This week I was listening to a great podcast conversation between Ian Gill and Andy Bold, focusing on Andy’s area of expertise, which is devops. (Disclaimer: I am part of the Agility by Nature network, of which they and that podcast is a part.)
Andy’s work is often about improving an organisation’s delivery capability on the infrastructure side—less about writing the code, more about getting that code into production fast, making sure it’s secure, that it can scale up, that costs are managed, that system changes can be applied quickly and reliably without disrupting the service, and so on.
One of the rhetorical questions Andy and Ian ask early on is, “What’s the point of doing all that work if the rest of the organisation won’t change?” The delivery of the system into production is one link in a chain, one section of a pipeline. Before that point there is the creation of the software system. After that point is the impact these changes have on the rest of the organisation—new sales opportunities, different kinds of support calls, internal operational changes needed to handle the new things users are doing, the new data that’s coming through, and so forth.
With that picture in mind, we can see that the infrastructure improvement has a diminished impact if the sections of the pipeline (or links in the chain) before and after can’t respond accordingly. IT might have been a bottleneck before, and no longer—which is great—but now there is a new bottleneck somewhere else. (This is an observation that leads to the Theory of Constraints.) That is why Andy says his work is often not just in the world of technology infrastructure, but also looking at other parts of the organisation to ensure that his optimisations have the maximum impact.
This is not just true in devops. It’s true for any part of an organisation that can be seen as a section in a pipeline. I’ve seen it in software delivery; sales people have seen it in sales.
So what happens when we discover the next bottleneck? The scenarios are numerous. Ideally the organisation will have anticipated the challenge and be keen to turn its attention to that. Repeating this creates a continuous cycle of improvement. But sometimes there is surprise or resistance to change elsewhere. Perhaps it does happen eventually. Perhaps the people who succeeded in the first change become disillusioned with the limits of the organisation. Perhaps they are at least satisfied that their work has improved.
Optimising one part of an organisation often throws the spotlight onto other parts next. That’s why local improvements can have much bigger consequences.