Dave Snowden

SAFe: the infantilism of management

RSS Feed

I gave the opening keynote at the Agile conference in Brno today. A good audience in that they paid attention and were thoughtful. I tried a slightly different approach to the subject starting with some of the biology and cognitive stuff, dealing with complexity in a simplified format and majoring on SenseMaker® as a way to create continuous feedback loops between users and developers with the intent of enabling both co-evolution and exaptation. The slides and podcast are now loaded so readers can judge for themselves on the value of the approach. One interesting side effect of getting fit is I have more breath available and I hear reports that my voice has changed as a result. its not clear if it is for the better though!

Keynote over I packed for the trip to Hong Kong later in the day and then settled on attending a track session on SAFe. Now I have heard of this particular new approach, generally in derogatory terms from various people I respect in the Agile community. However I had not had a chance to hear an advocate present it before so I took it. It is only fair to see that I was appalled by the retrograde nature of the approach. To be honest the criticisms I had heard were mild compared with some of the tweets I launched during the session. One comment makes my overall point SAFe is to Agile as sick stigma was the BPR . BPR in its early days had clear utility, it drove a needed change to horizontal product and customer focus in contrast with the previous hierarchical models. But then it got applied inappropriately and finally taken to nonsensical excess with the sick stigma obsession with measurement.

Put brutally SAFe seemed to be PRINCE II camouflaged in Agile language. SCRUM as an approach was emasculated in a small box to the bottom right of a hugely overcomplicated linear model. The grandiose name of a dependency map was applied to something which is no different from a PERT chart and in general what we had is an old stale wine forced into shiny new wineskins. At the end of the session I decided to ask the speaker what explanation he had for the rejection of the technique by most major figures in Agile. His response was to argue that it worked for him and had utility. Given that was avoiding the question I asked it again in a discussion which carried on for some time after the session.

Under pressure the speaker fell back on two arguments which always irritate the hell out of me. The first was to say that he didn’t take the overall approach to seriously but used aspects when they were useful. Now that is pathetic. If he had put up SAFe and a few other approaches he had drawn on that might have been credible. But instead SAFe was front and centre, presented as the overall approach and solution. Personally I think he was simply seeking to jump on what is a successful marketing money generator, but try and retain some personal integrity in the process. That he did not achieve. The second was to argue that he was above religious splits within Agile. Under pressure got him to switch from his pejorative use of religion to an acceptance that different philosophies were at the core. My strong and increasingly passionate argument was that SAFe is not only a betrayal of the promise offered by AGILE but is a massive retrograde step giving the managerial class an excuse to avoid any significant change. OK its a obey making machine but the same applies to snake oil salesmen and the South Sea Bubble. People will get damaged by this nonsense and it needs to be hamstrung at least, garrotted at best.

Such excuses abound and allowing these false linear models to perpetuate themselves is a form of infantilism, a failure to carry through on the need for change. In particular the failure to realise that software development needs to be seen as a service and as an ecology not as a manufacturing process.

  • Pingback: buy backlinks()

  • Pingback: From Agility and Scale to Antifragility | Si Alhir (Sinan Si Alhir)()

  • Pingback: Avondale Estates GA()

  • Pingback: Fairburn Root Canals()

  • Peter Saddington

    Dave, your comments are spot on. I appreciate the candor in which you wrote this. I attended a SAFe training over a year ago with my own money so I too could speak about it intelligently… here is my review of SAFe: http://agilescout.com/scaled-agile-framework-safe-review/

    • Jack Caine

      Attendance at 1 presentation on a framework or tool, given from 1 person’s perspective, experience and knowledge, may not be an indication of the value of a tool (e.g. SAFe). Formal training might inform the person more. Actual application (use) of the tool might inform even more.

      • Dave Snowden

        I don’t need to go on NLP training to know that it is a pseudoscience, neither do `I need to waste time on SAFe training when the overriding construct is based on aggregation and constraints not experimentation. Its not AGILE ….

        • Jack Caine

          I’m curious…have you ever changed your opinion about anything? You seem awfully rigid :)

          • Dave Snowden

            Read the blog, you will find I do based on evidence. But I don’t intend to change my mind in favour of flat earth theory or SAFe

          • Jack Caine

            So you know what you know and give an opinion (not facts)…hmmm…let’s call your opinion a hypothesis instead of an opinion or fact. What empirical data would you need to see to invalidate your hypothesis? What experimentation have you done to test your hypothesis?

