How to compare search and network conversion
Contents
Share
You cannot compare search conversion and network conversion as one number. Clicks and conversions of display campaigns on search are not compared with performance campaigns: the demand is different. And even inside one campaign, the day of the conversion depends on which report is open.
The checked guides do not give a normal gap of “search minus networks”. Below is which pairs can sit side by side, and which dates have to match. A single CTR for a campaign, mixing places, is covered in what a normal CTR is. This article is about conversion, not click-through.
Different demand does not go in one column
The guide to display-campaign statistics says it directly: clicks and conversions of display campaigns on search are not compared with performance campaigns, because the demand is different.
That is a ban on the comparison, not a benchmark. The page does not say by how many points network conversion “should” sit below search. If one line of yours mixes a display campaign on search and a performance campaign, a falling or rising combined figure is not a diagnosis. Split the types before you name the networks as the cause.
A placement CR from the Placements report is already networks: the report is filtered to the type Networks. Putting that CR next to a search conversion from another report and calling the difference “search versus networks” does not work until both sources are named. How a placement row is built is covered in suspicious placements in YAN. How to assemble your own slice in the wizard is in how to build a placements report.
One conversion lands on two days
The report wizard attributes a conversion to the click date. The pay-per-conversion report attributes it to the conversion date. In Yandex’s example a click on 3 July and a conversion and charge on 5 July land on different days in those two reports. The example is from the strategy guide, not your campaign’s calendar.
| Where you look | Which day the example uses |
|---|---|
| Report wizard | 3 July, the click day |
| Pay per conversion | 5 July, the conversion day |
Comparing search and networks “for 3 July” works only inside one dating rule. Otherwise a network whose conversion arrived on 5 July looks empty on the day search already showed the click. How a week then looks empty of leads is covered in clicks without leads.
The Metrica summary is already without the removed clicks
If you read search and network conversion in Metrica, the report Direct, summary is based on clicks that passed fraud protection. Other slices do not repeat that filter. Comparing a CR from the summary with a CR from a report that still includes removed clicks means comparing different bins.
The summary path and the “Other campaigns” case are covered in how to read the Invalid traffic report. Here it is enough not to mix the “after the filter” window with the “all visits” window and call that the difference between search and networks.
On one or two conversions the price can differ sharply from a stable one. A thin network row with one conversion and a thick search row with dozens are not a pair for the conclusion “networks are worse”. That is the placements guide’s caveat, and it applies here too.
A check on the site does not calculate search and network conversion and does not count money. The decision on a visit is not labeled search or network. If there is no answer, the visit is let through.
Order of comparison:
Take display campaigns on search out of the pair with performance campaigns. The guide tells you not to compare them.
Use one report and one dating rule: either the click day or the conversion day.
Label a placement CR from the Placements report as networks, not as “all traffic”.
Do not put a post-antifraud summary next to a slice that still includes removed clicks.
Leave a row with one or two conversions out of a “which is worse” comparison.