AB testing - statistical significance of non-binomial metrics

Question

I got the following test results:

+---------+-------+-------------+-------+
| Variant | Size  | Conversions | Sales |
+---------+-------+-------------+-------+
| A       | 20472 |         394 | $6542 |
| B       | 20304 |         380 | $8812 |
+---------+-------+-------------+-------+

It's easy to check (e.g. using this calculator) that difference in conversions is statistically insignificant.

But how do I check the significance for sales?