          • Dave Snowden

            No, somethings are known – for example that NLP is a pseudoscience. There has been enough research to establish that. Something we know about complex systems – they don’t decompose, they don’t have linear causality etc. So if a method purports to deal with complexity contradicts that knowledge it can be dismissed a priori. Which is not to say some people won’t get aspects of it to work. I’ve use astrology to get teams to talk about differences, but I’ve never asked them to buy into astrology. Otherwise you might want to read the blog post I wrote on the dangers of empirical approaches in social systems and the need for praxis.

          • Jack Caine

            Let’s say that for thousands of years that the data presented told us that the world was flat. The best minds might have said this for generations. In order to change mindset, must there not be a suspension of disbelief, some kind of curiosity towards experimentation or must we simply accept the opinion of some of the best minds that we have? I give you the example of three models: (1) Bohr model; (2) Quantum mechanics; (3) String theory. These models are all incomplete and in conflict in 1 way or another yet they came from some of our best minds. Do we speak in absolutes when they are taught in school? I think we often do to the detriment of the students…we teach the beginner the Bohr model, the college freshman Quantum mechanics and the senior college student string theory as if what was learned before was crap and what what we know today is the new absolute (which it is not).

          • Dave Snowden

            And Quantum mechanics creates an essential paradox with relativity theory. String theory is a theory. We know that they are not complete, but we also know that the idea of fire, earth, air and water is not the way to proceed. Evolutionary theory is full of unresolved issues which will get resolved (for example exaptation allows us to understand the speed arising from repurposing of traits) but creationism is incoherent therefore not worth pursuing. We can exclude some things absolutely while maintaining ambiguity about others. SAFe is incoherent to what we know about complexity, move on

          • Jack Caine

            What about “magic”? That which we don’t have the answers to yet. Should we discount it because we don’t fully understand it yet?

          • Dave Snowden

            Aspects of it yes, we can dismiss, other aspects of distributed cognition are mysterious and we should be more open. But sticking pins into dolls containing the hair of our enemy is not something we should take seriously, unless in the context of psychological belief due to lack of education (another SAFe example)

          • Jack Caine

            Hahaha…that’s funny Dave. I wonder if the witch doctor of days past who was sticking pins into dolls (an example of the latest technique) thought his technique/method/understanding to be “solid” or “mysterious” or “rigid” in belief.

          • Dave Snowden

            No idea, but I’m pretty sure Dean would have got him involved in an accreditation programme if it made him feel more comfortable

  • Pingback: The Scaled Agile Framework (SAFe) - A Review - Agile Scout()

  • blun0te

    Dave, sometimes people do have this unique gift of expressing things other people also think but cannot express. With this article you captured exactly what was born in my head throughout my recent SAFe class. Nailed it!

    • Dave Snowden

      Thanks …

  • A-R

    Any decision maker has the right to make errors of judgement, mistakes, utterly stupid decisions and so on. If SAFe is taking advantage of the above then who are we to disagree?
    In poker is a say:

    “It’s Immoral to Let a Sucker Keep His Money”.
    It should become SAFe’s motto :)

  • Spark

    The intentions on origin were right, the path taken to push it got murky and the partnerships created to cross the chasms defined its new intent. And yes now it is something that needs a plan to be contained like a epidemic.

  • Jochän Dinder

    Thanks Dave.
    That’s exactly what happens with SAFe. I had the bad luck to see it in practice. What it makes to people, organisations and agile. I’m glad, that you found the right words to express it.

  • Tor R.C. Ganslandt

    Don’t use SAFe if you don’t need to scale. But if you have a value stream where more than 50 people or so are trying to get new features to work across a few plattforms, I think it is great tool. Mind you, SAFe combines practices from LEAN and Agile and it forces management (if conveyed correctly) to rethink the command and control principles. I think you are using a lot of adjectives without substance Dave. What is it actually that you reject/object to? That the framework looks to big to comprehend? It isn’t as complicated as it looks. It just describes a few things that you probably already now if you’ve worked agile before.

