Google Content Experiments allow you to create experiments for A/B testing two pages, which then slowly converge on the "better" page depending how the goals are met.

However how is this convergence affected by filters? For example we have a filter to exclude internal traffic and two Views ("All Traffic" and "Excluding internal traffic").

As the experiment can only be generated on one view how does this work?

If I set up the experiment on the "Excluding internal traffic" view, and visit the page internally then I presume I will still call the experiment code, and potentially still be redirected to B page, but my page visits won't be available to view in GA using the normal GA code (for either A or B page).

So my question is, is this internal traffic view excluded from the experiment reports (and particularly the convergence calculation) for experiment success? Or does the fact that the experiment Javascript is called by my browser, mean it's included even though my traffic is not included in other GA stats under that view?

  • What metric are you using in you experiment? I think this could change the answer to your question, eg. bounce rate versus some other goal. Slightly off topic, when we have run experiments, I think it picks you your IP and always sends you to the same experiment page, i.e. keeps your experience consistent.
    – walbuc
    May 19, 2015 at 11:06
  • I was using a goal completion metric for a sales funnel. So serving up a number of different home pages with aim of seeing which page lead to best conversion. As it was the home page I was concerned that traffic from inside the company (where no sales will be made) would affect the experiment. I agree the same version is sent for each subsequent visit but that still doesn't answer the question. May 27, 2015 at 6:15


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.