Fundraising is important. Polling and messaging, too.
But in the end, the Democratic U.S. Senate primary fight between U.S. Reps. Ed Markey and Stephen Lynch will come down to the map.
Who can win enough cities here, and wrangle enough towns there, to claim victory?
Most observers, at this point, are betting on the more liberal Markey. The latest indicator: a Suffolk University/WHDH poll that gives him the edge in three bellwether towns — Sandwich, Swampscott and Newburyport.
But the race is still competitive. And Lynch's camp says it can pull off the upset — says it can build a winning map.
So what would that map look like?
Campaign officials and independent observers say it could look an awful lot like the map a prominent Republican — former U.S. Sen. Scott Brown — put together in his special election upset of Democrat Martha Coakley in 2010.
Brown's coalition three years ago: the South Shore, large swaths of Cape Cod, big chunks of Essex and Middlesex counties in the northeast part of the state, the Worcester suburbs in the middle of the state and the Springfield suburbs in the southwest corner of the state.
Conor Yunits, a spokesman for Lynch, says a Democrat can replicate the Republican's success.
"People seem to forget that Scott Brown didn't win because there were suddenly a lot more Republicans in Massachusetts," he says. "He won because a lot of [independents] who lean Democratic swung their vote to him."
Of course, it's not clear that independents will vote in large numbers. The Senate special election has been a ho-hum affair from the start. And the Boston Marathon bombings have smothered the campaign in recent weeks.
Indeed, Massachusetts Secretary of State William Galvin is predicting a small turnout in the Democratic primary.
But the Lynch camp is confident it can exercise a Brown-like pull on centrist voters.
The Urban Vote
John Walsh, chairman of the Massachusetts Democratic Party, warns that a comparison between the two races is imperfect.
Lynch and Markey are competing in a Democratic primary, he notes, while Brown and Coakley faced off in a general election; those are different electorates.
But there are some notable parallels.
Take the urban vote. Markey, like Coakley before him, is surely counting on big wins in Boston, Cambridge, Worcester and Springfield.
And Lynch, no doubt, hopes to replicate Brown's success in small and mid-size cities like Attleboro, Lowell, Taunton and Quincy.
Lynch has some natural advantages in his urban push: blue-collar roots in South Boston and a congressional district that stretches into parts of the capital city.
Indeed, Lynch hopes to improve on Brown's poor performance in Boston.
Peter Ubertaccio, a political scientist and director of the Martin Institute at Stonehill College, says Lynch will have to perform well throughout his congressional district if he is to pull off the upset.
And that doesn't just mean winning cities and towns on his turf — it means building wide margins of victory.
Here's the problem for Lynch: Outside of Boston and possibly Brockton, he doesn't have a lot of room to build on Brown's 2010 performance.
Brown, after all, racked up big wins through much of Lynch's district: 60 to 39 percent in Hingham, 66 to 33 percent in Abington and 72 to 27 percent in West Bridgewater.
Markey appears to have more upside in his district, which stretches through the suburbs north and west of Boston.
Coakley lost several towns there — Ashland, Holliston, Melrose — and claimed narrow victories in several others: 53 to 47 percent in Framingham, 51 to 48 percent in Sudbury and 50 to 49 percent in Natick.
That's plenty of room to grow — especially for a Markey campaign that says it will contact 500,000 voters with calls and door knocks by the time the polls close.
If Markey performs well in his district and other strongholds, he'll be tough to beat.
In the end, Ubertaccio says, Lynch could win all the cities and towns Brown claimed in 2010 and still lose.
It is, in short, a tough map Lynch has to put together. But he's put together tough maps before.
This program aired on April 29, 2013. The audio for this program is not available.