stack is to help work out what impact the decision you are taking has and shape your decision making approach accordingly. In an organisational context the stack translates to mean:
- Participatory layer: The question is highly local and only effects the people in that team
- Deliberative layer: the question is system wide and actually requires tradeoffs between teams
- Representative layer: the question is global and requires a change to the operating context or conditions of the organisation.
The important thing about the democracy stack is the fact that it radiates out from the empowered individual — you can’t set the boundaries at the global and cut them down to fit the team — you need to set the system up to grow from the individual outwards. The heart of the stack is participatory decision making — that is where it grows from — but it accepts that this does not scale to system level.
While we might use different methods for different types of questions this works best when you do full stack decision making. If you just take the representative layer you are back in the world of HIPPO based decisions with the highest paid person taking the decision (hello NHS leadership….). If you get stuck in the deliberative layer then you end up in decision paralysis as you can never effect the actual work or change the global context and if you make everything participatory you end up in a round of endless negotiation.
Netflix are often given as the example of what it means to deeply embed participatory decision making which has embedded in it huge agency for individuals as well as the idea that you need protocols and guidance to avoid keep overwriting each other decisions (if you have not read the netflix culture overview you should) but its the fact that they apply innovative thinking to their global governance that makes them revolutionary. Netflix redesigned their board meetings to emphasise the need to inform and upskill the leadership as a way of making sure that they make good quality decisions in the realms where there skills are most appropriate. Throughout the Netflix model you see high quality decision making at different levels all of which rely on trust and information exchange. There is no passing up the food chain — its about an active conversation within a dynamic system which senior decision makers participate in rather than simply try to control. Netflix are applying full stack decision making.
It often feels unfair to apply the learnings of an organisation that is both digitally native and also digitally focused — essentially a green field site in the network society to organisations which were born in the Industrial Age but as you start to see agile practices becoming normal and those highly participatory teams ready to start operating at organisational scale its time to do so. We undoubtedly need to make sure that we are rethinking our board rooms and senior forums to better resemble what Netflix are doing but I want to argue for a minute about the need to put our energy into building the deliberative layer with our organisations.
In the organisational change domain we spend a lot of time talking about how difficult it is the change middle management. How about instead of criticising them for doing what we have asked them to in order to maintain the status quo we gave them more power to actively manage the tradeoffs between the competing needs from their teams. How about we give then the space and empower them to deliberate? But lets make it explicit — they are not simply advocating for the answer that best suits their team they are deliberating to find the best solution for the system. This could be shifting budget decision from senior to middle managers or it could be about formalising the role of these groups in work prioritisation or standards setting. You probably can’t do this all at once because you have almost certainly spent the last decade training your middle managers to be cogs in the industrial machine but its where you should be heading. And if you don’t then they will start doing it without you — forget shadow IT — how much shadow decision making do you have going on in your organisation?
Information is key to high quality decision making and one of the great revelations of the agile mindset (as with participatory democracy ) is that decisions need to be taken where the best quality information is. Agile teams have long advocated for the need to empower them to make their own decisions and thats right — but only in the context of a system that also empowers people to effectively deliberate the tradeoffs and also provides an effective mechanism for setting the global context. Each layer in the democracy stack is expert in different aspects of the system, bound together with both the social networks which operate between these layers but also the flow of information — its about the right decision being taken in the right way and with the right information. One of the big adjustments that this approach is the need for agile teams to make their decision making more transparent and a need for is all to rethink organisational knowledge management in the context of a living breathing decision making system and not simply stored documents and KPIs (have a read of this Stanford Social Innovation Review article if you want a taster of what this could mean).
Lets try an example: A decision about a new UX feature is undoubtably better taken at the participatory level by the team who is closest to the user. But that feature should then go into a pattern library and if it gets outcompeted by a more effective feature — as defined at the deliberative layer which looks at what good looks like across the system — the system would expect that team to update that features. However a decision about a whole new class of user would need to be taken in the representative layer as this would change the context of the organisation. Does that work? I think it could.
A more familiar example to democracy types; a decision about what a community need to make it easier to connect to other people locally is best taken exactly there — locally. A decision about how to shape a shared space like a town centre needs to be done deliberatively. Decisions about national infrastructure need a representative approach.
The democracy stack is a theoretical model and we are some way away from full stack democracy but applying the lens of system thinking to something that has previously been an either/or discussion is important and helps theory move into a more practical realm. Theory helps us see our way through the messiness of the real world and while its never going to be achieved it provides the scaffolding for change that is needed if you are pursing a paradigm change. In an organisation context the democracy stack is a way of thinking about how we categorise decisions in a way which means we do have the right people taking the right decision at the right time.
And circling back to the policing example and also the Netflix intent to ‘avoid rules’; be prepared for the democratic system that you build to start making sure that your decision making process adapts to the context of the decision.