Dave Snowden

Scaling: context is all

RSS Feed

In my second post on scaling I talked about about two types of scaling and the last two posts handled the issue of scaling up within the enterprise.  The essence of yesterday's post was that we need to allow cluster patterns to emerge from finely grained objects in three areas: strategy, capability and needs.  Each of those patterns allows a new emergent potentiality to shift to realisation.  This evolutionary approach is all about sustainability and resilience, designing capability not product.  In this post I want to pick up the other type of scaling, namely this worked how do I get more of it type of scaling.  

It is worth repeating that a complex adaptive system is neither aggregative nor reductionist in nature so assembling processes into complicated control structures is not the way to go.  This basic error, exemplified by SAFe in software development is wonderfully satirised in this post.  We saw something similar in sick stigma,  BPR as a method had reached its limits and was (and is) failing in its application to services.  Such techniques work in closed systems but not in open systems.  The problem is that some people have been seduced into perversion by the attractions of order and can't cope with this limit.  So instead of recognising that they have something useful within boundaries they take the lets do it harder with more structure, more rules and more enforcement approach; in the case of sick stigma we also get the cult of belts!  SAFe is even worse as there is a perfectly good body of work for waterfall development in PRINCE2 and the like which all have some foundation and are not just slick marketing operations.

Either way, minor rant aside, lets get back to scaling success.   Now this is a common issue in development and government in general.  Something works and everyone gets excited and wants more of the same.  Being a little cynical I might say that the desire is in part driven by the general lack of success creating a culture of getting excited by pearls but forgetting the number of oysters required to produce them and the health consequences to pearl divers, let alone the oysters!  We saw it with the Grameen Bank the first well known case of micro-lending.  It was a success, but it was the end point of a decade plus evolutionary process in a specific cultural context.  Over time people found that replication in other cultural contexts was problematic to say the least and even in the sub-continent the loan sharks and banks found ways to incorporate the vision into their profit orientated structures.

The context word is key here, the success of things in a complex adaptive system is the result of a unique evolutionary process which can never be replicated exactly.  The success is also in general self-reported (particularly during the critical early years of said process). You can't replicate the end point of an evolutionary process, which by its success has changed the context anyway.  However we can seek to copy the starting conditions and enable the emergence of new contextually unique successes that learn from it.   If we are really clever we we will also ensure imperfect copying in order to enable and enact learning.   To do that we need to understand what can be replicated or copied, and what measures we need to use.   That will be the subject of tomorrow's final post in this series.

PS: I' pleased to say that the course on Friday which will cover all of this in an Agile context sold its minimum numbers yesterday, the first day it went on sale.   If you want to come the link is on yesterday's post.

Top