What’s the right structure for a large organisation?
What’s the right structure for a large organisation? There have been a few interesting blog posts about this in recent days, looking specifically at Apple and whether it is equipped to support its own scale.
It’s a topic that I started to tackle a few years ago with Enfield Borough Council when the chief exec asked me what a future-ready council should look like. I tackled it again when a £250m turnover logistics business asked me to help them to be future-ready. Neither of these businesses is Apple-scale. But I think the answer holds as you scale up.
My answer in both cases was similar. It starts with perspective. Every org chart is created from the leader’s perspective, and so has the leader at the top or in the middle. But we know that the most important person in any organisation is not its leader. It’s the customer.
Reorganising the businesses around the customer gave us a model of concentric circles: a common interface layer, a common network layer for data, and a common external interface for partners. This, in turn, led to a clear need to define the interfaces at each of these layers very carefully to minimise friction — another way of maximising the accessibility that Steve Yegge talks about in his famous platforms rant.
Where things get really interesting is when you diminish the friction not just between layers in this model, but between different services and functions.
Functional vs Divisional
In the classic org models, you would divide the business up by functions (HR, finance, sales, marketing) or by business line (product A, product B, product C). Most companies do a bit of both. The attempts to rationalise this hybrid often end up with a matrix: people have a functional reporting line (more senior people of their discipline) and a divisional reporting line (more senior people responsible for the delivery of products or services).
In our model, we proposed more of a service model. Part of this fits with the description of Amazon’s data-driven approach as Yegge describes: there are lots of interfaces that are currently based on high-friction, human-to-human interaction where little would be lost and lots gained by changing them to low-friction digital interactions. But there are also interactions where the human-to-human component is absolutely vital: creative, decision-making, or complex communication. Here we often suggested more of an agency/client approach.
Such relationships in the past may have been very high friction, hence the perceived need to co-locate people or set out the org chart to define hard connections. But technology has enabled very low friction interactions — e.g. shared documents, instant messaging — that supplement the richer face-to-face interactions that these relationships require.
Define interfaces, not connections
If you focus on defining the interfaces between units in an organisation and then optimise those interfaces, you can map it much more easily and more importantly, maintain that mapping. You can see through the organisation and understand value flows and interactions. You can keep goals coherent.
In 2017 I hope to expand further on this premise — the basis of our Stratification model. Perhaps through research and perhaps through direct application to an organisation.