The general form of 'applying a process from some other company' seems to be:
1. Read the process.
2. Dig for surrounding context around the creation of the process, and think critically about the differences.
3. Pick a small change to introduce as an experiment.
4. Iterate.
Conversation
Skipping any of these steps seems to lead to bad things:
Not doing (2) means not picking a small change intelligently (3). Not iterating (4) increases the odds of process rejection.
1
1
Anyway, I'm mentioning this mostly because I just read a post on continuous integration/delivery in software, and how it's is a universal, unalloyed good. And I'm comparing it to an anecdote from a friend at Apple, who said slowing down releases improved OS software quality.
1
4
Slowing down releases to a manual process* improved OS software quality.
You really do need to check the context from which 'best practices' emerge; the vast majority of the time, the writers from that context argue in universals, and don't consider adjacent contexts.