On Mon, Jul 20, 2009 at 6:04 PM, Robert Rohde <rarohde(a)gmail.com> wrote: It's also necessary to control for seasonal traffic (and thus donation) variations. I note that the first three button tests are at the end of summer while the fourth coincides with the beginning of the school year. It could be the case that there is no variation, or that the variation is highly significant. Since nobody has looked there is no way to tell if the test results are valid.

On Tue, Jul 21, 2009 at 11:02 AM, Andrew Gray <andrew.gray(a)dunelm.org.uk>wrote;wrote: Rotating them would seem like a more viable solution than randomised - We don't want the situation where every new page in WP someone reads there is a new/different coloured donation button where last week there was none at all - to go from nothing to that would be almost as bad as a flashing "donate here, now!" banner. I suspect that if there are any complaints about this new button system they will be about whether it makes us look like we're giving the 'hard sell' to people. Finding the middle ground between making the donation button clear/prominent and making it subtle/non-distracting will be difficult but also very important. We want people to know that we depend on donations but we don't want them to be annoyed by being constantly distracted by a big colourful push for their wallet (especially if it keeps changing). It's a tricky line to tread - but I'm sure we can do it! -Liam [[witty lama]] wittylama.com/blog Peace, love & metadata Sent from Sydney, Nsw, Australia

On Mon, Jul 20, 2009 at 9:19 PM, Erik Moeller<erik(a)wikimedia.org> wrote: For this the normal procedure is to give users a session cookie of some kind (either one handed out by the server or one just generated on the client) and base the selection on that. For caching reasons I suppose you'd just want to do this all client side. Should work fine. Alternatively, someone rigs up the front end caches to do this substitution based on IP at serving time. This would be non-trivial with squid. It would be much easier with varnish, alas. In any case, I strongly agree with the argument against running them sequentially. Not only do you get the uncertainty from changing habits over time but later buttons will suffer from the influence of prior ones. Whatever can be done to avoid sequential testing should be done.

What about changing it every hour? Then you have sufficient randomness over time, and no flashy buttons. -- eia 2009/7/21 Erik Moeller <erik(a)wikimedia.org>

On Mon, Jul 20, 2009 at 11:28 PM, Gregory Maxwell<gmaxwell(a)gmail.com> wrote: Yeah, I'd suggest it is better to show each person one button type and stick with it. One can approximate that behavior either by assigning each user randomly and tracking their type via a cookie or by selecting based on higher order bytes in the IP address and assuming those are fairly stable (e.g. using BBB mod 4 where the address is AAA.BBB.CCC.DDD). Neither approach is foolproof or totally without bias, but one can do fairly well. -Robert Rohde

on 7/21/09 10:33 AM, Birgitte SB at birgitte_sb(a)yahoo.com wrote: I agree with you very much, Birgitte. Both of these messages sound like threats. Marc Riddell