    • Dave Snowden

      The point is that is that it doesn’t force management to change, it allows them to say that they have changed but stay the same. If you know anything about complex systems and scaling then one look tells you it is mistaken in conception and wrong a priori. You don’t scale a CAS by aggregation or replication. Its pretty clear what I object to and your final sentence – sorry that is almost comical but I assume sincere

      • Tor R.C. Ganslandt

        As all the change management theories state, management must lead the change and BE the change. You need to teach management to understand the principles behind SAFe before trying to implement it. I agree, otherwise it will just become another facade. But that goes with any change regardless of the magical methods or abbreviation behind it. My last sentence is sincere since SAFe might look grand and terrifying but it is quite simple and pragmatic. To me your objecting to the fact that any change is futile since we can’t change management.

        • Dave Snowden

          I suggest you read what I say. Change is possible but not with an approach like SAFe I also teach management and have two papers with the highest citation history and my comments are informed by that experience and knowledge SAFe does not look tonne either grand or terrifying. More a loosely ill thought out assemblage of anything that might be deemed fashionable

          • Tor R.C. Ganslandt

            When I 2005 was in a crowd with PM’s, PMO representatives and teams of developers, I had a hard time conveying how to change work to become effective and efficient. I tried different approaches and the more successful ones looked a lot like SAFe. I do believe that SAFe is a help to bridge some of the gaps when entering a larger enterprise of software development. To say it is harmful, well a garden tool can be deadly if you use it with ill intention. Fashionable? What is? LEAN has it origins in the 50’s. It is what you make out of it that counts. As a practitioner I object that SAFe would be a ill thought out assemblage and I object to your statement that software development can’t be seen as an industrialized process.

          • Dave Snowden

            SAFe is not a garden tool it is a collection of used agricultural tools held together with bailer twine. Feel free to object but try and argue a case and deal with the argument rather than assuming moral umbrage

          • Tor R.C. Ganslandt

            Look who’s talking. Argument. As I said there is no substance to your article, rather a series of adjectives on how bad something is or how uneasy the presenter became once you subjected him/her to the criticism that you’ve heard. Anyway, to use your own theory on the spectra simple to chaotic I would say that SAFe, correctly implemented, moves complexity to complicated, or complicated to simple, without loosing the experimental attitude and short cycles to fail fast and then improve. And, yes, used in the wrong way it is just another way to polish the turd, but then you have other problems that needs different attention.

          • Dave Snowden

            Start complicated,assumes complicated/obvious. Aggregates any method or tool which has ever had the agile level attached to it. See my comment on used farm instruments. The defence of ‘implemented properly’ and ‘do what works’ are classic snake oil tricks. As to adjectives – why exclude them?

          • Dave Snowden

            Oh and if you think software is purely an industrialised process you have my sympathy

          • Tor R.C. Ganslandt

            I said “can’t be seen as”, it does not exclude other options. There is a long way to “purely”. Also, I don’t need your sympathy. I would usually condone that kind of statement, but given your record on leadership theory and practice, I’d say it is not “deemed fashionable” to go into master suppression techniques. On the other hand, I haven’t written any papers to prove that.

          • Dave Snowden

            I was never worried by fashion.

          • Tor R.C. Ganslandt

            😉

          • Jack Caine

            My sympathy goes to all of the programmers that will be replaced by AI in the next 2 decades for basic programming activities. We went from the industrial economy to the creative economy and next is going to be the AI economy.

          • Dave Snowden

            Agree, hence the need to avoid traps like SAFe and start to tank more widely about design

          • http://ScrumTrainingSeries.com mj4scrum

            Software development can be seen as an industrialized process if you don’t mind losing your business to creative people who aren’t bound by 19th Century restrictions.

          • Jack Caine

            I agree Tor. SAFe is simply a tool…to be used for good, neglect, or evil (notice the neglect part as this is where most tools are used because “they know not what they do” and do not necessarily have the intent to do evil) :)

          • davesnowden

            Its not Scrum or SAFe and we could do with more XP. Happy to buy Jeff’s percentage. But none of that is to the point. Scrum is a technique which carries out a specific task. It is not an over arching framework

          • Jack Caine

            Does LeSS live up to your expectations of a properly designed scaling framework? (or de-scaling framework)

          • davesnowden

            It’s makes less extravagant claims that SAFe and didn’t (as far as I know) adopt the pyramid selling scheme. However I think ALL the frameworks assume an engineering approach rather than the looking at creating an ecosystem

      • Jack Caine

        You cannot “force” someone to change. Real change, just like real quality, is what happens when no one is watching. You can; however, put structure in place for behavior (and perhaps culture) to follow in a certain direction but there is no guarantee in this either. I like Craig Larman’s “informed consent” model; however, most people won’t make the plunge into transformational change with both feet if they understand upfront what they really have to do (too painful for most people due to fear). Just as you cannot force someone to change, you cannot mandate bravery. Many a transformation have failed by trying to “force” something.

