My introduction to data and analytics came courtesy of the brilliant, whimsically skeptical folks at Baseball Prospectus.

While broadcasters rhapsodized about the virtue of sacrifice bunts and hailed players with inflated RBI totals, BP chipped away at such conventional thinking. Sacrifice bunts, the site’s writers argued, gave away a team’s most precious commodity — one of its three outs — while runs batted in were less a function of a hitter’s particular skill set than of his great fortune in coming to the plate when there were runners on base.

BP didn’t make its case like traditional late-1990s sportswriters, though. If the words “heart” and “intangibles” were part of its staff’s collective vocabulary, you never would have known. Rather, it backed its conclusions with cold, hard data.

Take the suggestion that batters shouldn’t bunt. In 2013, per BP’s database, teams with a runner on first base and no outs could expect to score 0.826 runs over the course of that inning, while teams with a runner on second base and one out could expect to score 0.637 runs.

What made BP’s analysis so compelling is that the site wasn’t dogmatic about it. “Bunts are self-limiting” may have been the headline, but it came with the caveat that occasionally dropping one down would keep defenders honest. BP didn’t marry itself to its convictions.

Similarly, it didn’t view analytics as the sole lens through which every question should be filtered. The site’s detractors often claimed that it wanted to rid the game of the oft-romanticized “human element,” notably the scouts who’d long lorded over the player-development process. To that, BP writer Dayn Perry responded with my all-time favorite metaphor: To paraphrase, when somebody asks if you want beer (statistics) or tacos (scouts), the proper answer is “both.”

The September print edition of MM+M, its fifth annual Data Issue, shows that medical marketers have finally gotten the memo. They’re deploying data in more useful and creative ways than ever before, whether to better inform the industry’s underperforming chatbots or, in tandem with AI tactics, to eliminate some of the guesswork from the drug discovery process. They’re poised to take advantage of data streams from untapped channels, such as ad-supported streaming, and to rid the industry of potential abuses, such as inferring pregnancy from purchasing patterns. The state of the data union may be splintered, but it’s strong.

Medical marketing has beer, in the form of its army of data scientists and technologists. It has tacos, in the form of its empathetic and savvy creatives. In the months and years ahead, this should make for one hell of a satisfying meal.