AdWords
5.5K members online now
5.5K members online now
Use AdWords conversion tracking and reporting to measure your results. Have a question about Google Analytics? Post it here, on the Google Analytics Community!
Guide Me
star_border
Reply

Split testing landing pages with GA's Experiments bug

Follower ✭ ☆ ☆
# 1
Follower ✭ ☆ ☆

Hi,

 

I've done some AB Testing with GA's Experiments on Adwords campaign landings. But it seems to me that there's some technological bug or something that lowers conversions numbers on Google Adwords for the time that the Experiment is ON.

 

Do you have any data about it?

1 Expert replyverified_user

Re: Split testing landing pages with GA's Experiments bug

Top Contributor
# 2
Top Contributor

What is the AdWords conversion code tracking? If you are measuring a conversion on the landing page itself - is the code on the experimental page as well?

Re: Split testing landing pages with GA's Experiments bug

Follower ✭ ☆ ☆
# 3
Follower ✭ ☆ ☆
The conversion code is in the Thank you page. Both the original and the experiment page goes to the same Thank you page

Re: Split testing landing pages with GA's Experiments bug

Top Contributor
# 4
Top Contributor

Sorry - I'm not too clear on what's going on here....

 

Let me check what's happening.... you are running the experiment within Analytics - so it's a Google content experiment, right?

 

The content experiment requires an Analytics goal as a measure of success - not an AdWords conversion.

 

So there should be a number of goal completions in Analytics.... and you are suggesting that parallel to this the conversions in AdWords has dropped... does the AdWords conversion use AdWords conversion script on the page or is it an imported goal from Analytics?

 

If these numbers don't match there are a couple of possible explanations...

 

- the Analytics goals imported into AdWords have a couple of days delay - so if the experiment has only run for a few days there will be a discrepancy...

 

- if you set up the Content Experiment in Analytics then it will track all traffic, not just the AdWords traffic - so any goal completions from organic, direct or referral traffic will be included - but will not record as conversions in AdWords...

 

Sorry this is a bit rambling - I'm kind of thinking aloud here!

Re: Split testing landing pages with GA's Experiments bug

Follower ✭ ☆ ☆
# 5
Follower ✭ ☆ ☆
Thanks Stick, great explanation.

This is a landing page for adwords campaign so it can't mix with other kind of traffic.

Adword's conversion are tracked with Adword's tracking code and Analytic's conversion with Analityc's tracking code.

But I'm not meaning there's a difference between both but that Adword's reported conversions goes down when I start the experiment and then goes up when experiment stops.

This is the path:

Ad -> Landing -> Thank you (conversion)

And when experiment enabled...

Ad -> Landing 1 ->
Landing 2 ->

Maybe when it moves from ad to Landing 1 and then it redirects it to Landing 2 ('cause of the experiment) something happens...

I don't really know.

Re: Split testing landing pages with GA's Experiments bug

Top Contributor
# 6
Top Contributor

Strange - and as you say it makes no sense...

 

Are you gettign the same number of people getting to the landing page  - is it just that they are not reporting or is the traffic noticeably underperforming when the experiment is running? That would be an even bigger worry than what appears to be a tracking issue....

 

Don't know how easy it would be to duplicate the paths and make it accessible only to Adwords traffic - that might help isolate the problem...

Re: Split testing landing pages with GA's Experiments bug

Follower ✭ ☆ ☆
# 7
Follower ✭ ☆ ☆
I just launched a new experiment. I think there are 2 solutions. To do the experiment right from the ads (directing traffic to landing A and B) or try with another A/B software.