The hidden cost of bad subtitles (and how to measure it)
Mistimed lines and over-literal translations don't announce themselves. They quietly bleed retention, one confused viewer at a time, which is exactly why they're so easy to under-invest in.
Bad subtitles rarely trigger a complaint. Viewers don't file a ticket; they just drift. They rewind once, then twice, then decide the show is "hard to follow" and open something else. The damage is real but invisible, until you decide to measure it.
The churn nobody attributes to subtitles
When we survey lapsed viewers of localized titles, a surprising share trace their drop-off to the reading experience, even when they don't name it directly.
What "bad" actually means
"Bad" is not one problem. It's a stack of small failures that compound:
- Timing drift, captions that lead or lag the audio by more than a beat.
- Reading speed, lines that exceed comfortable characters-per-second.
- Literalism, word-for-word renders that ignore idiom and register.
- Line breaks, clauses split in places that fight comprehension.
A quality score you can track
You can't manage what you don't score. We grade every file on four dimensions and roll them into a single index, so quality becomes a number a producer can actually watch over time.
The retention it buys back
When we re-timed and re-adapted a backlog of underperforming episodes, no new footage, same translation budget, completion rates moved measurably.
The takeaway isn't "spend more." It's "spend visibly." Put a number on subtitle quality and it stops being a matter of taste, it becomes a lever you can pull on purpose.
