Interesting. For me, the biggest learning from this new study is that writers and journalists get it wrong, and don't realize it. Unless your overlords are doing their own A/B testing to see what readers actually prefer, they're quite possibly giving you bad orders. I wonder if the "headline analyzing tools" are set for SEO optimization. If that's the case, then the game is quite different, as the goal would be to have the articles rank well in search — that's a totally different strategy than getting people to click on a headline within a publication. I should've noted that in my article (and will add something now, so thank you!).