So now they back tracked on their original decision which was based on opinion and instead use the one that actually has good data behind it and wound up making them millions upon millions more. Sometimes they'll restrain colors to stay within a theme knowng that the best they'll get is a local maxima, but when you've got the number of users that Google does, it makes sense to try 41 shades of blue, because you'll sometimes find that just a handful significantly outperform the others, and the probability of you having picked them as a designer with no testing or data is pretty small.
Designers hate how Google does things, but the results speak for themself... The methods used work really well.
Should you cater to the lowest common denominator and be the Microsoft of the web, or try to do the Best Thing and have a smaller market share but of much higher taste and disposable income, aka the Apple ?
And do you really think there's a socio-economic divide along the lines of blue vs. yellow text ads? Because that's pretty ridiculous.
Who loses here, besides the person who thinks yellow is "bad design?" Your customers have the final say, not the designer.
That surely is an overkill. Why do they have to decide between 41 shade of just blue? Instead they could have had a hackday kind of event and asked the designers to come up with his/her own designs based on a basic layout and then combine the best features (and colors maybe) of each design.