Rich-text Reply
Highlighted

Conflict from running multiple experiments at the same time

ScottWalter 01-07-15
Accepted Solution

Conflict from running multiple experiments at the same time

[ Edited ]

Can running multiple expierments on the same site cause the results to influence each other? That is, will the buckets be containmaniated by other variations which could influence some of the same metrics? Or does it allow only one expierment and variation per visit at any one given time? 

greg 01-07-15
 

Re: Conflict

There is no (practical) limit to how many experiments a user can be bucketed into during a visit. So yes, be mindful of which experiments you're running, what goals you are tracking, and how they can affect one another.

 

You can use a custom Javascript audience (if your plan supports that) to exclude visitors who are already bucketed into other experiments. See instructions from user @nap0leon : https://community.optimizely.com/t5/Developers/Custom-Audiences-include-exclude-other-experiments-am...

Level 2
Amanda 01-07-15
 

Re: Conflict

Great question. @greg is exactly right. You are bucketed into a variation independently for each experiment. So, a visitor could easily see the original for one experiment and a variation for another experiment. 

 

Here are some other discussions that highlight a similar topic. Both provide some great insight and link to some great support articles! 

 

https://community.optimizely.com/t5/Using-Optimizely/Simultaneous-Testing-Running-two-different-test...

 

https://community.optimizely.com/t5/Using-Optimizely/Overlapping-Experiments/td-p/1432

 

Let us know if you have questions! 

Optimizely
Hudson 01-08-15
 

Re: Conflict

Scott and Greg, I recommend you check out the answer posted by my colleague Ryan Lillis in the first link that Amanda shared; its a good synopsis. The key point is that unless two tests are affecting similar items, the randomization of bucketing visitors between experiments should mitigate interaction effects in most cases.
Optimizely