Conversion rate stands as a pivotal metric in digital marketing, e-commerce, and user experience design. It forms a ratio that juxtaposes the number of users accomplishing a desired action against the total user count. The nature of this desired action can vary contextually but often entails actions such as making a purchase, signing up for a service, or clicking a link.
As an illustration, consider an e-commerce website where the interest lies in the conversion rate for purchases. This can be evaluated by dividing the count of users making purchases by the total number of users visiting the site. If, for instance, 100 users visit the website and 5 make a purchase, the conversion rate becomes 5/100 = 0.05 or 5%.
Within the Statsig framework, you can craft a custom metric to track conversion rates. To elaborate, you might define the numerator as unique users triggering a 'purchase' event, and the denominator as unique users triggering an 'add to cart' event. This formulation yields a conversion rate reflecting the proportion of users progressing from cart additions to purchases.
It's pivotal to comprehend that when deriving the numerator, Statsig exclusively incorporates users who also experienced the denominator event on the same day. Consequently, a user solely engaging in a 'purchase' event without a corresponding 'add to cart' event on that day will not contribute to the numerator.
Nonetheless, it's worth noting that while working with ratio metrics like conversion rates in experimentation, prudence is recommended. The possibility exists for observing an uptick in conversion rate coupled with a net reduction in total conversions. This scenario may arise if the count of unique users viewing a product (the denominator) diminishes. Hence, it's advisable to monitor both the numerator and the denominator as independent metrics when overseeing conversion rates.