It makes a-lot of sense. You are saying that one must view the full set of digital platforms as a single system, and whenever making a change to that system, one must consider the full scope of what is impacted, and design a team (or teams) that can handle that full scope.

That’s a big ask, but I agree with it. Assuming that organizations will get it right half the time, that means there needs to be a safety net for mistakes, which implies some kind of design-time or test-time validation. E.g., Google is known for having massively comprehensive integration test suites that kick off automatically.

There is also the issue of trying to do better than “get it right half the time”, particularly with the ability of the teams to work across the entire range of the system. Suppose one has a IoT app for which millions of events are arriving every minute, which need to be collected and logged, passed to a stream processing system, and ultimately data presented to users in dashboards. To work across that entire stack, the team might need to know the onboard stack (say, C++), as well as event and stream processing, as well as microservices and the Web/mobile progressive that present data to end users. That’s quite a-lot. What would be your thoughts on a team approach in that case?