Information about expected data variances between Target and Adobe Analytics when using and not using Analytics as the Reporting Source (A4T). A4T significantly reduces data variance.
With A4T, both Analytics and Target reporting of activities use Analytics data exclusively, so there is little variance between the solutions in Target activity reports. In some circumstances, however, customers might compare Target data to Analytics data outside the scope of the A4T integration and, thus, experience the variance issues described below.
Here are a few scenarios in which you might experience expected data variance:
A4T allows for the possibility that a Target hit (top of page) occurs, but no Analytics hit (bottom of page) occurs. An example of this would be if the user loads the page, but closes the browser before the Analytics call triggered. In these cases, A4T excludes the Target hit from our data. The reason for this is that allowing Target hits (again, top of page) to count as Analytics hits in the absence of an actual Analytics call would create inconsistencies with the data set in Analytics (visitor inflation, etc.).
If a redirect test is set up in Target to split traffic 50/50 (or 25/25/25/25, etc.), user behavior might not be divided evenly. If you see an uneven split, it simply means that one group of users failed to execute an Analytics call on the landing page more than the other group(s) did. This failure to execute the Analytics call for one group caused the Target hit for that user to be excluded, creating the unevenness.
This is something we hope to address in the future as we work toward A4T on the Adobe Experience Platform. Our teams are working through how best to handle these different events occurring at different times on the page.
A known issue exits that is causing a limited number of customers using redirects with A4T to see a higher percentage of unstitched hit rates. See Known issues and resolved issues.
Suppose you create an Auto-Allocate activity open to all visitors to a particular page. Because Auto-Allocate activities don’t support A4T, all of the activity data is collected by Target. You might expect that the visitors to the activity in the Target reporting should match the visitors to that page in the Analytics reporting for the same date range. This is a scenario in which the variance described below is expected.
For a complete list of activity types that support A4T, see Supported Activity Types.
Variances of 15-20% are normal, even with similar data sets. Systems that count differently can result in much higher data variances, as much as 35-50%. In some cases, variances can be even higher.
Although actual data can vary significantly, trends are usually consistent. As long as the differences and trends remain consistent, the data remains valuable and useful. If the differences and trends are inconsistent, it could mean that something is set up incorrectly. In this case, contact your account representative for assistance.
Analytics uses a system based on visits and transactions, but Target uses visitor-based metrics. That means that whenever a visitor opens a page, it counts as a visit in Analytics, but Target does not count the visit until the conditions set in the activity are met.
Reports in Target show performance based on the conversion mbox selected when defining the activity, but this conversion mbox data is not sent to Analytics, which has its own conversion variables as defined by your Analytics tagging implementation. In cases where you might expect identical data (for example, if a retailer’s order confirm page contains both a conversion mbox and an Analytics purchase event), data might differ due to the placement of these tags. In general, trends in the two products’ reports should be the similar.
Expected data variances can be caused by both technical and business variances.
The following can cause data variances based on technical differences:
The following can cause data variances based on business differences: