There's a big difference in analytics used in MLB and the NBA versus the NFL. The difference is in the number of attempts to gain the data, a 2 pt shot vs a 3 pt shot, the trade off works because of the sheer number of shots taken and higher total points scored. Miss a single 3pt shot and it doesn't change the complexion of the game or how the coach strategize after that miss, unless it's the last 2 minutes of game. Because of the relatively low number of real scoring opportunities in the NFL versus NBA/MLB, if you fail the consequences are much harsher. The NFL also has so many fewer games, much less margin for error if you lose a game. If this team truly embraced analytics, it would be built around defense first and on offense it would be built around the passing game, the rule changes over the last several years benefit passing game vs rushing game.
Tomlin got his team into trouble at times he didn't have to in the past using/misusing analytics. He was aggressive in going for 2 pt conversions, because the data showed it should produce better results. It works until it doesn't and changes the way he had to coach to make up for those failed attempts. Analytics is useful in building rosters, but using them during a game requires a feel for the game and moment. It also requires the ability to prepare and adapt for if/when it fails.