This message was deleted.

Phases are meant to be used when you changed something in the implementation of the experiment and want to exclude people who were included in the initial version

Thanks for the explanation! However, In my case the targeting didn’t work as you describe it. This means I filtered out a value for a dimension and still got participants in the experiment, who had that forbidden value in the dimension. I’m not sure whether I can’t easily share this example with you, since the data used is internal.

hi Peter - you can DM me and perhaps we can meet to see if I can help solve the issue

Ahoi @fresh-football-47124. I just created a bug report here with some example dataset: https://github.com/growthbook/growthbook/issues/1383

thanks! we’ll take a look