Confirmed users, Bureaucrats and Sysops emeriti
1,737
edits
Test Pilot/2017Q1: Difference between revisions
m
no edit summary
mNo edit summary |
mNo edit summary |
||
| Line 55: | Line 55: | ||
| KR1 | | KR1 | ||
| At the end of Q1 double the percentage of new sessions enabling an experiment on intial Test Pilot install | | At the end of Q1 double the percentage of new sessions enabling an experiment on intial Test Pilot install | ||
| | | <span style="color:orange;font-weight:bold;">0.61</span> | ||
| Q4 2016 was 6.6% . [https://datastudio.google.com/#/org//reporting/0B6voOaUZL-jwTTlmcVZ1R3RCb2c/page/FZH source]; page 3, "% of New Sessions Enabling an Exp" | | Q4 2016 was 6.6% . [https://datastudio.google.com/#/org//reporting/0B6voOaUZL-jwTTlmcVZ1R3RCb2c/page/FZH source]; page 3, "% of New Sessions Enabling an Exp".<br /><br />There is a [https://datastudio.google.com/#/org/Ra9O2QIYT8We-1pOmnS_6w/reporting/0B6voOaUZL-jwR2VwSUNaQmJCZjQ/page/COH new dashboard] which puts this number at 10.9%, with some qualifiers: If you look at the performance by traffic medium, you'll see 65% of our traffic was from snippets, but snippets have the lowest percentage of sessions enabling experiments (7.4%). That means any time we run a snippet campaign we tank this number, so basing it off of a percentage of new visitors was a bad idea and not a great OKR. The proper way to do this would have been an A/B test.<br /><br />.61 credit for the increase in the number, just realize that it's heavily weighted by other forces. | ||
|- | |- | ||
| KR2 | | KR2 | ||