The author's takeaways are pure speculation as to causation, but he presents them as if they are facts. I could claim that people just don't like the color of his photo, and I would have as much evidence to support this has he has for his other claims re: authorship placement. And he doesn't mention that since he is not doing a split test (since it's not possible) and he's not measuring confidence intervals that his "A/B test" may be meaningless.

His intuition makes sense, and his conclusions are probably correct in this specific case, but this is shoddy analysis and shouldn't be presented in a generalized manner as he's done here.