Analytics
5.8K members online now
Understand information in your reports and troubleshoot reporting issues such as self-referrals, (not set) data, and inaccurate information
 
Guide Me
star_border
Reply

Experiment self-referrals conversion probelm

Visitor ✭ ✭ ✭
# 1
Visitor ✭ ✭ ✭

Hi everyone,

 

I created experiment so I could run A/B test for complete registration page.

Experiment code is implemented on original page, and universal GA code is set both on original and test page.

 

Both pages are on same domain.

Original: www.example.com/registration.php

Variant 1: www.example.com/registration_.php

 

 

The problem is that since the experiment have started all conversions are shown as self referral, so I cannot track real source of conversion, and performance of campaigns.

 

I have noticed that when someone comes directly to registration page GA ads Experiment ID and Experiment key to the URL. But when someone come from other page to registration page it ads utm_referrer my domain, and the last page that was visited. At the end that's what I see as a conversion.

 

Does anyone what can I do to see the real source of conversions?

 

Thank you in advance.

1 Expert replyverified_user
Marked as Best Answer.
Solution
Accepted by topic author Boris P
June 2016

Re: Experiment self-referrals conversion probelm

Top Contributor
# 2
Top Contributor
Hi :-)
I dont have much experience with experiments.
Having said that..
1) is your domain name (www.example.com ) on the Referral Exclusion List as it should be when using Universal Analytics.
2) Do all website pages on the domain have the Universal Analytics tracking code added to them
Bronwyn Vourtis, Google Analytics Top Contributor
Was my response helpful? If yes, please mark it as the ‘Best Answer.’ Learn how here

Re: Experiment self-referrals conversion probelm

Visitor ✭ ✭ ✭
# 3
Visitor ✭ ✭ ✭
Hi,

Thank you for help.

The problem was because example.com was not added in Referral Exclusion list.
Some conversions still appear as self referral, but I suppose it's because of Campaign Timeout in Session Settings is set to 6 months.

Regards,

Re: Experiment self-referrals conversion probelm

Top Contributor
# 4
Top Contributor
Hi Boris, happy to hear you now have this working as expected

Thanks
Bronwyn
Bronwyn Vourtis, Google Analytics Top Contributor
Was my response helpful? If yes, please mark it as the ‘Best Answer.’ Learn how here

Re: Experiment self-referrals conversion probelm

Visitor ✭ ✭ ✭
# 5
Visitor ✭ ✭ ✭
Is this a new thing? I've had the same issue with an experiment the last week, but I've done many experiments in the past and they never lead to self-referrals.

Quite frustrating that needing to add your own domain into the referral exclusion list is necessary. For one, you'd think that would be default behaviour. For two, you would think that would be mentioned when setting up an experiment.

Re: Experiment self-referrals conversion probelm

[ Edited ]
Top Contributor
# 6
Top Contributor

The Referral Exclusion List is relevant to Universal only.
When a new Account is created and a Property is generated, or another Property is created within an existing account,
the domain is added to the Referral Exclusion List by default.

 

Where it can get a little tricky, is if for some reason the domain name was removed from the Referral Exclusion List
Or if it is an older Property that was not originally Universal and had not been upgraded to Universal analytics.js by the Account owner,

and instead the Property was auto migrated by Google.

With auto migration of the Property, its not compulsory (though certainly recommended) to update the tracking code on the applicable site(s). A Universal Property is capable of tracking and reporting from the older ga.js tracking code.

So at this stage, it is up to the Analytics Account owner to ensure the domain name gets added to the Referral Exclusion List at the time they decide to upgrade the tracking code from ga.js to universal analytics.js on the applicable website

Bronwyn Vourtis, Google Analytics Top Contributor
Was my response helpful? If yes, please mark it as the ‘Best Answer.’ Learn how here

Re: Experiment self-referrals conversion probelm

Visitor ✭ ✭ ✭
# 7
Visitor ✭ ✭ ✭
Ah, that explains it. The last experiment _may_ have been (must have been, given what you describe) when we were still using ga.js. Here's hoping the Analytics team may address this so any other people making the update in the future don't lose hours of head scratching on the same issue.

Thanks.

Experiment self-referrals conversion probelm

Visitor ✭ ✭ ✭
# 8
Visitor ✭ ✭ ✭

 

Hey Boris,

 

Six months is a long time to wait for the self referrals to fix for returning customers.

 

Do you know what the negative impact would be of reducing this down to close to zero to solve the problem?

 

Would we lose our Google Optimize data? Any other negative effects?