Allocating 100% of traffic to one version but still seeing the other version
Maybe it's a caching thing, but I set one version to get 100% of traffic, and on some browsers (days later) I could still see the version that should have been getting 0% traffic. Any ideas??
In particular, this section might be useful for you:
"Changes made to the overall experiment traffic allocation only affect new visitors. Existing visitors (whether or not they were bucketed in a variation) will keep seeing the same variation, even if change the traffic allocation. This also means that visitors excluded from the test will always be excluded, even if you change the overall traffic allocation to 100%."
When you say you "set it to 100%", was it ever active and *not* at 100%?
If so, then any user who saw the experiment before you changed it to 100% will continue to see the "now inactive" version of the experiment.
To truely set an experiment as the default variation for ALL users, you should
1- clone the experiment,
2- remove all obsolete variations,
3- deactivate the original experiment,
4- activate the new clone
Analytics and Testing Guru