Hi everyone.
I have question about running experiments with growthbook: how to create depend experiments?
My case:
• I have experiment X: 50% variation-1, 50% variation-2
• I want to create experiment Y with following constrains: “only for users in experiment-X variation-1”
How to properly implement this scenario using growthbook?
r
rhythmic-agent-34208
01/23/2024, 10:23 AM
Hi Bogdan, we don't currently have functionality for sub-experiments or dependent experiments.
You could potentially use Namespaces to create overlap between experiments; otherwise, any dependencies would need to be written into the application code.
👀 1
l
little-lighter-38371
01/24/2024, 9:45 AM
Thank you August.
Seems like namespaces isn’t suitable for my case.
For example: if user-1 entered experiment-X variation-2 i want to add this user to experiment-Y to maximize speed of experiment (and time to market).
-------
What i can do with namespaces:
• create namespace namespace-1
• create experiment-X in namespace-1 (0 — 0.5)
• create experiment-Y in namespace-1 (0.5 — 1)
• got a results: users in experiment-X can’t get experiment-Y and I lose time & money
r
rhythmic-agent-34208
01/25/2024, 5:21 PM
Hi again Bogdan, I think you're right that Namespaces won't work in this case. I was wrong about that.
if you want to specifically include in Experiment B users who saw a certain variation in Experiment A, you would need to track which variation a user saw and then use that information as a targeting condition for the second experiment.
This could involve:
1. Storing the variation assignment for the user when they are exposed to the first experiment.
2. Using this stored information to target the user for inclusion in the second experiment.