Field notes · 2026-05-12
Reading a completion curve without panic
Finish rates look alarming until you separate opening friction from mid-story abandon and true ending skips.
A steep early drop on a completion curve is not automatically a failure of the title. Many media apps see viewers sample the first minutes, then leave because the promise of the episode did not match the thumbnail or the previous episode’s cliffhanger.
Start by marking three windows: the opening sample (first few minutes), the mid-run hold, and the final stretch. Compare those windows across similar runtimes. A documentary with a slow cold open will never mirror a short-form comedy’s finish shape, and treating them as peers hides the real story.
When the mid-run hold collapses, look for interruptions that belong to the product experience: mid-roll prompts, forced sign-in walls, or bitrate stalls that coincide with the abandon spike. Editorial teams often blame the cut when the stall sits on a network handoff in evening peak hours.
Ending skips deserve their own note. Viewers who jump the credits still completed the story. If your event marks “completed” only at credit end, your finish rate understates satisfaction for titles with long credit rolls or post-credit stingers.
Bring one concrete question to each review: which window moved, for which title cohort, and what changed in the product or the cut during that window. Panic fades when the curve is read in pieces instead of as a single percentage.