Come one guys, if you submit to Hacker News at least put a little effort into your posts.
Update: Don't mean to sound brash but I'm tired of folks telling me what I should/should not do without giving me some detailed CR datasets, including some demographic data (if possible).
But UX improvements about button calls to action? Come on. Anyone here can test this, in less than two minutes of work. You could have an interesting discussion about metrics selection, but arguments which collapse to "I have an aesthetic preference for this alternative" leave me a little unsatisfied.
It's also important to remember that just because it's a best practice does not mean it will apply for every design in every case, which is why testing should be done on your own with your own site if you feel the need.
We believe that picking ideas apart through and through with the use of experience, logic and intuition with brilliant design thinkers is the only way to arrive at something substantial and meaningful.
Simply accepting something because it has been tested or researched is not good enough, and does not engage the mind. Independently thinking about an idea with other independent thinkers is the only true way to arrive at something meaningful. While this is probably quite difficult, we are up to the challenge.
Amazon hides sitewide navigation during checkout. I don't go quite so far, but I've tested a few variations, including A: sitewide nav and B: one-off nav which includes only things that I feel are likely to increase conversion.
B crushes A in tests on my site, to the tune of "would cost me thousands of bucks a year to do A instead". Do I get a subsidy from the Right Honorable Guild Of Independent Thinkers for doing it the correct way? Because doing it the way that actually works is really tempting.
I could show you 10 thorough and detailed product pitches from seasoned and experienced people for Farmville. 8 out of those 10 pitches turn out to be utter failures in the field.
Hacker News has little patience for armchair theory. Why do you think patio11 is so highly regarded? It's because he shows us why our rational thinking about users is wrong straight from the trenches.
The article isn't saying if you do your search this way you will sell a billion products. Of course if that was the claim some statistical proof would be necessary. All the article is doing is offering a simpler and cleaner approach to search. That's all.