Every plan runs on two clocks: effort and calendar

Every modernization plan runs on two clocks, and mixing them up is the most common estimating failure we see.

Effort time is engineer-hours: how much work the migration is. Calendar time is when things are allowed to happen: the freeze windows, the vendor's certification cycle, the quarter-end nobody deploys near, the auditor's schedule, the retirement date of the one person who can bless the cutover. Effort time responds to headcount. Calendar time does not care how many people you hire.

The failure pattern: a plan priced in effort time, presented in calendar time. "Six engineer-months" becomes "done by March" on a slide, and March arrives to discover that the payment processor only certifies integrations in two windows a year, both missed, or that the change freeze around fiscal close ate five of the plan's weeks. Nobody underestimated the work. They never listed the calendar.

So list it. Day one of any program, before the architecture work, one page titled "when things are allowed to happen":

1. Freeze windows, all of them, including the unofficial ones.
2. External certification and approval cycles, with their actual dates, from the vendor's mouth, not from hope.
3. Compliance and audit gates that can veto a cutover.
4. The people calendar: retirements, sabbaticals, the contractor whose contract ends in June.

Then schedule backward from the calendar constraints and forward from the effort estimate, and let the two meet in the middle. Where they cannot meet, the plan has found its truth early, which is the cheapest place to find it.

Adding engineers shortens one clock. Only starting earlier shortens the other. Most late programs we are called into were never behind on effort. They were behind on calendar from the day the plan was signed, and nobody had drawn the second clock where the sponsor could see it.

Draw both. Date the wall.