Rich-text Reply

PHP multivariate testing

fooyourhomesuit 10-07-16

PHP multivariate testing



How do I set up a multivariate experiment with 2 variables? Is it easy to add another variable in the near future? Is this feature supported in the php library?



Reneevl 10-11-16

Re: PHP multivariate testing

Hi there,


When you want set up a Multivariate Experiment you can use Optimizely Classic to set this up.

However multivariate testing is not available to all plan types. If you don’t see this feature in your Optimizely account, or want to learn more about what’s available, please refer to our pricing matrix. Optimizely X does not support this to set up.

Also the SDK you seem to be reffering to is not one officially supported by Optimizely so I am not able to recommend a fitting solution in this specific case.

Perhaps you could aks the person who created the SDK?


I hope this helps!



Renée van Leijen
Technical Support Engineer
robertchan 10-11-16

Re: PHP multivariate testing

You might want to reference this Optimizely article on creating multivariate experiments:

For a basic set up, the break down would be creating an experiment, going to a particular variation to set the multiple elements you want to keep track of, and then adding another page. The new page added should contain the same elements that were being tracked on what you initially created. You can update and add another variable in the near future, but you'll want to consider whether your experiment's already running or not, if it is, then users who were previously bucketed into a particular variant will continue to be bucketed as such.

Looking at the repo you sent, it's definitely possible to implement this. You would simply use the "Variations" section to target the appropriate variants in your experiment then loop through adding tracking on the elements you're looking to track.
Robert Chan

Experimentation Hero
juliofarfan 10-12-16

Re: PHP multivariate testing

Also, a good practice regarding MTV test (best when you havent made any yet) is to think about each "Section" as one test with its own variations.
I.E MTV test that includes changing the layout out of a button, the header style of the page and adding images to the promotion.
Section 1: Button changes. var1: Original Var 2: Layout Proposal
Section 2: Header changes var1: Original Var2: New header style
Section 3 Promo images Var1: oroginal Var2: Promos with images.

Also when you get the results be sure to turn of sections to see if a mix of the changes in certain section is conclusive or not. I.E The mix between 3 sections is not conclusive but maybe its conclusive the mix between sections 1,3.