How appropriate do you think it is to run a split test for a charitable campaign where the bulk of the donations will happen quickly and the opportunity for longer term optimization is low? Did the fund approach you because they were interested in a test or did you suggest it? How quickly did they adopt the winning candidate as their only version?

The article notes "this one’s a bit of a no brainer" and four out of five random people on the Internet predicted the optimized version. Given these facts, it's hard not to interpret the results as having cost a substantial amount of donation money by testing something the designer involved already knew the answer to.

Perhaps you could elaborate on the benefits to Hatians or the non-profit of running this test?