When the Deadline Is Fixed and the Estimate Doesn't Fit
A conference date, a contract, a regulatory cutoff, a competitor's launch. Some deadlines are real and immovable. Then the team estimates the work and the total comes to eleven weeks against eight weeks of calendar.
What happens next determines whether the team's estimates stay worth anything.
The response that destroys the practice
The tempting move is to send the team back to "re-estimate more aggressively". The numbers come down, the plan fits, and everyone relaxes for a few weeks.
This teaches the team that estimates are negotiable, which means they stop being estimates. It also doesn't change the work; it only removes the warning. You'll arrive at the deadline with the same eleven weeks of work and three fewer weeks to react. This is the most expensive of the estimation anti-patterns because it burns the instrument you'd use to catch the problem early.
The four levers that actually exist
When effort exceeds calendar, only four things can move:
- Scope: ship less. Almost always the right first lever, and the only one that reliably works late in a project.
- Date: ship later. Off the table by definition here, but worth explicitly confirming rather than assuming.
- People: add capacity. Slow, and on a project already underway, often negative in the short term.
- Quality: ship worse. Sometimes legitimate for a genuinely temporary situation, always a loan against later sprints.
Notice that "estimate differently" is not on the list. It isn't a lever; it's a way of pretending you pulled one.
How to run the scope conversation
Bring the backlog, the estimates, and a line drawn at your realistic capacity for the remaining calendar. Ask the product owner to order everything above and below the line. Not "what can we cut" (people can't answer that) but "what order do these go in".
Then say plainly: everything above the line is what we expect to ship; everything below it is what we expect to leave. If that set isn't a launchable product, the problem is the deadline, and the earlier that's established the more options exist.
Say it in ranges, once
Don't offer a single date you don't believe. Give the range your estimates support ("at this scope, six to nine weeks; the fixed date needs the six-week version to go perfectly") and let the person who owns the deadline decide with real information. That's the whole job of an estimate.
Re-check weekly, not at the end
Fixed-date work needs a much tighter feedback loop than normal. Compare completed points against the line every week, and raise the scope conversation again the first week you're behind, not the week before launch. See using historical velocity for how to set the line honestly in the first place.
An estimate that clashes with a deadline is doing its job. The clash is the information; deleting it doesn't move the date.
