There is an online store with a high proportion of deferred conversions (a visitor comes, looks, goes, and returns several times during the month and, finally, purchases). The source of the last visit is often different from the source of the first visit (the first time man walked on contextual advertising, and when did the order of bookmarks). Of course, this spoils the performance for certain sources.
The I metric is a report on the sources with the cutoff for "returning visitors", but it is not convenient for real applications. Let me explain why.
Let's say in September I spent X amount for contextual advertising. I want to count how many orders I got for the money.Ya. Metrika -> Sources -> Summary (sampling in September) | Aim "Order":
Transitions to the is -> 10 target visits | Returning: 15 target visits
Total, it would seem: 25
IT is NOT, for 3 reasons:
1) the Users that I drew in September, and the order they did in October, the report got
2) The "returnees" were some users that were received until September.
3) If the user and the first and the last visit was from clicking on the advertisement, he counted 2 times (as a normal conversion visit returned). So stupid to put these 2 figures is impossible.What to do?
You need to design a report that would have calculated new visitors, broken at the FIRST traffic sources
. It is not difficult to do in the report designer. And here is how to count only those visitors who made the order (what order may be made in the following visits, with other sources or even in the framework of the visits falling outside the reporting period).
That's the task until it is impossible to solve the I metric. Any idea how to do it using the same Metrics or Google Analytics?