Outage windows vs lead times: who owns the risk?

In substation upgrades, the critical path is rarely the steel and wire. It is the outage window and the parts you cannot rush: relays, CTs, breakers, and the occasional transformer. Ops gives you two weekends in October, procurement says 40 weeks, and the program wants the feeder back before summer. I see a lot of schedules that ignore that math.

What has helped on my projects: lock outage windows with operations early and drive the plan to those dates. Place long lead POs off a functional spec at 30 percent design and manage changes instead of waiting for perfect drawings. Keep a single source of truth for protection settings and firmware, tied to test plans. Bench test panels before site work. Readiness gates with go or no go checklists before mobilization.

Where I still struggle is risk ownership. Do you carry float with a mobile sub, spares, or sequenced cutovers, or push vendors to carry liquidated damages? How do you manage firmware drift and late field conditions without burning the outage? What has actually worked for you?