Percent of converters
Percent of converters is the ratio of "everyone who ended up converting" compared to the fraction of them that also did a given behavior beforehand. It indicates how much of the conversion population a given behavior accounts for.
Definition
The share of users reaching the commercial outcome event who also performed a given candidate behavior, computed as the candidate's converter volume divided by the baseline cohort's converter volume.
The coverage read in multiplier analysis
Percent of converters is one of the two analytical reads in multiplier analysis. Multiplier measures how strongly a behavior distinguishes converters from non-converters at the individual level; percent of converters measures how much of the converter population the behavior actually covers. A 10x multiplier on a behavior 3% of converters did is a much narrower lever than a 3x multiplier on a behavior 60% of converters did, even though the per-user signal is stronger.
How it diverges from multiplier
The two reads can move in opposite directions, and the divergence pattern is diagnostic in its own right:
- High multiplier, low percent of converters: a niche lever with strong individual signal but limited reach. The behavior sharply distinguishes converters, but most converters didn't do it.
- Low multiplier, high percent of converters: widespread but weakly distinguishing. Most converters did the behavior, but so did most non-converters — it doesn't separate the populations.
- High multiplier, high percent of converters: both strongly predictive at the individual level and covering a large share of converters.
Consulting only multiplier risks chasing niche levers that can't scale; consulting only percent of converters risks optimizing behaviors that aren't actually distinguishing. The two reads inform each other, and the multiplier-percent quadrant is where they get read together.
Example
In a multiplier analysis on a signup cohort against first paid conversion, "invited a teammate" might show a 6x multiplier with 9% percent of converters — strong per-user signal, narrow coverage. "Completed onboarding tour" might show a 1.4x multiplier with 72% percent of converters — wide coverage, weak per-user distinction. Multiplier alone makes the teammate invite look dominant; percent of converters alone makes the onboarding tour look dominant. Reading them together is what surfaces that the two behaviors are doing structurally different jobs in the conversion population.