Of course the stark reality is that A/B testing on users (typically
readers, not editors) during the annual Wikimedia Foundation fundraiser has been a major component of the Wikimedia Foundation's growth.
In part that's a myth. The income has been increased simply by making the banners larger, brighter, naughtier and alarming (we're in danger, bla bla). Sometimes they take more space than is left to the article; sometimes they can't be dismissed.
Hi Nemo - I can't agree with this at all. The banners from the 2013 campaign (the last I can readily find) are no bigger or scarier than those from 2011. On the whole they are much less interruptive, as they are displayed less consistently; and they attract far less third-party attention than the "Jimmy banners".
The increase in efficiency through the banner campaign has been truly remarkable!
If there was a way to get the same kind of result on (say) the number of new editors who stick around and contribute more that would be great.
Chris