A very interesting post from Nick Malik over at Vanguard EA recently. Go read it first and then come back, I’ll wait.
I was touched by his (almost) cri de cœur at the end…
Today… I do not know
The post has drawn me back here to write something for the first time in ages because a) the subject resonated very strongly with me and b) I wanted to cheer him up a little!
All models are wrong…
…but some are useful. This quotation from statistician George Box is a touchstone for my more formal architecture work. The sentiment being that architecture models inevitably are a simplification of the real world in order to try and generate insight. The balance that always need to be struck is whether the gain in insight is offset by the loss of fidelity compared to the real world.
A variation of this looks to the fact that the business is constantly evolving underneath you and sometimes you don’t even know all of the change that is happening. I have heard architects bemoaning the fact that businesses don’t come with configuration management and version controls – OK, I own up, I’ve also done this myself! But unless the change is “architecturally significant” does that matter?
So, in the previous two paragraphs, I’ve talked about two things that are not susceptible to formal, rigorous architectural modelling: balances needing to be struck; and, defining something as architecturally significant. This, hopefully, sets the theme for the rest of the post and starts to speak to the title as well.
In his post, Nick makes the following observation…
…it should be a clarion bell to any architect who drinks their own lemonaid and believes that an enterprise is a construct. It is not. An enterprise is a community of organic elements, each of which grow and change according to organic principles.
I’d like to buy that man a drink for saying this. Although presumably not lemonaid?
No business change or architectural activity happens in a vacuum; it is always happening in the enterprise’s current “context”. I’m using context here to refer to the (often) subjective, people related, not necessarily rational, sometimes temporary, emotive and/or political aspects of the organisation. Where is the people and leadership headspace at right now?
Sometimes, elements of this context need to be challenged because they will block the enterprise from realising the strategy; other times the game isn’t worth the candle and we’re better off just accommodating it. What is common to all of them, though, is that I don’t think they are easily susceptible to formal modeling or rigorous, rational analysis.
As a side note, it might be an interesting exercise to see if some of these aspects could be modeled using something like the OMG Business Motivation Model. However, I suspect that semantically these aspects are probably internal Influencers which, in my mind, come across as second class citizens, compared to the first class artefacts of Means or Ends.
In my mind, it’s the mark of a really good architect that they can successfully navigate these waters in the absence of the models that explain them. And, no, I don’t necessarily count myself in that group as, while I can recognise them, I’m not always successful in handling them well!
Is it the enterprise or is it the people?
Nick’s post talks about the self organising enterprise but I think it might be worth thinking about whether the enterprise is organising itself or whether the people in the enterprise are organising themselves via behaviours that are manifesting as enterprise changes.
In either case we, as architects, would do well to acquaint ourselves with seminal works such as Mullins’ Management and Organisational Behaviour and Handy’s Understanding Organisations. Neither of these offer formal modeling approaches and I’m not aware of any works out there that do model enterprise level behaviour. If there are any I would be very interested if somebody could point me at them.
On the other hand there is wealth of material out there modeling people behaviour. Here’s just one example which, while dealing with the domain of healthcare, outlines at least 3 behaviour modeling approaches. Two key challenges will exist in adapting any of these:
- Is it possible to adapt these models to extrapolate the impact on enterprise behaviour? and
- What would be needed to integrate these new models with more established EBA modeling techniques and artefacts?
Now, answering these challenges is a whole heap of work and I don’t pretend to have any answers here or even know whether this would be a fruitful avenue to pursue. What I do know is that I agree with Nick when he worries that “our guesses are missing so many of the influences that actually drive enterprise structure and behaviour”.
In the meantime…
We are where we are and we would do well to heed Nick’s message that our current state of the art has significant flaws that we ignore to our peril.
I don’t know that we need to be too disheartened though. Reflecting back to our building architecture forebears, I have never seen an architecture blueprint for a building that explicitly includes cultural, societal, behavioural or political impacts on the design. These are the contextual elements that have informed the architecture blueprint but they aren’t actually shown in the diagram. Are we any different?
Just some thoughts from a rainy and windy UK inspired by people thinking better and more deeply than I do.