Dave Snowden

Fidelity in complex systems consultancy

RSS Feed

It been a week since my tirade against the abrogation of leadership by falling back to the crutch of a mega-consultancy report. Since then a mixture of bad flu, trans-pacific, trans-americas & transatlantic flights have all conspired against my blogging. Returning yesterday from my first TedX in Cyprus I was chatting with Sonja Blignaut about some upcoming activities and differences between approaches to sense-making and complexity. In the course of that I realised that small consultancy groups and individual consultants are not entirely innocent of the same issue, frequently focused on retention of utilisation rather than risking any ill feeling by saying or enabling something which disturbs their client. It’s a wider issue, if we look at policy making in Government there is a general tendency for policy advisors to focus on being invited back to the table rather than to tell a hard truth, or provide all the facts. Now it is an understandable issue; if you work for yourself then securing revenue is as vital and shooting, or failing to rehire the messenger is all to frequent. So its a hard problem, if not the Hard Problem, and it is especially so when you are working in a paradigm shifting area such as complexity theory. Changing paradigms can be deeply unsettling and, like all pioneers, if you are doing your job well then the chance of arrows in the back is high. You won’t have a problem at the time, but ….

So lets start this with a quote from Cicero, of the ancients one of the most relevant political philosophers to our current condition not least for his opposition to use of populism to impose despotic rule.

Nothing is more noble, nothing more venerable than fidelity. Faithfulness and truth are the most sacred excellences and endowments of the human mind.
Marcus Tullius Cicero

Now complexity is uncomfortable to any organisation (which in fact means more or less all organisations) as it challenges several key assumptions. In particular the assumption that you get set clear objectives for a desired future state, that things that work once will repeat as is, that cases from the past provide recipes for future action and so on. In a complex system we:

  1. start journeys with a general sense of direction but without precise targets.
  2. have to be comfortable with what is inherent and unavoidable uncertainty
  3. need real time feedback, with the flexibility to respond and change quickly
  4. can’t avoid some understanding of theory, which has to act as a enabling constraint on action
  5. are ethically responsible for the unintended consequence of any intervention

None of that makes life easy for event based consultancy which tends to assess its success based on evaluation at the end of a workshop. When I was looking into backcasting cases, most of the success was how people assessed what they would do, or who they felt at the end of the workshop. As I have pointed out in a previous post that is a very poor measure of change; what matters is what happens as a result, actions not intent. A related issue is that any workshop using complexity approaches is going to leave people with a heightened state of uncertainty and their brains may hurt. Thinking anew is hard enough, acting anew is even harder (Lincoln reference there).

Now we can mitigate the problem. Our approach using structured forms and portfolio management for safe-to-fail experiments shows fidelity to complexity principles, but uses a familiar form. If we use MassSense to present (and represent) material and ideas to desperate networks we can create an evidence based approach to understanding the patterns of uncertainty; we can measure coherence and indicate significant outliers. But we can’t hide the fact that this is different from what people are used to. I remember back in IBM days running a two day session for a company in a large group context. I was working with an IBM partner level consultant who panicked at the end of the day one that we had nothing resolved. I had explained multiple times that in complexity work it all comes together at the end of the session, but we are avoiding premature convergence. That means we are holding as many options open as possible. He intervened, it went wrong, I recovered it and, when he saw over 600 action forms at the end of day two he apologised and we ended up writing some articles together. Others were less able to learn, for them the final result should be starting to become clear early on or there was a risk that it might never emerge. I once said that the whole idea was the prevent that sort of solution from ever emerging but that didn’t go down well.

So coming back to my conversation with Sonja (and she is only responsible for simulating this not necessarily for my conclusions) there are some clear indicators that a consultant using complexity lacks fidelity:

  1. Their methods and tools have been rebadged with complexity language, but are not really different. They may just use the language to have all too familiar conversations focusing on consensus and clear agreements.
  2. The output is comfortable, couched in language and form that it is difficult to disagree with. They may not be out and out platitudes but you get the point
  3. The goals are idealistic, not progressive from where we are, but aspiration as to where we should be (I need to elaborate this more and the distinction between goals and directions will be the subject of future posts)
  4. Any actions arising lack contradiction, complexity work is about increasing heterogeneity not achieving homogeneity 
  5. They advocate happiness, which is not to say that I want people to be happy, but that is a consequence of downstream results not an objective of near term intervention.
  6. Interventions focus on individual behaviour change or goals rather than changing interactions.
  7. The consultant has been engaged in the content, rather than managing the process 

Now I need to expand on those, introduce some qualifications and so on but its a good starting point.

  • Piotr Maksimczyk

    Thank you for this piece of writing.

    Although I sometimes do not understand some of the concepts and wording you use, as I do not have any formal education in the complexity theories area, each time you are in Poland I passionately attend your lectures. Last discussion panel during Agile By Example was the most memorable of all. I also willingly read this blog.

