Obama's Moneyball Strategy

Brad Pitt in a room of Romney voters
Source: Columbia Pictures

Anybody who has seen “Moneyball” — or, even better, read the book — knows the story of how Billy Beane and the Oakland A’s outsmarted the baseball establishment by using new statistical techniques called “sabermetrics.” The movie portrays the confrontation between the grizzled scouts, whose time-honored player evaluation methods include rating the attractiveness of a prospect’s girlfriend (an indication of confidence…), and Brad Pitt’s Beane, who only cares about numbers. Not the traditional numbers, such as batting average and runs batted in, but the new sabermetric numbers with acronyms that don’t appear in the daily box score: WAR, OPS, BABIP, and of particular importance to Beane, OBP.

The idea behind sabermetrics is that the conventional baseball statistics are unrepresentative of a player’s true value. Critically, they often obscure or misassign causation. For instance, a batter gets credit for an RBI if a runner scores during his at-bat, even though the run was dependent on a prior hitter reaching base. And when the baserunner scores, he gets credit for a run that was contingent upon the batter’s hit.  At its most basic form, sabermetrics disaggregates the complex team sport of baseball into more accurate estimates of individual performance.

I have long argued that society can learn a lot from baseball, and sure enough, the Obama team has adopted a sabermetric model for the 2012 campaign. Slate got the scoop:

The Obama campaign’s “experiment-informed programs”—known as EIP in the lefty tactical circles where they’ve become the vogue in recent years—are designed to track the impact of campaign messages as voters process them in the real world, instead of relying solely on artificial environments like focus groups and surveys. The method combines the two most exciting developments in electioneering practice over the last decade: the use of randomized, controlled experiments able to isolate cause and effect in political activity and the microtargeting statistical models that can calculate the probability a voter will hold a particular view based on hundreds of variables.

The idea is to precisely evaluate the impact of individual political messages. The campaign starts by running test ads in various media markets. An intensive data-collection effort follows to gauge the effectiveness of the messages contained in the ads. This data is crunched and refined, so that the campaign can follow up with more targeted messaging based on the results. Through a number of iterations of this test-and-refine process, the campaign hopes to acquire empirical evidence for which messages work where and with whom. If all goes to plan, persuadable voters in key areas will receive tailored political communications during the heat of the fall campaign.

Campaigns have traditionally hired political consultants to tell them which messages work and which don’t. They often look like the paunchy, profane, tobacco-chewing scouts of “Moneyball.” Their credibility comes from experience, and, unsurprisingly, most of their analysis is based on a qualitative “gut” understanding of what went right or wrong in past campaigns. To the extent that they use numbers, they are the misrepresentative RBI-type statistics that appear in the top line of polls.

While two old school political vets, Jim Messina and David Axelrod, are managing the Obama campaign, they — like Beane and sabermetrics — have embraced the new models of electioneering. And just as Beane went outside the traditional coterie to deputize a young Yale economics grad, Obama has turned over the EIP programs to political scientists from the Analyst Institute.

For what it’s worth, the Romney folks aren’t exactly Luddites. Romney, a BCG and Bain man if there ever was, was noted for pioneering the use of data analysis in management consulting and private equity. He has brought that empirical approach to his campaign and will be using some of the same microtargeting techniques as Obama.

Nevertheless, the Democrats — like the Oakland A’s — will enjoy first mover advantage during this election. For all of the sophisticated trappings of 21st century baseball front offices, a team still wins by scoring more runs than its opponent. New types of electioneering haven’t altered the precept of successful campaigns, either. The goal is still to build a list of supporters longer than the opposition’s, and then to get those supporters get to the polls. Democrats have a sizable lead in both.

Every contact that recent Democratic campaigns have made with a voter has been recorded in Votebuilder, a powerful database built for and licensed exclusively to progressive candidates. The Obama campaign can instantly call up a voter’s candidate preference (on a 1-5 scale, from strong D to strong R) in each presidential and congressional election dating back, in some cases, to 2004. Information obtained from other sources — internet purchases, magazine subscriptions, club memberships, etc. — can be overlaid and cross-referenced to fit complex voter choice statistical models. The GOP has a similar database (Voter Vault,) but the McCain campaign and RNC were unable to match the Obama field effort in 2008 and consequently have little residual data to offer Mitt Romney. Obama may be tied with Romney in the polls, but he has a much better list of supporters.

Obama also has head start building the massive network of organizers and volunteers who will be tasked with mobilizing those voters on Election Day. Not only is much of the Obama infrastructure still intact from 2008, but the campaign has been able to ramp up staff and offices while Romney was fighting for his life during the primaries. And instead of pumping money into television ads, a consortium of left-leaning SuperPACs have pledged $100 million for door-to-door field operations this fall. The grassroots capabilities of the GOP once struck fear in Democratic hearts. Now it is the Obama campaign’s greatest comparative advantage.

The new empirical techniques will allow the campaign to sharpen the impact of its messages and maximize the efficiency of its resources. But just like the sabermetric revolution in baseball, if it is done correctly, the average spectator won’t notice the difference. She (since women vote more often than men, the median voter will be a she) will only think that the political ads seem a lot more relevant and that Obama is starting to look much more appealing.