Looking at that Analytics graph, it's pretty obvious the number of signups is around 5 on the high end, and 2 on the low end. So, in this case, increasing a conversion rate by 46% is the difference between 3 sign ups instead of 4.
The sample size is so tiny, it's hard to argue that any language changes truly affected conversions.
I'm also noticing a downward trend on the blue line over the month. Whatever was working earlier in the month stopped.