‘Snowden – Seddon’ Venn diagram

Back on St David’s day at the start of this month, Ben Taylor promoted a link to a post from Steve Schefer comparing my work with that of John Seddon.  It was interesting and given some current intersections with Seddon’s Vanguard Method, I thought I would write a response.  It’s also worth reading the exchange that followed Ben’s reposting and some of the links he provided.  I’ve always been puzzled by just what Seddon’s method is and why he labels it systems thinking when from all I have seen it’s basically a variant of Total Quality Management.  As Ben says: “It’s ‘normative learning’ (taking managers back to the floor to see for themselves), understanding demand and how the demand message gets screwed up, a form of ‘action learning’ (loosely) around following that through to see why, a slightly naff beliefs-behaviours model a la NLP, and a slightly more useful purpose-measures-method model, and lots of sensible stuff about how the view from management world is distorted, particularly around and because of targets.”  My own view over the years has been that Seddon is a pretty passionate consultant but that a lot of the claims made don’t stand up to scrutiny.   Ben more or less confirms that view and links to some articles that expand on it and there are frequent comparisons between Seddon and Old Testament prophets, hence by the inclusion of Michael Angelo’s portrayal of Jeremiah as the opening picture.

By way of a compliment to Ben, and I’ve been critical so this will balance that, he collects a lot of things in what he calls his tool shed and is an avid linker to interesting material on social media.  My own view is that his tool shed is a little to far on the messy side of messy coherence, (see the banner picture) but there is still a lot of useful stuff there.

This post is probably best read with Steve’s post open in another window – that is certainly how I am writing my response.  I’m also going to follow his headings.   I quite like the opening as the Venn diagram rightly shows intersections between my work, systems thinking and Seddon’s work, but it doesn’t attempt to subsume the three into one – something that too many people in this space do.  Differences are important.  I’m reasonably happy to be known for my precision in the use of language, it indicates that those years getting a degree in Philosophy were not wasted.  I’m less happy with the Newton/Leibniz comparison as I see only practice little theory in Seddon’s work as described.  But OK let’s live with that for the moment. Just for the record I’m on the side of Leibniz in that dispute, I was never a fan of Newton and the use of the Royal Society as a propaganda weapon.

So there are nine sections in Steve’s post and let’s go through them one by one:

On making sense of the problem space

Steve’s summary of Cynefin is not wrong, but neither is it right.  He makes the distinction between order and complex and then tries to match that with Seddon’s distinction between manufacture and service, with the latter being people-centred and corresponding with complex in Cynefin.  There are a few problems with this:

1. In general, defining the problem space is more about Estuarine Mapping and SenseMaker® than Cynefin.  The function of Cynefin is to indicate the types of decisions that can be made based on the nature of the system; not on the discovery of the system type.  This is a pretty common mistake so I’m grateful for the opportunity to point it out.
2. There are clear distinctions between manufacturing and service but both have complex and ordered aspects.  Its not a binary and all have people-centricity to varying degrees
3. The liminal aspects and dynamics of Cynefin are more important that its domains.

This first one really is a stretch to compare the two, but I’m happy to agree that we have in common an opposition to treating people like widgets in a manufacturing process

On working with complexity

Again it’s not wrong, but importantly in Cynefin parallel safe-to-fail interventions are important.  We’d also now emphasise micro-interventions at scale to change the energy gradient or dispositional state.   The description of helping people, building trusted relationships over time, and providing autonomy are all things that I would agree with and the suggestion that absorbing variety is a common focus also.   That said my view is that such behaviours are themselves an emergent property of changing the way that people interact.  And the various methods outlined in the Field Guide are all designed to work at scale with minimal consultancy intervention – something that may be a difference here.  And Snowden isn’t privileging a salt marsh over a sea wall, he is saying that both have validity,

On the importance of failure

From what I can see Seddon is trying to use failure as a means to get management’s attention and engage management in making change.  That is a little different from sharing failure as a learning mechanism.  We also have a host of formal methods – archetypes, future backwards etc. etc. designed to allow such sharing without creating undue stress in the system.

The failure to see a gorilla or inattention blindness is part of what we are, so we have to define systems that work with attentional blindness.   I am dubious that an explicit focus on enabling managers to see such things would really work.  Again at scale, we need to find ways for weak signal detection and distributed decision-making that work with the flow of how humans make decisions.

So again – yes there are similarities but the difference is our creation of methods and tools to do this at scale without dependency on consultants.

On facilitating change within the complex domain

Steve correctly identifies that we focus on the discovery of localised insights and the creation of micro-change projects which includes experiential learning.  But we do a lot more as well, including focusing on decisions about creating constraints and processes that allow such behaviour to emerge within an auditable setting.   I also think we depend less on the ‘interventionist’ than the Vanguard people but I could be wrong there,

On scaling change within the complex domain

Scaling by decomposition and recombination has been a theme of a lot of my writing here and also speaking.   The wiki probably needs some work to make sure it’s not a stub.  There is also an indirect reference to vector theory of change so I’m OK with the description of our approach, in part.   Rolling in rather than rolling out would fit within that approach.  But I think you have to do more – the work we are doing on distributed decision-making is to make this more than an admonition, but something that can be audited with appropriate transparency and which sustains itself beyond the consultancy intervention.  So yes broad agreement but very different approaches to achieving change at scale.

On measurement

Another right but also wrong comment here, most aspects of a complex system support vector rather than outcome/output measurements but it’s not universal.  You might set outcome targets around a constraint shit for example.  The purpose question is more complex – I’d agree with Steve’s analysis is correct in its summary of Seddon, but from my point of view there are theoretical and practical issues with defining emergent properties in advance of their emerging.  Until stress resulted in flight as an exaptive property of feathers we couldn’t really have set flight as a goal.

Combining quantitative with qualitative

I really disagree here – yes cases can be complimentary but they need to be authentic to micro-stories and not written post hoc.  Part of the criticism of Seddon in the linked articles is over claiming it is always a temptation.  Stories can’t really predict either – what they can do is help people predict and take short-term actions – the adjacent possible.   So here I think there are some real and substantial differences.

On the use of tools

I’m OK with this summary so nothing much to add

On ‘Systems Theory

I don’t argue against hard systems theory, but I do think its application is limited.  Seddon I think is all practice and I’m dubious as to what appears to be his appropriation of the term systems thinking, and I know I’m not the only one.  I don’t see any real theory in his work, although I do see passion, concern for the humanisation of the workplace etc.    So I think Steve is wrong here – systems theory is not an intersect

In conclusion

I have a little niggle here.  We publish everything, we put methods in open source, and we do a huge amount of training and lots of indirect help.  I don’t see the equivalent in Vanguard which does seem dependent on Seddon as a “guru’.  But I would fully agree that both of us are happy to call out nonsense even if people would prefer us not to.  I like to think I’m less Manichæian but …

But overall I’d like to thank Steve for this.    My own view is that Seddon and I probably agree on the main values, but we have different ways to make those real.   I’d suggest that our differentiator is that our methods can scale and our emphasis is not on a consultant lead intervention.  But if someone likes what they have done with Vanguard, we’re happy to pick that up and use methods and tools to apply that experience at scale.

The picture of the tool shed is cropped from an original by s w sourced from Unsplash and that of Jeremiah is from Wikipedia