The idea of an “iron triangle” is an outdated one. In mission administration, it’s the tradeoffs between price, schedule, and scope. “Decide any two” is a standard rule of thumb – that’s, you possibly can optimize for 2 however not three.
An essential variation is price, schedule, and high quality. The signal under (by way of Flickr) is a well-known model:
Nonetheless, when DevOps arrived, it appeared just like the iron triangle had run its course. DevOps aligned organizations had been delivering software program higher, cheaper, and quicker – unexpectedly! Typically, breakthroughs occur.
However as DevOps has matured and developed into product-centric group transformation, a brand new iron triangle appears to be rising. It begins, innocently, with Agile and DevOps values:
- Groups ought to be autonomous
- Groups ought to be “two pizza” in dimension (6-10 or so)
- Groups ought to be persistent; they shouldn’t be routinely damaged up and re-formed (compared to some mission administration fashions).
These three values can have lots of you nodding your heads. We’ve all heard them.
However they’re contradictory. Decide any two:
In case your workforce is small and lengthy lived, it can have vital exterior dependencies. You’ll be coping with coordination and queueing points to entry exterior companies important to delivering an final result. That is precisely the issue varied prospects of mine are actually scuffling with of their working fashions. Automation is the popular response, however quite a few in depth buyer engagements have satisfied me that there’s at all times a residual set of human-based dependencies that may trigger issues.
If you wish to prioritize “small and autonomous,” you’ll be rotating folks on and off of the workforce. Positive, you possibly can onboard vital experience, however if you wish to hold your workforce at an optimum dimension, it’s important to rotate another person off. This has a draw back by way of the workforce tradition, belief ranges (i.e., psychological safety), and for total workforce effectiveness that comes with deep understanding of colleagues’ data, expertise, and attitudes.
Lastly, in case your workforce is to be lengthy lived and autonomous, will probably be bigger than two pizzas can feed. You’ll most likely wind up with sub-teams and coordination issues, and maybe an total product supervisor with an excessive amount of on their plate. I’m listening to a number of circumstances of this as giant IT organizations experiment with outcome-orientation.
There are a lot of responses: automation (as talked about) and T-shaped professionals, for instance. However in vital environments, these are restricted. Automation helps, however it isn’t a silver bullet. Typically, the dependencies are on experience (e.g. safety session) that can not be simply automated. And generally the issues are tougher than a T-shaped skilled’s looking on Stack Overflow can reply.
So, my view is that there isn’t any “answer” to this. It’s the design tradeoffs for product-centric organizations. I’m very eager about how you’re navigating these! Please attain out or schedule an inquiry if that is one thing you’re actively engaged on.