Conversion Rate

It turns traffic into a result. A count of organic visits says how many people arrived; a conversion rate says how many of them did what the business wanted. Google Analytics defines a key event as an event that measures an action particularly important to the success of the business, and any event a property collects can be marked as one^[4]. The rate is the share of sessions or users in which one happened.

Google Analytics 4 gives two formulas, not one. Session key event rate is the number of sessions in which a key event happened divided by the total number of sessions. User key event rate is the number of users who performed a key event divided by the total number of users^[1]. A session with three key events counts once in the first, and one visitor with five sessions counts once in the second. For one month the two rates differ because their denominators do.

The word “conversion” has changed meaning. Google now uses it for an important action used to measure the performance of ad campaigns and optimize bidding, and calls an event that measures an action important to the business a key event. The two can differ in reports, for reasons that include attribution settings and lookback windows^[2]. A key event rate in Analytics is not the conversion rate shown in Google Ads.

Google Ads writes the rate another way: conversions divided by the number of total ad interactions that can be tracked to a conversion. It says the result can pass 100% when more than one conversion action is tracked or when “Every” conversion is counted, because more than one conversion can be counted for each interaction^[3].

Denominators change a share more than most people expect. In QueryRev’s study, the same 896 sites are 14.2% of the sites whose robots.txt could be read, 9.0% of the whole panel and 66.9% of the sites with rules for AI crawlers^[5]. A conversion rate carries the same trap: it is a share of something, and the something has to be written beside it.

Choose the denominator before the number. For a page that is meant to turn a visit into a sign-up, a session rate fits: of the sessions that landed here, how many ended in a key event. For the question of how many of the people reached signed up, a user rate fits. Write which one stands beside the figure, in every report.

Count the numerator the same way each time. In Google Analytics 4 the key events are the events marked as such, so adding one to the list changes the rate with no change on the site. Date every edit to that list, and read organic search visits on their own, by the channel group the report uses.

Show the counts beside the rate. On a page with forty sessions, one more conversion moves the rate by two and a half points. A rate on a few hundred sessions is a range in practice, so compare it with the same period a year earlier, which also takes out a regular seasonality in demand, and read it over several weeks.

Illustrative: an invented site, to show the method.

An online shop has 20,000 organic search sessions in a month, from 14,000 users. In 560 of the sessions a key event happened, a purchase or a newsletter sign-up. 490 of the users performed one.

The session key event rate is 560 of 20,000, 2.8%. The user key event rate is 490 of 14,000, 3.5%. Both are right, and they are different numbers for the same month. A report that says “conversion rate 3.5%” in one month and “2.8%” in the next, having switched denominators between them, shows a fall that never happened.