        • Dave Snowden

          You can put people in a position where the old ways of working cannot be sustained, then they change. I don’t shy away from terms like ‘force’ as I think we need to be more honest. Ideally the nature of the change should be a response to the change in constraints, but sometimes ….

          The point is that SAFe allows people to carry on as before but not change

          • Jack Caine

            Are you saying that not a single person’s mindset has changed for the better by being introduced to SAFe?

          • Dave Snowden

            No, but it won’t scale

    • EmigratedUK

      precisely

    • Moustachiarty

      Here’s a simple way to object to SAFe. Compare it with the Agile Manifesto.

      https://www.linkedin.com/pulse/does-safe-comply-agile-manifesto-peter-merel

      As Merel observes, it doesn’t comply. End of story, pick a better method – there are a dozen of them in the last Gartner survey.

      • Frank Bauer

        Which ones though?

        I personally hate SAFe for all the reasons why other Agile evangelists hate SAFe. Being against SAFe is a no-brainer. But I am a big fan of constructive criticism. The key principle of constructive criticism is that people end their criticism with a “And here is how to do it better” section. If I don’t find that section, then I consider the whole article to be worthless. Yea, I am kinda strict and dogmatic about that, because telling what is wrong is something my four-year-old can do. Giving solutions is the part I expect from intelligent people.

        And no, I don’t consider “Check out some survey” or “Google it” or “Well, don’t you know already?” to be valid answers.

        That is the funny thing about this whole SAFe discussion. I see a lot of people going on and on about how people are stupid to try and implement SAFe. Fine. But when asked for a non-stupid alternative, the answers become very very very (very!) fuzzy. At least you tried. Mr. Snowden’s article contributed exactly zero in terms of solutions.

        I used to work for a company where 1000 people worked on projects that culminated into an annually released product. Teams A and B waited for Teams C, D and E to finish stuff that Teams F, G, H, I and J were contributing to. And that was only 2% of the whole value chain. How do you transform that whole organization into something self-organizing that satisfies the Agile Manifesto? I have something in common with Mr. Snowden, all (highly experienced) agile coaches that came and went at that company, and as far as I can tell everyone else in the Agile community: I had no solution. My solution in the end was: I left that company (not kidding here). Yea, I ran away from a problem that nobody had a solution to. Shame on me.

        Where is the concept that gives us a full-fledged solution to having 100+ teams cooperating with each other in a self-organized manner? Where is the book that describes how that works? Simple: It doesn’t exist. No, really: It does not exist.

        I am not saying there can’t be a fully agile framework to deal with that problem. I am just saying that it is not there yet, and “Try a few elements of this” and “There are some conceptual ideas here” and “SAFe doesn’t work” and “Well, you just gotta be agile” is not the right solution. But that’s what you get whenever someone tries to present a solution: Very very very (very!) fuzzy BS that is ultra-heavy on jargon and very light on content. The kind of BS that gives Agile a bad name.

        Scrum has rules that help a small complex system to self-organize to solve a small complex problem. There is simply no equivalent yet to do the same for huge complex systems with huge complex problems.

        Now I am just waiting for the next cop out: “It’s a highly complex problem, and there is no one-size-fits-all solution to that.” I have heard that many times as well. Any argument seems valid, as long as people can avoid being constructive in their criticism. But I have yet to see someone say “Ok, I will sit down either alone or with other smart people, and I will come up with a framework that we can apply to solve this.” As long as nobody has that framework available, articles like this one along with this whole discussion thread are pointless.

        • davesnowden

          I’ll reply at length later Frank but this wa not an article it is a blog post reflecting on a presentation. The fact that three years later it is still referenced is significant. That side you seem to suggest that the only alternative to a massive overarching framework is a direct alternative. It isn’t and there are other options people can take now and their other collaborative non commercial initiatives in various stages of development that will provide different approaches.

