Hi Dave,

Your aversion to Six Sigma is well known and I'm no great fan of it myself.

However, I was wondering how you would resolve the apparent contradiction between statements like 'increasing rather than decreasing variability is often key to success' and the well-demonstrated power of checklists to reduce variability and improve outcomes (eg see this Heroic Checklist article)?

I can't see anything on checklists via that link Stephen, it just takes me to ast Company. However I don't really see the point. A checklist simply helps me go through a set of tasks, we all need that in complex or ordered environments.

Dave,

I wonder if you have looked a John Seddon's work of Vanguard (link > http://www.systemsthinking.co.uk/home.asp ).

I can't help seeing similarities in your work, but not sure if it's wishful thinking on my part. I have a feeling however that there may be synergies to be gained from a combination/collaboration between the Cynefin Framework and the Vanguard process.

My gut feel is that the work John does would fall into the 'Complicated' domain, possibly acting as a bridge from 'Complex'.

The whole goal of process engineering is to drive out variation. Yet in complex and even complicated work activities variation is not only a given but essential for agility and responsiveness. This is a major driver for moving toward network ways of modeling work activity. The goal of a value network orientation is to create a resilient and robust complex adaptive system that can handle whatever is thrown at it - and yet provide sufficient monitoring, tracking and indicators to achieve consistent outcomes and meet conformance requirements. The role based human-centric value network analysis (VNA) meets that complexity stretch requirement better than linear approaches. Dave I know you and I both have some qualifiers and reservations about VNA and its limitations - as it is only one of several lenses and tools that are needed. Nonetheless it provides a grounded way for people to move into the kind of mindset and approaches you advocate - and is completely compatible with narrative sense making as well.

Dave,

Re: Heroic Checklists

See http://www.fastcompany.com/magazine/123/heroic-checklist.html

Interesting and perhaps not entirely incompatible with your point of view?

The background is this example:

"Intensive-care units (ICUs) often use intravenous lines to deliver medication, and these lines can become infected, causing nasty health complications. [Michigan ICU] compiled a five-step checklist [that] saved about 1,500 lives."

They argue, "Even when there is no ironclad right way, checklists can help people avoid blind spots in complex environments."

What do you think? Is it more a case that the issue with infected intravenous lines isn't really a complex one, but a known-known that people ignore while working in a complex environment?

ICUs have many highly ordered aspects for which a check list is a useful thing and no environment (even a complex one) is without ordered aspects. The point is that there is no need for Six Sigma, even for order. THere is nothing you can legitimately do with Six Sigma that you can't do with cheaper methods without the high priests and their belts.

Hi -

Shouldn't it be six stigmata (plural)? In bridge I may bid ‘three spade.’ My partners always tell me it is three spades.

Just as Six Sigma becomes six stigmata, the holy information repository invariably becomes the information suppository.

See your brain on chaos –

http://bit.ly/bAm0u2

Thanks. Good post.

-j

Good points... I believe it is easy enough for the mass of people to understand that a checklist makes sense as long as patterns and trends remain in tact.

There are two parts to the "complexity" surrounding the use of a checklist as well I believe...

On one end, the checklist has to be "understood" similarly by all relevant people in terms of syntactics, semantics, and pragmatics... or it would not be effective / efficient.

As well, it has to be relevant to the situation...

So the people involved need to ascribe meaning to the checklist (and words used, etc...) that is relevant to the situation.

But the checklist should not be reified to the point where it's existence precedes the correlationary relationship it is trying to safeguard (which is where I think Six Sigma has gone)...

So far, I do not believe any science has proven the permanence of order... which is what would be required for something to have objective and absolute meaning and value to any particular situation. (this is where I think the concept of "bounded applicability" comes in)

I believe Verna has provided a compelling observation in "the whole goal of process engineering is to drive out variation." There are other forms of engineering that have a complementary goal of "accounting for variation." I think we have two sides of the same coin, and indeed, an industrial engineer may be charged with doing both within the same Six Sigma effort. If a team, and more importantly, the company's management, is purely focused on one or the other, then they are underestimating the complexity of their work. In the Cynefin framework, their efforts might be limited to a single domain, when in reality they should have some activities or parts of the team working across two (or more) domains. Traditional methods such as red-blue teams probably have resolved this dynamic and produced resilient organizations. VNA probably does something similar.

Isn't Sixth Sigma like so many band wagon management technique fads - it starts with a single and good project (technique for reducing manufacturing defects at Motorola)then packaged to be the single (and of course extremely expensive) approach to guide all aspects of every business organisation (and then to government). It is hard to disagree with using sound statistical approaches to improve manufacturing quality (even if in your organisation you create weird and wonderful black belts, blue belts as a creative way to do it) but it seems a despicable crime to generalise that to all business situations and even worse to layer that with an esoteric priestly class system to humiliate the justifiable sceptics who all the while will never be initiated into the system and therefore ridiculed or ignored. I could think of no better approach (apart from process re engineering say) to attract unscrupulous psychopaths into your organisation.