    What you write about consultancy, telling the truth, gathering real-time feedback, complexity, resonates with my thoughts I have after 20 years of software development. I see not only that the iteration lengths of projects/products are becoming shorter and shorter, but also this process becomes faster. After all those years I have intuitions (not formal knowledge) that product delivery is some kind, as I name it, “quantum delivery”. By saying this I mean we deliver the smallest possible quantum pieces of bits, software, information, human interactions, etc. Then we try to gather feedback in real time and respond. In that context any planning in advance seems to be waste of time (e.g. scrum kind of planning), any target-driven or goal-oriented software delivery becomes pointless. Assuming we are in the complex domain, we are completely unable to predict what will be the results of our doing including financial safety. This may be very uncomfortable for many us, as we are taught in school, then in companies and corporations to constantly try to meet the plans and schedules so we can pretend we are more safe.

    • Dave Snowden

      I think you have the heart of it there

  • Marco Valente

    Greetings Dave,
    I have been reading your blog for long enough to understand the caveats you use. On connecting to the client’s expectations and the usual want for certain outcomes, I found in my community of facilitators that managing expectations is crucial, as well as radical openness. It is honest to probe them in conversations about how comfortable they are / would be with unexpected outcomes, with uncertainty, with not knowing what the answers will be beforehand. There is I believe a “coaching” role in dealing what often are personal anxieties that can be associate with a client’s pressures: often we are not talking about clients ‘obsessed with control’ but rather about a whole system of incentives that privileges having the right answers vs asking the crucial questions, and there is a lot we can do in easing those personal concerns as consultants.

    • Dave Snowden

      I agree on the wider context of incentives etc. being a major part of the issue. My concern in this post was more to identify the dangers of not shoring radical openness and also to create some markers for to identify people using the language of complexity but not changing their approach

  • Hanno Jarvet

    Dave, why does one’s brain hurt after dealing with complexity issues?

    • Dave Snowden

      Don’t ask me, I see it as the science of common sense and since I engaged fully I’ve had fewer headaches. However the issue is frequently reported …

    • Michael Hill

      It’s like getting off the couch to exercise. It’s hard the first time after a long time on the couch. It’s easy if you’ve always been doing it (using your brain to think rather than follow).

  • Robert Falkowitz

    So much for the midwife theory of consultancy.

  • David Noll

    In many organizations the need to know all the answers very quickly means the ambiguity of the complex approach is extremely stressful for them, which is simultaneously difficult to work through and very entertaining to watch. The tension that we all have as consultants between telling the truth and making sure we have food on the table is tough. Often we can create environments where the client discovers the truth on their own which is the best approach if you can pull it off. That approach involves creating experiments which sometimes work sometimes don’t but you can learn a lot about the system based on how it responds. And lastly, Cicero had his head cut off which I suppose illustrates your point regarding the tension as well!

    • Dave Snowden

      I chose the example of Cicero with care – his execution was a bargain between the three tribunes, each had to sacrifice one favourite .

  • Asher Sterkin

    Great post, Dave, as always!
    The last “uncomfortable truth” about consultancy business: “The consultant has been engaged in the content, rather than managing the process” is the most challenging for me. While I understand the principle, and I do know it will backfire if violated, I’m struggling with implementing it in practice. Unless one has a very long record of success and large (normally by word of mouth) customer base consultants somehow need to establish some credibility before customers start taking them seriously. If I’m coming from the software business my first cohort of customers would be software companies and then naturally they look at my professional background and temptation, as well as at times tactical needs, to demonstrate that I do know what I’m talking about is too strong to resist. It’s not only about software technicalities but the subject matter itself. One practical trick, which I plead guilty to commit to, would be to keep this constraint flexible and context dependent: to demonstrate some knowledge to gain initial credibility and then to step back as quickly as possible once it did the job. Still, any advice about #7 above would be highly appreciated.

    • Dave Snowden

      If its advise they are asking you for, interview and write it up! Its the workshop dominance I find an issue as it may not tend to sustainable solutions

      • Asher Sterkin

        I’m trying to avoid workshop dominance as much as my personality would allow:-). I feel I need to establish some credibility at the first part of it in order to demonstrate that I do understand the company business and technical challenges and have some rough idea about what kind of solution it might be. Recently it went this way. I asked somebody from the audience to state the problem, then provided a recap in my own words asking to confirm that it is correct, asked some more clarifying questions and then proposed a solution direction. The problem was lack of common language between operations and developers and my proposal was to adopt a very lightweight variant of Event Storming. It went well, I hope, but it was possible only because I spent substantial time on learning about their domain and technology and what state-of-the-art solutions could offer. I was deeply involved in the content and it was hard to imaging how could I limit it to managing process only. So my question remains: was it completely wrong or there is some zone of maneuvering. The concern is, off-cause, that by getting into content too deeply I reduce chances of a serious followup.

        • Dave Snowden

          It really depends on context, establishing credibility then facilitating a process are separate. You might be an expert participant …