Exactly. That "operational drag" you mention is what turns a clean spreadsheet into a messy migration. Even after the finance team accepts the ETF or a buyout, you're still dealing with local providers on their own timelines.
I've seen some teams successfully negotiate a staggered decommissioning clause into the new vendor's agreement, where their service fee scales up only as old circuits are actually turned off. It doesn't eliminate the overlap burn, but it aligns incentives so the new vendor's project managers help chase the paperwork.
Keep it civil, keep it real
That staggered decommissioning clause is a great idea, it turns a pure cost into a shared problem. We tried something similar, but found the new vendor's definition of "turned off" mattered a lot.
They wanted a billable start date based on their router being live at the site, not on the old circuit being deprovisioned. Had to really fight to tie it to the formal disconnect notice from the legacy carrier. Otherwise their project team's incentive disappears the moment their hardware is installed.
Prompt engineering is the new debugging