        • Dave Snowden

          OK Frank, I originally thought you were responding to the Linked in Discussion, but now having a little more time and not working off the iPhone I discover you were actually commenting on the blog post. That means it is disingenuous of you to talk about my post as an ‘article’. If you want to read other posts here, or real articles or turn up at a lecture you will find that I, like many other people in complexity theory have many ideas, practices and experiences in alternatives to these highly structured approaches.

          But no I am not offering an approach that looks like SAFe because such approaches are plain wrong and it is more than legitimate to do so. There is a very simple alternative that existed before SAFe which is to pick up SCRUM, KANBAN some XP techniques and good old project management and make it work. Intelligent (to use your word) people are more than aware of that.

          As to people who play the BS game, my general experience is that they are covering up the fact that they don’t understand the language or the concepts that are in use. I hope that in those circumstances you would do what any intelligent four year old would do, which is to ask for elaboration rather than tantrum. Sorry to be direct but I repeat this is a post on a blog with near daily entries. There is a lot of material there if you bother to look. It is of course always possible that blogging is something with which you lack familiarity of experience in which case you have some excuse.

  • David Wells

    I respectfully disagree, and submit to you that proselytizing “pure” Agile to management teams responsible for multi-million dollar global strategic initiatives accomplishes very little. My experience as consultant, coach and trainer has taught me that pragmatism always trumps dogma, and you have to meet the client where they are if you hope to gain any traction with them in changing the way they operate. SAFe provides a great set of proven tools, and simply ranting that it is evil because it doesn’t conform to the orthodoxy of the cult leaders strikes me as beneath a man of your obvious intellect…

    • davesnowden

      Ah the Dean propaganda machine is in operation again with the standard phrases being trotted out. I’ll reply later and please respect the English language, you mean something other than ‘respect’ in this post

    • Dave Snowden

      Ok lets get to the meat of this …
      – firstly no one is proselytiszing “pure’ Agile and even if they were the alternative is not SAFe. Its a classic sales trick to create a crude dichotomy to make yours the only alternative. Agile needs to scale and needs to increase its relevance. That fact does not justify SAFe
      – but it is a classic line from the Dean propaganda line, the other common one (that you evidence) is to attempt to take the position that SAFe is pragmatic and any opposition is some form of dogma. Complete hokum, Pragmatism is about dealing with where the client is and moving them forwards, not allowing them to slip back into the comfort of something that claims to be Agile but lacks the flexibility of Prince II in its approach :-)
      – SAFe clearly is not proven, but it seeks to combine multiple proven approaches, as such it uses other’s success to validate its construct – very much in common with the pseudo-science NLP which takes some proven techniques but then weaves them into a mystical construct that claims efficacy
      – I’m happy if you want to say that snake oil selling is evil, but I don’t need to rant about it. Neither do I seek to confirm to the orthodoxy of cult leaders (that one was pathetic). I do seek to respect the contribution of the Agile manifesto and not see its key messages corrupted by mercenaries
      – The essence of Agile is small incremental steps, the essence of SAFe is one massive commitment up front to a set of dubious claims and borrowed practices. It is not Agile ….
      – SAFe is to Agile as Six Sigma is to Innovation and Sharepoint is to Knowledge Management A refusal to realise that the world is complex and an attempt to impose inauthentic order to development. It is in the interests of consultants to sell all three, none are in the real interests of the users

      • Jack Caine

        I’ve heard Jeff Sutherland say several times, and also talked to him personally about this, that less than 10% of all Scrum team are “high performing” (whatever that means).

        Is Scrum to be considered “proven” when less than 10% of the Scrum teams do not meet the level of “high performing”?

        A high percentage of projects are now using agile in one form or another. How does one designate “proven” or “good” or “better”?

        Do we do it based on market penetration? Scrum, for sure, is the most widely used agile framework; however, and I quote Jeff Sutherland again, that he’s never seen a high performing Scrum team that wasn’t also using XP engineering practices. Does this mean that if you aren’t using XP that you are wasting your time? (my mentor Ron Jeffries would agree with this comment) :)

        And VersionOne’s last report shows that SAFe is the most widely used of all the scaling frameworks. Are the people using SAFe just lemmings following SAI’s marketing plan? Hmmm…SAFe requires ScrumXP, not just Scrum. Is there any good to come from doing this?

        Must we have a revolution at all levels of an organization in order to see “good” or “better”? Or can we start in the middle (program level) and get teams to work together better (using Scrum and XP) by talking to each other in a structured way if they haven’t been able to do so in the past?

