Like every organisation I've ever worked for, we currently have issues joining up data -from different, older, siloed systems. Bringing it all together means operations improve, automation becomes possible, the dreamt of digital organisation becomes a real possibility, it opens up analytics to improve decisions, yes in the boardroom but in real time at peoples desks while they are taking calls.
The problem is it's hard.
I've lost count of the number of days I've spent in a room with the subject matter experts, a plate of donuts on the table, mapping system to system to system.
Hard slow grindey work, but with a huge payoff when it's done successfully.
So I've spent many years looking for magic.
Magic that will make this easier, quicker and involve fewer donuts.
We recently did a PoC witih a product that promised me magic. It promised that it could integrate our - very - disparate data quickly and easily and make it sing to our users from sales and fulfilment and supply-chain.
The problem is, I don't believe in magic so I was cynical but excited,
So after six months instead of the promised three and with a massive reduction in scope they declared the PoC complete and successful. Except there was no magic, only a talented and experienced data guy who had spent the last 6 month grinding through our data and joining it up by hand.
We won't have a problem walking away from the software: nobody liked it, but the data that guy joined up....that could be worth a lot.
So
I still haven't found magic , I still don't believe in magic and I've been right every time so far
But
Technology, especially AI is looking more and more magical every day.....so maybe one day....I'll be able to lay off the donuts.
Business Capability Modelling, some thoughts...
Business capability modelling is the easiest and hardest thing to do in enterprise architecture.
There's a lot of temptations and a lot of excessive expectations.
The temptations are to try to be too perfect, to get into too much detail, to get to the ideal model where everything is comprehensively modelled once.
The expectation is that by having the capabilities modelled in the perfect fashion the rest of the architecture will become easy.
As is so often the case, we need to go back to what we are - practically - trying to achieve via our capabilities and how do we navigate as close to that as possible.
We are trying to group what our organisation does into hierarchical categories, that, will help us understand and manage what we do and what we will do in the future.
That's not a perfect thing it's a practical thing.
Architects aren't always happy with imperfect things, they are also paid to have opinions (they like it too) which means that capability modelling can go on forever.
For a start Hierarchies are great and useful structures but they are rigid rigid rigid, the real world isn't like that. So use them practically, create the hierarchy that most closely matches what you do but don't sweat when there's a little duplication or subtlety to some of the capabilities. The generalised functions e.g. Finance and the more specific - project finance - often cause these problems.
Don't try to go too deeply. In a hierarchy, every time you split into another level you are attempting to include one lot of capabilities and absolutely exclude the rest (theoretically). So don't do too much of it. Leave some wriggle room for ambiguity, it won't be perfect but driving down down down won't solve it either. 80:20
Try to ensure that your capabilities sound like what you do. Not something completely generic, most successful firms have a competitive advantage, that should be clear in the model.
Finally, having a semi-thought through model is generally better than no model, if you have sensible but quite imperfect capabilities you can still attach systems and people and data and investments and projects to them. It's better than doing it by feel.
Add comment
Comments