You observe a conversion change after a product update. In the pre-change period, 18,240 of 152,000 users converted, and in the post-change period, 19,656 of 154,000 users converted. You want to determine whether the observed increase is likely a real change in conversion rather than random sampling noise, using a 5% significance level.
Is the observed conversion lift statistically distinguishable from noise, and how would you quantify the uncertainty around the estimated change?