curious if anyone’s run A/B tests on a low-traffic site (~500 clicks/week) by optimizing micro-conversions (e.g. ATC clicks, scroll depth) instead of final purchases. Have you tried this approach to ramp up the number of tests per month, and what kind of results did you see?

For low traffic page (and I'll do the same for high traffic too), I'll use heatmaps by device type and individual session screen recordings, which I'll manually sample to get a sense of the user's use/confusion with the site. For A/B tests, I typically like to have a couple thousand sessions with anonymous users through a funnel to make any decisions based on singular outcome/data points. I'd couple getting down the funnel/micro-conversion with final purchases data to get a sense of customer interest leading up to purchases. Note that it is good to filter out noise (bounces/bots/internal tests) so you are only isolating real users only.