This is the last in a four-part series on the challenges of working in baseball. Part 1 looked at where front office workers come from. Part 2 looked at the cost of internships. Part 3 considered the implications of the Ivy League's takeover of front offices.
This is the final installment in a series that has examined the major-league front office. If big-league players are actually just guys in funny pajamas, then the front office workers are just (mostly) guys in funny polo shirts. But who are the people wearing those polo shirts (and how can I become one)? We’ve spent the past few articles looking at that question—in BP style—using a data-driven approach.
To this point, we’ve seen that despite the trend toward bigger front offices in baseball, the actual odds of getting a job in one are slim, and the process is very reliant on the internship system. While the internship system is theoretically just a way for young talented individuals to prove their mettle before a team commits to them, it also often requires paying out a few thousand dollars in expenses to live in another city for a few months. We’ve also seen that front offices tend to be sadly lacking in diversity on a number of fronts, and that leads to the question of whether teams are cultivating enough diversity of thought in their decision-making processes.
The final question worth asking in this line of research is what to do about all of that.
A league that celebrates Jackie Robinson with his own day every year still hurts for diversity in the positions of true organizational power. While the sort of horrific institutional racism that kept deserving stars out of the league for a large portion of its existence isn’t around today (and good riddance!), it’s easy for a set of seemingly reasonable decisions to end up producing a system that is still flawed. Baseball is no exception to the invisible forces that shape the world.
Let’s return to the issue of internships. As we outlined in Part 1, more than 70 percent of front office employees who graduated from college in the past 10 years started their baseball careers as interns. Again, internships make sense from a surface level. Teams can audition new talent (and get free or cheap labor) with a minimal commitment. Baseball teams have the luxury of having a lot of people who want to work for them, even for free. In fact, by the numbers, the pyramid of getting to a high-level front office position is even steeper than the one to become a big leaguer—there are, after all, only ~30 major-league GMs, and at least ~750 major leaguers. The proliferation and popularity of fantasy baseball has produced an entire industry of people (and an entire section of this website) dedicated to pretending to be a major-league GM.  With this glut of willing and eager potential labor, baseball teams—who at their core are just businesses—do what any sane business would do: They lower their wages and increase competition. It’s something of an open secret that baseball teams pay relative peanuts to their employees, particularly compared to what employees with those same skill sets would command elsewhere on the open market. However, if people are willing to endure lower—or no—wages for the thrill of working in baseball, then that’s how a free market works. See? Perfectly logical decision!
Still, we’ve seen that it’s only a select group of people who can work for free, and that group is shaped by things that we might expect and some we might not. For example, geography plays a role. If you want to become an accountant, there are businesses that need accountants in every city and there are probably dozens of them. But there are only 26 cities with major-league teams, and most of them have only one team in them. In order to take an internship, someone needs to be able to move to that city for what could be a very short period of time. Even if it is short, that’s still too long for someone with children, or even a person married to someone with a job that doesn’t allow for relocating. These considerations don’t even bring in how much it costs to move across the country just to try to follow a dream.
The internship market is also shaped by U.S. labor law. U.S. law states that most employees must make minimum wage. But if a position meets certain qualifications—specifically that it is a “training opportunity”—exceptions can be made. Most companies use this exemption with varying degrees of truthfulness. While there are internships that are genuine training opportunities (and this includes inside baseball, too), some are just pure cost-cutting job mills.
On top of that, there’s the Affordable Care Act. An internship is not likely to come with health insurance, but according to the ACA, a parent can keep a child on his or her health insurance plan up until the end of the child's age-25 season. If you’re 26, though, you have to provide that for yourself, no matter if you’re only making $1,160 a month for ten months. From the beginning, the structure of an internship tilts the pool toward college-aged or recently post-graduate applicants who are from well-to-do families that can support them.
