I'm running an A/B test right now of a rather subtle paywall copy tweak, and the results are very surprising.
This is off the back of a previous paywall experiment which increased conversions by 39%
Which represents a 119% improvement in 30 days... if I've done that maths right