First traffic source and conversion analytics

Question

Conversion

There is an online store with a high percentage of delayed conversions (the visitor enters, looks, leaves, returns several times during the month and finally buys). At the same time, the source of the last visit often differs from the source of the first visit (the first time a person came through contextual advertising, and when he made an order, already from bookmarks). Of course, this greatly spoils the efficiency statistics for some sources.

Y.metrica has a report on sources with a slice for “returning visitors”, but it is not convenient for real use. I'll explain why.

Let's say in September I spent X amount on contextual advertising. I want to count how many orders I received for this money.

Ya.metrica -> Sources -> Summary (selection for September) | Purpose "Order":

Ad clicks -> 10 targeted visits | Returnees: 15 targeted visits

Total, it would seem: 25

THIS IS NOT SO, for 3 reasons:
1) Users that I attracted in September, and they made an order in October, were not included in the report
2) Some users who were attracted more were included in the "returned" until September.
3) If the user's first and last visit was from an ad click, it will be counted 2 times (as a regular conversion visit and as a returned visit). Therefore, it is impossible to stupidly add these 2 numbers.

What to do?
It is necessary to design a report in which new visitors would be counted, broken down by FIRST traffic sources. This is not difficult to do in the report designer. But how to count only those visitors who made an order (moreover, the order can be made within the next visits, with other sources, or even within the visits that do not fall within the reporting period).

Here is such a problem, until it is possible to solve it in Ya.metrica. Any thoughts on how to do this using the same Metrics or Google Analytics?

Answer 1

If the buyer entered through contextual advertising, then he was assigned to it by cookies. So if he returns and the cookies are saved, it will be counted.
In Google Universal (new analytics), you can even track if a visitor came from different devices to complete the conversion.
The second option, which is sometimes used to ensure that cookies are not overwritten if several advertising campaigns are launched in one system, is to send an additional GET parameter from advertising campaigns, which will be written to cookies for an unlimited period of time. Each RK has its own cookie file name. Therefore, they will not be overwritten and can be counted.
If the user thought for a month before buying, then you will see him with a monthly delay, otherwise there is no way.
If the first visit did not complete the conversion, and the last visit completed the conversion and these were ad clicks, then it will not be counted twice.
The conversion will be calculated only for the last one (RK cookies are overwritten with fresh ones - for Ya.Direct for sure).
Still, I recommend tracking Google Analytics with the new universal code (you can use the old and new at the same time and compare statistics)