User talk:CKoerner (WMF)/Work/A-B Test Results 1

Question

There were two test groups. Did the two test groups see the same thing? If not, how were they different? Whatamidoing (WMF) (talk) 18:12, 2 February 2016 (UTC)

Thank you, I've added in a bit about how the test groups were different CKoerner (WMF) (talk) 23:30, 10 February 2016 (UTC)

Minor suggestions

• Personally I would say "accounts for" rather than "counts for". I don't know if it's actually more correct or not.
• When you say "1.5-2% of our traffic", I would suggest being more precise about which traffic you mean. All views of all wikimedia projects?
• Should "were likely to end in a clickthrough" really be "ended in a clickthrough"? In other words, is it really non-deterministic as the current phrasing suggests?
• At the end, it is probably worth noting that the change can't simply be deployed as-is. The test was run with some hasty code (you'll probably want to check with several people to get the right adjective there), so the code has to be brought to production quality before it can be rolled out. I don't want people to think it will happen tomorrow, or to be frustrated that it won't.

--KSmith (WMF) (talk) 21:38, 3 February 2016 (UTC)