Case Study
An education funnel in fitness had unstable trial-to-paid conversion due to mixed channel quality. The team rebuilt source selection around lookalike discovery and strict validation before scaling paid placements.
+24%
trial-to-paid conversion rate
-22%
early unsubscribe rate
5 weeks
to stabilize funnel acquisition
3 waves
used to calibrate source mix
The team saw acceptable trial volume, but paid conversion fluctuated heavily between waves. Acquisition came from channels with inconsistent audience quality.
Operators seeded Similar Channels from top trial-converting sources, validated output in Channel Analysis, and launched with unique attribution links for every channel group.
A pool of 110 discovered channels was reduced to 36 after quality checks. Two waves later, 9 sources were removed for high early churn and replaced with higher-retention lookalikes.
Within 5 weeks, trial-to-paid conversion improved by 24% and early unsubscribe rate dropped by 22%. Paid funnel quality became stable enough for predictable weekly planning.
The media and funnel teams moved to one shared KPI board: qualified trials, early churn, and paid conversion by source cluster.
Lookalike expansion preserved audience fit, while hard quality gates prevented noisy channels from entering scale phases.