Multipage Experiments

Understand Multipage Experiments

Setting up multipage experiments or funnel experiments are possible and fairly easy in Convert Experiments. We don't have a wizard for this but with the document below you will be able to set it up fast. 

  1. Start an A/B or Multivariate experiment as normal.
  2. In the editor make the changes on the first page as normal.
  3. Click the on the gear icon and select "Change Editor Page".


    .
  4. Enter the URL of the other page you want to include in your experiment and edit changes on:



  5. Make the changes on the second page of your multi-page experiment.

  6. Check your generated code by first clicking on the pencil you get when hovering over the variation name. Then click on the Variation Code menu item.



  7. Then you can see the code like below:



  8. Optional you can secure that the two changes we made (one per page in this example) will only run on a specific page by wrapping them in this code:

    
    if(window.location.href=="http://www.google.com") {
    //code here 
    }
    
    
  9. If needed, you can apply certain changes on pages that have certain custom Page Tags, in the example as follows:

    
    if(_conv_page_type=="product") {
    //add some JS code here that does changes on the product pages
    }
    if(_conv_page_type=="checkout") {
    //add some JS code here that does changes on the checkout pages
    }
    
    

     
  10. When you Continue to the Experience Summary make sure all the pages that you like to run this test on are included in the "Site Area"



  11. Done
Have more questions? Submit a request

Comments

  • Avatar
    Byron Reed

    Does this mean that a person who gets the variation B on the first page of the experiment will also get the variation B on the second page?

    Let's imagine that there is an email signup on 3 different pages. We want to run an A/B of two approaches to that email signup. We want to test (Page 1, Page 2, Page 3) vs Page 1 Variation, Page 2 Variation, Page 3 Variation) so that we can see the aggregated impace of the Variation. Does the setup outlined in this article (205160345) support this? Thanks!

  • Avatar
    Claudiu Rogoveanu

    hi Byron, yes that's exactly what it does.