Most operational breakdowns are visible in retrospect. The warning signs were there for months — patterns in how work got done, or didn't get done, that pointed to structural problems before anything actually collapsed. The hard part is learning to read them before a customer complains, an auditor asks, or a key person leaves.
Sign one: you find out something wasn't done after it should have been done. Not in real time — after. The onboarding that nobody completed because everyone assumed someone else was handling it. The approval that stalled because the usual person was on vacation and no backup was defined. When your system for knowing what's happening is 'someone would have said something,' invisible work is already accumulating.
Sign two: the same process produces different results depending on who runs it. The new hire onboarding looks completely different depending on which manager is doing it. The client kickoff has five variations across the CS team. Inconsistency isn't a performance problem — it's a design problem. The process was never specified well enough to be repeatable without interpretation.
Sign three: you can't answer 'was this step done?' without asking someone. If verifying execution requires a conversation, you don't have a system — you have coordination. And coordination doesn't scale. At some threshold of team size and process volume, the coordination overhead becomes the job, and the actual work starts slipping.
Signs four and five: new hire quality depends on who's available that week, and you're thinking about hiring someone to manage coordination between people who are already there. Both are symptoms of the same structural gap. Before you add headcount to manage the symptom, it's worth checking whether the underlying process can be made self-managing first.