I've been having a lot of conversations, both academic and practitioner, about the distinction between these two words over the last few months in the context of strategy, risk and sustainability. Traditional approaches to managing future levels of uncertainty has been to focus on a strategy of robustness, future approaches will need to focus on resilience. The picture illustrates a strategy of robustness that has failed, and it doesn't just happen with levees. Think of IBM before it posted the then largest loss in American corporate history, or ENRON with Andersen still publishing case studies on its success after the collapse had started. On a wider political scale Empires tend to collapse quickly, popularity in politics is fickle and can change radically overnight.
So what is the difference between the two Rs?
- Robustness means we attempt to prevent system failure by managing the level or risk and investing on the basis of probability. Here the ubiquitous gaussian distribution comes into play, any thing six degree of separation from the mean is considered improbably. Those distributions provide a seemingly objective approach to risk management, Typical approaches here include scenario planning. A range of scenarios cover the range of the probable and we plan accordingly.
- Resilience focuses less on preventing failure than on enabling rapid recover following failure. One of the drivers for this is the realisation that in nature most events follow a pareto rather than a gaussian distribution. The import of this is that events that appear to be improbable on a a gaussian distribution turn out to be far more likely that we thought (the New Orleans levees come to mind). Given this the cost of robustness is too high, so we switch to resilience. This means that we need to manage for the possible rather than the probable.
I'm planning to write more on this tomorrow, but the basic three principles of complexity apply. We need to use more finely grained objects, distribute cognition and remove mediating layers of interpretation between what is happening on the ground and decision makers. By way of setting the scene three comments:
- An organisation that can dissemble and reassemble quickly has greater capacity to survive than one which attempts to maintain its original form, the reassembly will not be the same thing, although it will have common elements.
- Relaxation of constraint is easy during a crisis, but difficult post crisis when the need for flexility is greater. When in Australia I heard of a case where attempts by a community to recover from a bush fire were destroyed by government concern about liability insurance; thats an example of a failure of constraint relaxation
- Symbiosis generally develops from parasitism, radical change in our environment can transform the nature of relationships.
More on this tomorrow
Comments (6)
Please go on Dave - such an important topic - maybe the most important of our time
Rob
Posted by Robert Paterson | August 30, 2009 1:41 PM
Posted on August 30, 2009 13:41
What does it take for an individual to be resilient for example in overcoming or living with a chronic disease or a trauma? Antonovsky http://bit.ly/wHTdv suggests there are three elements:
Comprehensibility - Do I understand what is happening?
Manageability - What can I influence and how? and
Meaningfulness - Why does it matter, why is it worth it for me.
The latter is considered the most important and is context dependent.
How does this translate to organizations? I have found in my work that the way you engage people to move forward and overcome the "event," involves opening up meaningful dialogue using all sorts of facilitative tools to get as many people as possible to share their points of view and have themselves synthesize it in an emerging organizational point of view with emerging actions. Such large group interventions require courage from the leaders in the organizations in a time when the event is unfolding (for example, rapid deterioration of financial results with the danger to go bankrupt). Often the skills to organize large group events are absent. However, when the courage is present it is organized well, it has the potential to make the organization truly resilient in the sense Dave describes:enabling rapid recovery following failure. The challenge for leadership is to let go of contro; and to accept what is posible.
In the process of these large group events people begin to make sense, they begin to comprehend, they begin to see what they can influence and they find meaningfulness. Precisely what Antonovsky suggested.
Posted by Boudewijn Bertsch | August 30, 2009 2:18 PM
Posted on August 30, 2009 14:18
Hi Dave ... I do agree with complex theory ... but you guys make it so complicated!
eg ... you wrote ... We need to use more finely grained objects, distribute cognition and remove mediating layers of interpretation
come on ... I think ... that if you want more people to listen ... you need to find a way of dropping this.
Cheers Jon
Posted by jon thorne | August 31, 2009 10:32 PM
Posted on August 31, 2009 22:32
Your statement regarding managing for possibilities instead of probabilities really resonated with me. I am currently focusing much thought on the notion of how synthesis can effectively negate analysis, as well as possibilities of the present effectively negating probabilities of the future...
Furthering in that pattern of thought, we are constantly negotiating the risk between assumed certainties and accepted uncertainties.
Posted by Mark spivey | September 1, 2009 12:55 AM
Posted on September 1, 2009 00:55
Sorry Jon, but those three ideas (finely grained, distributed cognition, removing mediating layers) have been a theme over many courses and blogs and have been explained simply.
Posted by Dave Snowden![[TypeKey Profile Page]](http://www.cognitive-edge.com/blogs/dave/nav-commenters.gif)
|
September 1, 2009 10:39 AM
Posted on September 1, 2009 10:39
I have worked with robustness in two contexts where it was largely decoupled from probability.
In some scenario work, strategies are developed independently and then stress-tested using the scenarios. Those strategies that appear to be a good idea across a range of scenarios are considered robust - they would be good things to do under most circumstances (it reminds me a little of the concept of dominance in decision theory), and therefore have low risk, regardless of probability. I get a sense that, when a safe-fail probe shows context independence in this way, it could be an indication that it could be considered to be moved into the complicated domain ...?
The second is in assessing risk, where the seriousness of the outcome can be a more important consideration than the probability. In some risk methods a serious outcome would prompt you to develop so-called "contingency plans", i.e. plans to mitigate the effect of the event, should it occur. In maintenance strategy planning, if a failure could result in a fatality, regardless of the probability of the failure, the only option open to the planner is redesign, as such a serious outcome is to be avoided at all costs.
There is a nice range of tools to work with here - robustness (preventing failure) considering probability or severity of outcome or both, complemented by contingency planning and resilience, should failure occur.
Posted by Marietjie Vosloo | September 2, 2009 7:11 AM
Posted on September 2, 2009 07:11