When deciding which metric conversions to include in an experiment analysis, the only things we use are user ids and timestamps. If the metric conversion happened after the user saw the experiment (within the conversion window), it's included.
future-teacher-7046
09/23/2022, 6:59 PM
You could create a new one-off metric for this experiment that is basically a copy of your global "Add to Cart" metric, but adds an additional WHERE clause to limit to specific routes
e
echoing-memory-34743
09/23/2022, 7:10 PM
How does that play with the attribution model?
It looks like the All Exposures model implies that the same experiment for the same user can result in multiple outcomes, if the experiment is triggered multiple times