Not because of incompetence or bad intent, but because early decisions tend to harden faster than understanding does.

A few observations that seem to repeat:

Early schedules become social facts Initial timelines are usually created with limited information, but once shared upward they quickly stop being provisional. They become anchors for funding, reputation, and confidence. Later evidence is forced to fit the date, rather than the date adjusting to evidence.

Risk is recorded instead of managed Risk registers are often thorough and sincere, but the act of documenting risk can substitute for actually changing decisions. Addressing certain risks would require revisiting scope, sequence, or assumptions — which is often seen as destabilising rather than responsible.

Governance filters reality Reporting structures tend to optimise for reassurance. Bad news is delayed or softened, not out of malice, but because it feels unconstructive without a solution. By the time issues surface clearly, the remaining options are usually expensive or binary.

Complexity is deferred, not reduced Early approvals tend to reward simplicity. Interfaces, dependencies, and operational constraints are minimised to get things moving. The complexity doesn’t disappear — it just shows up later, when flexibility is lowest.

What’s struck me is that many projects don’t so much “go wrong” as proceed logically from the assumptions they were built on - which were often incorrect from the outset.