There are going to be some readers of BP who have dreamed of working in a front office who might be surprised to know that teams even have internships. A lot of teams don’t go out of their way to publicize their internships or low-level job openings. If they did, they’d likely get hundreds of applications for one spot, in addition to the resumes that float in on their own every day. This is where the beast that is “connections” comes in. That’s where a “good word” from someone else comes in handy. In a vast pool of random strangers, having one of your friends say one particular candidate among the thousand resumes on your desk is “a good worker” could tip the scales. These things have to be filtered somehow.
So, how are teams to sort among the myriad applicants? One way is to use academic prestige. Out of our American League–sized sample, we previously determined that 16 percent of front office staff went to an Ivy League school. It’s hard to believe that those eight schools are the only ones capable of producing smart people. One other possibility is that the person in charge of hiring the interns might only put the word out to her/his own alma mater or a small selection of schools. We don’t have data on what schools interns themselves attended, but among the 159 full-time front office workers for whom we were able to find undergraduate data, there were 12 “paired” cases in a front office, where two people just happened to have gone to the same place for undergrad. That’s not an epidemic, but it is suggestive of a trend.
Suddenly, not only do you have to be young, relatively mobile, and have access to some financial backing, you might need the right connections or to go to the “right” college to get that internship. In a society where access to “elite” colleges and wealth are sadly correlated with race and class, it’s not surprising when a system that inadvertently selects its candidates on these criteria ends up looking the way that it does.
It comes down to this: If baseball is going to get serious about diversity, it’s going to have to re-examine the internship system.
The frustrating thing is that there’s a good argument to be made that it doesn’t have to be this way. The base assumption powering a lot of these practices is the idea that a team should prioritize keeping costs to the absolute minimum. That assumption, though, might not actually be correct for baseball. Research from 538’s Ben Lindbergh and Rob Arthur has shown that early adopters of sabermetrics within the front office collected several wins worth of value from being willing to embrace a new idea before their competition caught up to them. Work by former BP writer (and current Phillies analyst) Lewie Pollis suggested that the differences between front offices might be worth millions to a team. Analysts are cheap (even adding in the cost of data feeds and computing power) compared to the million-dollar players whom they help to pick out. We’re talking about a bunch of business and finance majors. Shouldn’t we be looking at these things in terms of return on investment, rather than pure expenditure?
It seems strange that within a system where information and ideas are relatively cheap and powerful, teams would begin with an assumption that they would simply find the lowest-priced ones, rather than the best ones. What if someone with great ideas can’t afford to work for free?
How, then, can we tell what teams are looking for? By analyzing their “help wanted” ads. Earlier this year, Sean Dolinar of FanGraphs looked at all of the job postings that had circulated from teams, as well as a few other organizations, such as Trackman and Baseball Info Solutions. The kinds of positions included in his data set were largely analytical and R&D type jobs, but Jonah created a word cloud showing the concepts that came up most often. At first, it might not be surprising that the qualifications section of the cloud included words like SQL, statistical, data, statistics, mathematics, and server. These are, after all, jobs largely in analytical departments.
What’s more telling is what’s not in that cloud. Technical skills make up nearly the entirety of what teams specifically ask for in job postings, with little-to-no mention of critical thinking, much less an active seeking for someone willing to question the underlying assumptions about the way things are currently done. Teams do often include additional questions for the people who make it past the initial screen, some of which assess for problem solving and creativity, but this doesn’t resolve the fact that they’re still drawing from a pool pre-limited based on (teachable) technical skills.
To draw a baseball analogy: When scouting for position players, teams were once guilty of looking only for players who had obvious hitting skills and ignoring the ones who had particularly good defensive skills (unless they also had a big bat). We now understand that both sets of skills can provide value to a team, but if you only look for excellence on the hitting side and treat fielding as a nice bonus, you miss the guys who are legitimate 80-grade fielders in premium positions who would be able to hit enough to make it work. Once we understood how much value a good fielder could bring to a team (thanks to those same analytics departments!) teams changed the way they approached scouting. Interestingly, though, when these analytics departments go looking for their own talent, it doesn’t appear that they’ve yet applied the lessons they learned on the field. Technical skills are most certainly valuable, but are they the skills to look for?
