Instead of explaining why a database migration is hard, I frame it in terms of business impact: "We can rush this in a day, but it increases the risk of system downtime and slows down the next three features."
When deadlines are unrealistic, I switch to a variable scope menu: "We can hit that date, but we have to drop features X and Y to make it happen. Which one matters less for launch?" Turning a flat "no" into a collaborative trade-off conversation changes everything.