Optimize
2.6K members online now
2.6K members online now
Learn how to create your first experiment and get tips on creating A/B, redirect, and multivariate tests.
Guide Me
star_border
Reply

Analytics Content Experiments vs Google Optimize

cdb
Visitor ✭ ✭ ✭
# 1
cdb
Visitor ✭ ✭ ✭

We have been using Analytics Content Experiments until now. We want to migrate some of our experiments to Google Optimize. We used to detect which changes to make to the UI using the Content Experiments API, namely we used to call cxAPI.chooseVariation() . This would return a number and we modified our UI accordingly.

 

The only way we can see to do this in Google Optimize is by using the "Run Javascript" feature and adding a window variable which we can then detect in our code. Is there a way we can call Optimize to tell us what variation we should display?. Setting window variables is not a clean way to do it. It does not play along very nicely with React components

1 Expert replyverified_user
1 ACCEPTED SOLUTION

Accepted Solutions
Marked as Best Answer.
Solution
Accepted by Jason O (Community Manager)
May 2017

Analytics Content Experiments vs Google Optimize

Google Employee
# 6
Google Employee

Correct, that is a limitation today that we are working on improving.

View solution in original post

Analytics Content Experiments vs Google Optimize

Visitor ✭ ✭ ✭
# 2
Visitor ✭ ✭ ✭

Just wanted to second this- we are also using the cxApi chooseVariation method to run javascript AB tests in our single page application (documented with more detail in our blog).

 

I also explored replacing this functionality with Google Optimize, but it is not very clean, as cdb stated, and also doesn't let us explicitly set variables for the control group. In our case it is useful for us to detect if a user was explicitly assigned to the control group, or is just there because they are adblocking the google analytics script, and for our own internal event tracking it is useful for us to be able to make that distinction.  

 

If someone has found a good workaround to replace the cxApi chooseVariation method with optimize, or a programatic way to access the Google Optimize test variations in JavaScript please let me know!

Analytics Content Experiments vs Google Optimize

Visitor ✭ ✭ ✭
# 3
Visitor ✭ ✭ ✭

I think currently there is unfortunately no solution for server side experiments (as it was previously called in GA experiments) in Optimize. 

There is a page Https://support.google.com/360suite/optimize/answer/6315892?hl=en that explains that there is currently no option with Optimize to run server-side experiments. 

 

But I'm also exactly looking for this, I have a use case where I'm using the measurement protocol for all the tracking and therefore there is no client-side tracking/js possible. Still I am able to do A/B tests with GA and content experiments right now but I would love to see them in optimize. 

Analytics Content Experiments vs Google Optimize

Google Employee
# 4
Google Employee

Stay tuned for more documentation around this capability.

 

However, I will briefly mention here that it is possible today to use the measurement protocol with Optimize. You would create a new experiment in Optimize with the right number of variants, pick your experiment objectives and probably a targeting condition that always evaluates to false. You can then start the experiment and note the experiment ID in the Optimize experiments detail page. You can then set the value of that experiment ID as the &xid parameter and the chosen variant in the &xvar parameter on your server side tracking beacons. Those measurement protocol parameters are mentioned here: https://developers.google.com/analytics/devguides/collection/protocol/v1/parameters#experiments

Analytics Content Experiments vs Google Optimize

Visitor ✭ ✭ ✭
# 5
Visitor ✭ ✭ ✭

Thanks Jim, this is a really useful tip!

The experiments in optimize have to be started to get the experiment id, though. 

 

It would be great to have it in the future as we have it in GA so that we can prepare everything in the server side code, release it and then click start in the experiment fragt in GA and then the experiment starts. 

Marked as Best Answer.
Solution
Accepted by Jason O (Community Manager)
May 2017

Analytics Content Experiments vs Google Optimize

Google Employee
# 6
Google Employee

Correct, that is a limitation today that we are working on improving.

Analytics Content Experiments vs Google Optimize

Visitor ✭ ✭ ✭
# 7
Visitor ✭ ✭ ✭

Hi Jim, 

 

Our organization is currently evaluating A/B testing solutions, and Optimize has a lot of upside but the lack of an API for managing server-side experiments is a bit of a dealbreaker. If there's something on the horizon, we could keep our options open until then or potentially even invest in Content Experiments as a stopgap, but if not we'd need to go another direction.

Is there a roadmap/timeline that Google can provide, any information at all for their existing customers to know what to expect in this regard?

 

Thanks!

Analytics Content Experiments vs Google Optimize

Visitor ✭ ✭ ✭
# 8
Visitor ✭ ✭ ✭

Seconding this question! The management API currently only works with Content Experiments - will this be updated to work with Optimize before Content Experiments are deprecated?  Is there a timeline for when that deprecation will occur?