"Facebook should test its ideas better before full-scale rollout."

Most of the time they do, and I imagine they did with this, as well. Did I miss any evidence in the article that suggested that they didn't? Just because they are turning it off now doesn't mean at the beginning it wasn't successful.