The evidence actually suggests that technical skill can really only take you so far. We don’t know what’s happening behind the curtain among team analytics departments, but we do have a few lines of research in public that have been going on for many years. The oldest is probably the pre-season projection system. It answers a very important question (what can we expect out of Player X next year) and there’s no shortage of very smart people who have tried their hand at it, including the people who are responsible for our own PECOTA system. The models, some of which are open-sourced, some which use proprietary formulae, are without a doubt very complicated things filled with all sorts of code. And when you test them, none really consistently excels over the rest of the field, and none of them really does much better than the no-frills Marcel system. Then there’s research on what might broadly be called ERA estimators. Since the publication of Voros McCracken’s initial DIPS paper, suggesting that for pitchers we should largely excuse their performance on what happens on balls in play and focus on the three “true” outcomes, there’s been a desire to create a measure of how good a pitcher “really” is. Those measures have undergone multiple iterations with increasing statistical sophistication: from simple component models (FIP), to models beginning to incorporate additional factors (xFIP), and then eventually interactive components (SIERA), and finally the recently released cFIP, which uses a mixed model approach. We can run retrospective tests on cFIP and see that it is demonstrably better than its competitors. But it’s worth noting that even then, it’s only incrementally better than what came before it.
It’s not that increasing complexity in handling data isn’t valuable, but how much? What happens when big data tells you mostly the same things that little data did? For example, imagine that a team had discovered Mr. McCracken’s idea of DIPS before the public knew about it and created a rudimentary component-based ERA metric for itself. That one insight would have been a boon to their player evaluation, and probably more valuable than the amount of additional value they would have gotten if someone had handed them cFIP the next day. Again, this isn’t a case of either/or. If you can encourage great insights and world-class statistical analysis, then great. But teams seem much more invested in those technical skills.
It does seem likely that much of what we have discussed here has probably come up in the board and war rooms across America’s ballparks, and as the practical sort, we acknowledge that change takes time. There are a lot of moving parts involved in creating a front office, and none of them can be tossed aside—and we are loathe to suggest that they be. However, with the proliferation of “sports analytics” majors, the growing national student debt load, and many other factors, neither is this the time to sit back and let this sleeping dog lie.
We are, of course, guilty of gross over-simplification here. There are 30 front offices, each with its own unique story and mix of characters. But in the aggregate, there are some patterns that are emerging in the data. There is danger of stagnation in the front office writ general, or at least danger of front offices not being as fruitful as they could be. Yes, it’s great that the analytical movement has made inroads into the front office, but to survive, “analytics” needs to be fed good ideas. It’s not that front offices are places bereft of ideas, but are they creating a garden where great ideas will sprout all over the place?
There are good ideas—ideas that analytical types might have scoffed at a few years ago—making their way into the game. Teams are paying much more attention to nutrition and diet, even among their minor leaguers. Teams are hiring coordinators who simply help players deal with the pressures of being a pro baseball player. The neat thing about there being 30 teams is that each controls its own hiring practices and can even write its own book about The Way in which they conceptualize themselves. If a team wants to change things up, they can!
However, the data suggest that they might need a prompt to take a fresh look. We started off with a simple question of what the odds were that someone might eventually end up in a front office job. It turned out that the answer to that question seemed to depend at least in part on factors that weren’t really directly related to baseball decision-making. Maybe the new operational inefficiency is to take a look at the people in the front office itself and to re-think how they get there.
 Often at great personal cost, as we discussed in Part Two.
 So many people. So very many people.
 This is not a condemnation of fantasy baseball. Many people who want to work in baseball don’t play fantasy baseball! Some do, and some wouldn’t think they were qualified if they didn’t.
 (and cracker jack!)
 28, if we’re counting Oakland and Anaheim as separate metropolitan areas.
 Don’t ask me how I know. -K
Enough being a varying term, of course. What is enough one year, ain’t enough the next.
 Now is the time to note that these great jumps in analytical development may not continue with the increased privatization of data.
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.Subscribe now