As my colleague Sasha Issenberg writes in his wonderful forthcoming book, The Victory Lab, candidates and political parties started paying more attention to the science of campaigning after the election of 2000. That close race proved that when the outcome is tight, every little bit helps. So campaign strategists dropped the folk wisdom and started paying attention to the eggheads to learn whether a house visit, a phone call, or a piece of direct mail is more effective in identifying and turning out voters, or whether all three can be just as effective if deployed at different times and in different ways. …

Campaigns spend years putting together lists of their voters, identifying who they are, and noting what issues motivate them and, in some cases, what methods are most successful in getting them to vote. When early voting starts, most states release a daily list of who has voted. Campaigns can check these lists against their party lists. They don’t know how a person voted, but because they’ve usually been in contact with the voter—or they identified them at an all-night Romney rally—they can make a pretty good guess. …

Or, campaigns can know when to stop working a state and spend time in another. “In 2008, I knew that Obama had won the election a week before,” says Michael McDonald, a George Mason professor who has studied voter-mobilization techniques and used the same analysis the campaigns did. “We took a look at the early vote in Colorado. and it was implausible that McCain could come back.” (Obama officials hated this speculation because they worried it would suppress early turnout.)