Ask most institutions to describe their operating model and you'll get a reasonably coherent answer, usually accompanied by a chart. Ask them how much of that model was actually designed, deliberately, versus accumulated through a decade of individually reasonable decisions, and the answer is usually far less coherent.
Operating models rarely get built from scratch. They accrete — a reorganisation here, a new reporting line added to accommodate a departing executive there, a process exception carved out for one difficult client that quietly becomes the template for how a whole category of client gets handled. Each individual decision was reasonable at the time. The cumulative result was never designed by anyone.
This matters because institutions frequently try to fix operating model problems by adjusting the visible chart, without recognising that the chart is downstream of years of undocumented accretion that the chart doesn't actually capture. You can redraw the boxes as many times as you like; if the underlying accreted decision patterns aren't addressed, the new chart will accrete its own exceptions within months.
The diagnostic work that actually helps is tracing the accretion — understanding which parts of the current model were deliberately designed and which parts are historical accident that nobody has revisited in years, simply because revisiting it would mean reopening decisions that feel settled even though they were never actually made on purpose.
Most operating models are not badly designed. They were simply never designed at all, in any comprehensive sense, and treating an accreted structure as if it were a deliberate design is the mistake that keeps transformation programmes from addressing the actual problem.