Its simple. The tools are only as useful as those who wield them. Using data analytics to execute decisions is sports clearly has flaws.
This is a terrible and probably wrong take. Find me an analytics based argument that would suggest you should keep batting a terrible hitter in middle/high spots in the lineup.
I can't think of any off the top of my head, but I can easily make an old-school non-mathy explanation: the idea that Springer is important and needs to get going so the team decides to force feed him at-bats possibly in "protected" spots in the lineup until he rebounds. Similar case for Kiermaier except not high in the lineup because he has no history as a dangerous bat. We've seen these sorts of things happen and heard old-school guys like Buck argue for them in the past.
Not saying the latter is what's happening, but it at least makes more sense than blaming "analytics" for no really definable reason besides the fact that the team is known to use math in its decision making process and it becomes an easy target especially for traditionalist fans.