Rich-text Reply

Integration With Adobe Analytics multiple experiments in 1 page

ppc 02-16-17

Integration With Adobe Analytics multiple experiments in 1 page

Hi,

 

I just integrate my optimizely account with my Adobe Analytics.

Everything is working well.

I set up my Prop to be a list and an eVar.

 

My question is, on some pages of my website I can have more than 1 experiment at the same time. But in my prop and eVar I only have 1 value. I should get at least in my list prop all the experiments that are running on the page.

Or do I need to have 1 prop/eVar per experiment?

 

Thanks

 

PPC

ppc
Level 2

robertchan 02-16-17
 

Re: Integration With Adobe Analytics multiple experiments in 1 page

You don't need 1 prop/eVar per experiment so you I'd suggest keeping the list. If you've not got a large set then perhaps you could work with the 1 to 1 you mentioned.
Robert Chan

Experimentation Hero
ppc 02-16-17
 

Re: Integration With Adobe Analytics multiple experiments in 1 page

Thanks Robert for your answer.

But even with the list prop when I have multiple experiments in 1 page I only have 1 experiment value on it.
Do I need to do something more than only setup more Prop as a list?
ppc
Level 2
robertchan 02-16-17
 

Re: Integration With Adobe Analytics multiple experiments in 1 page

Anytime, from my understanding, you could take the approach of working with a data element. Then, within each experiment, you could adjust the value of your variable according within your code editor.
Robert Chan

Experimentation Hero
robertchan 02-16-17
 

Re: Integration With Adobe Analytics multiple experiments in 1 page

This is if you're operating through DTM.
Robert Chan

Experimentation Hero
ppc 02-16-17
 

Re: Integration With Adobe Analytics multiple experiments in 1 page

ok thanks. I will check this
ppc
Level 2
JDahlinANF 02-16-17
 

Re: Integration With Adobe Analytics multiple experiments in 1 page

When the first experiment runs, Optimizely sets the vlaue of that expeirment/variation.

When the second experiment runs, Optimizely sets the vlaue of that expeirment/variation.

When the second experiment runs, Optimizely sets the vlaue of that expeirment/variation.

 

Now that all experiments have been processed, the value of that singular variable is the value of the last experimet that processed.

 

If you can allocate a couple props and a couple eVars, you should.  That way you can have multiple experiments running on any page.  You'll need a calendar to keep track of which variables are in use so you don't accidentally use the same one twice at the same time on the same pages.

 

An alternative, is you don't do the Adobe integration at all and instead read the optimizely object to determine which experiments are running on the page and which variaiton the user is in and set a list prop or list var with the appropriate values based on what you find.  If you choose to go this approach, keep in mind the character limits of the props and eVars.

ppc 02-16-17
 

Re: Integration With Adobe Analytics multiple experiments in 1 page

Thanks JdahlinANF
ppc
Level 2
ppc 02-21-17
 

Re: Integration With Adobe Analytics multiple experiments in 1 page

Hi JDahlinANF,

 

I did the implementation by reading the optimizely object. It works perfectly thanks.

 

However I'm wondering if you know where I can fiind in optimizely interface the id of the varition. For now I put in my prop list experiment id + variation friendly name. But because of the 100 characters limit of the list in some case my value is truncate. So I think a better option would be to put the variation id instead of the friendly name.

 

Thanks

 

Pierre

ppc
Level 2
JasonDahlin 02-21-17
 

Re: Integration With Adobe Analytics multiple experiments in 1 page

Hi @ppc,

If you are using Classic Optimizely, the variation the user sees is stored in "optimizely.variationIdsMap".

 

Use "activeExperiments" to get the list of expeirments the user is seeing on the page, and use variationIdsMap to see which variation of each experiment the user is seeing. (This contains the map for all experiments, not just the ones on this page.)

 

If you go with recording the Variation IDs, you'll need a document of some sort to either create your SAINT or classification rules.  Sicne those are only available in the reporting tools and not available when viewing the page directly (like, when you are validating the experiment), I've found it helpful to simply name the variaitons like "A_control" and "B_test".  This keeps the number of characters short and lets you know which bucket the user is in.

--Jason Dahlin
Analytics and Testing Guru Smiley Happy
Experimentation Hero