The article is a little biased. Indeed, users who use social login spend more time on their website. It doesn't involve a causal link between time on website and social login: perhaps users are not the same, with different habits of navigation.

And users who use social login will perhaps spend more time if it's not provided.