Highlighted

A/B Test with redirect: Control version counted multiple times

christiank86512

04-07-2018

Hello,

we have started an A/B/N Test with overall 6 versions of a landing page (incl. control version). Each of the test versions (B-F) redirects to a different url (landing-page-b.html, landing-page-c.html, etc.) 

It is set up in Target, that the traffic is distributed by 20% for A and 16% each for the 5 test versions (B-F). However, according the Adobe Analytics tracking, roughly 80% of the traffic go to A and 20% to the 5 other versions (4% each).

Does anyone have an idea where the problem might be?

Thanks in advance!

Replies

LJ_Jones

05-07-2018

How in Analytics are you reporting the test?  Are you using A4T?  If you are using A4T, then you have to be using the at.js implementation of Target.  Otherwise you may have problems reporting in Analytics.  Have a look at the link below.

https://marketing.adobe.com/resources/help/en_US/target/a4t/c_a4t-faq-redirect-offers.html

Highlighted

christiank86512

11-07-2018

Thanks for your Feedback LJ Jones​. Yes we use A4T and the at.js (we ensured that it is the latest version).
We partially solved the problem by also setting a redirect for version A of the test (linking to the default landing page). Now the traffic is split among the experiences as defined in the test setup.

However, the numbers between the Target Report and e.g. the Pages Report in Analytics are still very different. E.g. the number of unique visitors in the Target Report is by approximately 90% lower than in the Pages Report (same for the visits).

Highlighted

LJ_Jones

12-07-2018

When comparing two different reports, its important to make sure that you are comparing apples to apples.  In other words, for the two reports that you are looking at, make sure that you are looking at the same segments, audience and that its a true comparison.

You said Page Reports.  Are you just looking at the page report?  Meaning, are you looking at all traffic to those pages?  Or are you looking at a Target Activities report that would show the visitors in the test.  My point is that if you are looking at a Pages Report, you could be looking at people who visited the page you aren't in the test. This would explain why its lower in Target.

I prefer to build out a view for my Target Test in Workspace. Something like this.  A saved report like this saves time and is easy to reference.  Plus you can look at a lot of different metrics.

1526696_pastedImage_0.png

Highlighted

inges98599238

16-07-2018

Hi,

I'm facing exactly the same issue as described.

See below the screens the report in Target and below the report in a workspace. There is one different but the numbers are nealry the same.

Bildschirmfoto 2018-07-16 um 16.24.26.png+Bildschirmfoto 2018-07-16 um 16.25.44.png

So far so good:

Now we checked the number of the UV for the page (experience D). The number of the UV for this page is much more higher? At the same time if i apply the segment to see how many visitors of this page were part of the success - the number is lower than in Target. The only way to get to this page is to be redirected to it. And as we're talking about UV there is in my opinion no why how the same visitor could be counted twice in AA?

Bildschirmfoto 2018-07-16 um 16.31.37.png

Any ideas?