Rich-text Reply

Running test for repeat visitors

avarshneya 11-13-14
Accepted Solution

Running test for repeat visitors



I want to run a test where only the repeat visitors are subjected to the test. Is there a way I can do that natively in Optimizely or do we have to write cookies for such a test?





Amanda 11-14-14

Re: Running test for repeat visitors

Hey @avarshneya - Check out this Help Article on New vs Returning Visitors:


Note that whether a visitor is counted as new or returning is session based, not necesarily experiment based. Does this sound like it will do the trick for you?


If you are looking for something more specific, please let us know and we can suggest another option. 

avarshneya 11-20-14

Re: Running test for repeat visitors



Hey Amanda,


Good call, that is what I'm looking for. However, I want to make it a bit more complicated and see if the page they see on their RETURN impacts conversion rate. So in other words, 100% of New visitors will see the control page. Upon returning, I want to do a split test between the control page and Variation B.


My Variation B page is much simpler than the control and I want to see if that added focus on the form will help conversion rates amongst returning visitors.


Is that possible?

Amanda 11-21-14

Re: Running test for repeat visitors

[ Edited ]

The method mentioned above should still work, just set the audience targeting to returning users only.  This will cause your test to work as follows:


1. The first time the visitor arrives at your site, he/she will be counted as "new" for the whole session duration and thus won't meet the targeting conditions and the experiment won't run. The control version of your page will be seen as a result

2. If the user comes back again, he/she will be counted as "returning" and the experiment *would* run. The visitor would be randomly allocated into the control or variation in accordance with your traffic allocation. 


Does that sound right? 

avarshneya 11-21-14

Re: Running test for repeat visitors

That's perfect, thank you!