Hi, my team has been working with GB for a few weeks now and we've ran into the following issue running a namespace targeted experiment that uses a combination of feature flags and visual editor changes: We see that users in the variation group of experiment 1 are being shown the visual editor changes of experiment 2 (both are in the same namespace). We expect these to be mutually exclusive, and what's unusual is how this is reflected on the Growthbook devtools: 1. Image 1: Shown on initial load of page --

> Skip because of filters

on experiment 1 is logged and settings do not include

namespace

info, but do include a

filters

2. Image 2: Shown on initial load of page --

> In experiment variation 1

on experiment 2 is logged and settings do not include

namespace

info, but do include a

filters

(has same info as

namespace

). Note this is the experiment that should not be shown. 3. Image 3: Shown on refresh of page --

> Skip because of namespace

on experiment 2 is logged but the visual editor changes for this experiment are still being applied on the page. 4. Image 4: Shown on refresh of page -- >

> In experiment variation 1

on experiment 1 is logged and settings include

namespace

info. Any support here would be greatly appreciated!

Are these all from the same user?

These are all from the same user. The question is: why is namespace being ignored causing experiment 2 visual editor changes to be shown?

I think it might be the hash version

Hi Saori, checking in on this. Were you able to verify the different hash versions being used?