Two Checkout Pages, One Verdict

An online shop is testing a redesigned checkout page. Every visitor is randomly shown either the current page (A) or the redesign (B), and the shop records whether they complete a purchase:

The null hypothesis $H_0$ says the redesign changes nothing: both pages share one true purchase rate. The test statistic compares the two sample purchase rates $\hat p_A$ and $\hat p_B$:

$$z = \frac{\hat p_B - \hat p_A}{\sqrt{p_0\,(1 - p_0)\left(\frac{1}{n_A} + \frac{1}{n_B}\right)}}$$

Here $n_A$ and $n_B$ are the visitor counts, and $p_0$ is the single purchase rate that $H_0$ says both pages share, estimated from this experiment. (Each visitor is a Bernoulli trial with variance $p(1-p)$, and the variances of two independent groups add — that is where the denominator comes from.)

What is the value of $z$?

Round your answer to 2 decimal places.