    • Moustachiarty

      Where are the numbers, David Wells? You assert that SAFe provides “a great set of proven tools”. Okay, let’s have that proof then.

      Is SAFe less disruptive than other enterprise agile frameworks? Does it pay down technical debt faster? Does it raise business throughput quicker? Does it reduce risk or seize opportunity more effectively? If it’s effective at something, what is that thing?

  • Pingback: SAFe: the infantilism of management – Cognitive Edge | SAFE()

  • marcdanziger

    This may be the pluperfect Agile piece on LinkedIn; it encapsulates much that’s wrong with the discipline, and as such ought to be addressed (skip to the end for that commentary).

    I’ll start with the question of SAFe; in reality it has become the ‘elephant in the room’ as the executive cadre begins to see ‘Enterprise Agility’ as the equivalent of SAFe. I’ve done SAFe (before I’d studied it) and studied it and see it as fraught with problems if it is viewed as a ‘solution’ – which many people do.

    As such, it’s an antipattern for dealing with complex adaptive systems and the wicked problems they present.

    Is it useless, or even toxic, as the piece suggests? As a framework/toolbox, not necessarily.

    I first brought it to a client that had problems driven by culture and by a leadership that was constantly and arbitrarily replanning and changing direction and couldn’t understand why the tech team had trouble getting things done.

    My belief was that simply establishing a strategic portfolio – and watching the churn in that portfolio – would temper their behavior, and that establish a coarse-grained planning cadence would keep their changes and the ability of the teams to deliver in alignment.

    This was based on my reading of Dean’s book, not any specific training. And it worked rather dramatically.

    This was the opposite of being a “SAFe practitioner” and implementing SAFe as an industrial template; and in this context, tools derived from SAFe were quite useful.

    SAFe clearly is sold in a different way – but it’s interesting to note that when directly asked the question, Dean is clear that it’s a framework/toolbox, not a solution.

    So by all means, let’s have a discussion of SAFe, and let’s deconstruct what, if anything, is good/useful in it and what, if anything, is dangerous/stupid.

    I’d be thrilled to sit in on a serious conversation about that question which wasn’t “we rule, you drool” which is what these conversations tend to devolve too far too quickly – and I’ll bet that dozens, if not hundreds of serious practitioners would be interested and willing to participate as well.

    This piece, OTOH, falls neatly into that category. The author seems to have missed the room set aside for argument – I don’t see a lot of them – and moved directly to the rooms for disagreement and abuse.

    When challenged, he steps back to argument from authority (…the weakest form of argument on the basis of the authority of Boethius – an old Aquinas joke).

    I’m genuinely interested in what we ought to do about driving more agile behavior into large organizations. I’m even genuinely interested in what the author has to say about it.

    This piece is unfortunately mostly interesting as an example of how not to discuss these things. I’d advocate a healthy dose of humility for everyone involved and a reboot.

    • Dave Snowden

      OK so you read a book by Dean and found some stuff you could use. Great, buy that. Given that much of SAFe is a loose aggregation of other methods and tools once would expect that. The issue with SAFe is its attempt to treat a complex system in an ordered way, to avoid the very basics of agility and complexity – namely small safe to fail interventions. SAFe for me fails a priori because of that construct. I also think Dean is cynically saying “use what works’ so that the overall framework is less likely to have its flaws exposed. Otherwise I am happy to argue from authority in respect of human complex systems and if you have a problem with that fine. Go study the subject and come back with a knowledgeable argument and I’ll listen. Otherwise this is a blog post reacting to a specific event not a considered academic argument (although that would come to a similar conclusion as I have studied it a lot since I wrote this). The fact it constantly gets referenced and liked three years later is interesting, and the fact that SAFe apologists regularly attack it with almost identical arguments is a tribute to the being ‘on message’ in the marketing. If you want a wider discussion of complexity in Agile then take a look at any number of recordings I have made at conferences.

      • marcdanziger

        “Go study the subject and come back with a knowledgeable argument” – thanks so much for peacocking and making my case. I’ve been studying bounded rationality and complex systems since the late 70’s with people like Horst Rittel. I’m pretty comfy discussing it with an Agile coach. Want to try again to make a serious case about complex systems? Because I don;t see one in your arguments.

