You work on a work-management product and run an A/B test on a small onboarding change in the project creation flow. The treatment shows only a modest lift on the primary conversion metric, and some stakeholders argue that the gain is too small to matter while others want to ship it.
How would you explain why a test with a small lift can still be worth shipping? What factors would you use to decide whether that lift is meaningful enough to launch?