        • Dave Snowden

          Then answer the points from the perspective of Complexity science not your personal experience. Then I’ll take you seriously

          • Jack Caine

            Are you saying that it is like a perspective of resolution? I mean, like fractals, as you get closer or move farther away, the resolution changes and the things become revealed or hidden but the entire system cannot be fully appreciated. Is that how complex systems work?

          • Dave Snowden

            fractals are a part of complexity but yes I think I agree with that. Also you only understand the system when you interact with it, hence the need for experimentation before massive structure. Complexity gives us a natural science based understanding of systems which means we can move forward under conditions of uncertainty. We also avoid the trap of allowing a few cases to determine method (which is what happened in SAFe)

          • Jack Caine

            It seems that all of the scaling frameworks (SAFe, LeSS, etc.) incorporate hierarchy (differently in each, but there nevertheless). Does the use of hierarchy prevent the scaling of systems? Must we solely scale via networking to allow the benefits of complexity (e.g. innovation, creativity, etc.)?

          • Dave Snowden

            Not per se, its is more about the nature of constraints on interaction. Hierarchy can be an enabling constraint so it can allow complexity to happen.

          • Jack Caine

            Thanks for that. I have another question specific to something that you wrote above….”The issue with SAFe is its attempt to treat a complex system in an ordered way, to avoid the very basics of agility and complexity – namely small safe to fail interventions.” I’m curious what you consider “small” and what you consider “safe” with regard to experimentation? How might we experiment on a complex system without some constraints (order) that allow for that experimentation? What parts of the SAFe flow do you consider to prevent the experimentation?

          • Dave Snowden

            The issue is not parts of SAFe but the overall construct. SCRUM for example is a complex to complicated conversion technique. Nowt wrong with waterfall if high constraints. Timeboxes good in complicated. True complex requires experiments that take hours or days and run in parallel. Its an eco system not a system diagram

          • Jack Caine

            Would you say that SAFe attempts to convert complex to simple and this is main thing that you dislike? Is SAFe trying to dumb down reality?

          • Dave Snowden

            It reduces complexity rather than embraces it so its not that I dislike it per se its just wrong :-) Dumbing down reality is OK will go along with that, but ignoring reality better. Its creating a construct that is far too comfortable for people who want to appear to change but actually stay the same

          • marcdanziger

            Oh, piffle. You’re not doing complexity science; where are the masses of data you’d have to analyze about an organization? Where’s the computation to define edges and hubs and to weight relationships? _You’re using complexity science as a metaphor_. And in doing so, you’re – in essence – burying a set of assumptions and belief in jargon (which is why you’re so defensive when challenged).

            More dangerously, you’re skirting the issues of agency and responsibility within organizations that must be dealt with in order to trigger real change.

            Thanks for your time, sorry I intruded.

          • Dave Snowden

            If there is onev thing any modern organisation is not lacking then its data. Also that is a requirement for modelling which yes is being done. In terms of understanding how a system operates and making judgements about a design or process we don’t need to restrict ourselves to metaphor. If explaining some basic science to you and asking you to respond (given your claim to knowledge) is being defensive then it’s your use of language which is metaphorical at best. If you don’t want to engage with the argument then yes, you are wasting your time

          • Dave Snowden

            Oh and what issues of agency and responsibility are being avoided? A problem with SAFe is that agency becomes bureaucracy and control, it’s not agile

        • Dave Snowden

          Oh by the way Rittel is one of the pioneers of Design Thinking (and a good one) and techniques like Issue-based mapping are very useful. But like a lot of techniques from that period which fall within Systems Thinking they tend in various ways to assume that there are drivers and linear material causality, but we may not know it. Complexity, and in particular anthro-complexity or the study of complexity in human systems deals with systems that are modulated and which have dispositionality not causality. I think when Rittel used ‘complex’ in the context of the definition of ‘Wiked Problems’ the use was more in the moral English language sense off the word. From that perspective you might see the argument differently.

  • Andy Wootton

    A PERT chart was always a dependency map. They were largely discredited in the 1980s in the ‘Software Engineering Journal’ for use with software projects, due to concerns about use in complex problem domains :-) though a couple of days ago, I discovered that the idea of ‘Software Engineering’ was already being challenged by Edsger Dijkstra in the late 1970s. He proposed replacing it with the idea of ‘scientific design’. Sounds like real agility to me. We should have pivoted earlier.

Top