The Washington PostDemocracy Dies in Darkness

How to read the leaked Michelle Nunn campaign plan

The National Review's Eliana Johnson got her hands on a remarkable document: a thorough outline of Georgia Senate candidate Michelle Nunn's (D) campaign plan as it stood at the end of last year. As an entrant into the canon of Documents that Turn the Tide of Political Campaigns, it is a non-starter, full of generally benign (if occasionally blunt) descriptions of the electorate. But as an example of Documents that Show How Political Campaigns Are Run, it's stellar.

The document (which you can read here) starts, as all campaigns start and all campaign plans should start, with the numbers. We made a graph that breaks down how the consultants figure Nunn gets to the magical 50-percent-plus-one number -- that is, a majority of the votes in the race.

The memo doesn't mention it, but there were about 7.2 million people over the age of 18 in Georgia in 2010. Of that number, about three-quarters are actually registered to vote, though as the memo notes, that figure varies. The campaign figures that 5.3 million will be registered by November.

This is the first assumption made in the document, and a key one. It's critical to know how many people are registered to vote in the state with some certainty, since all of the other numbers flow from that.

For example: The memo assumes 52 percent turnout in November. Applying the 52 percent to that 5.3 million number gives you a total voter pool of about 2.75 million. To win, a candidate in a two-person race needs 1.38 million votes. So the Nunn camp set her target at 1.4 million, to provide a cushion. That's more than most Democrats not named Obama have gotten in the state; usually, the memo says, Democrats get about 1.2 million votes "with strong campaigns."

Where'd the 52 percent figure come from? That's precisely the turnout the state had in 2010, although the memo says that turnout that year was 51 percent. (The consultants who drafted the memo added one percentage point, assuming increased turnout.) It's not clear if this is the final figure that Nunn is using, but you can be sure it's one that her opponent is paying close attention to.

Next, the memo outlines where those 1.4 million votes will come from. One key argument is that the campaign hopes to "discourag[e] the kind of Republican surge we saw in 2010," to keep the vote total at about 2.7 million. Otherwise:

The large slices are groups that went with Democrat Roy Barnes' gubernatorial bid in 2010: 412,000 white Democrats and 700,000 black Democrats. The former figure was 23 percent of the demographic's vote; Nunn's team figures they can hit 40 percent, adding another 160,000 votes or so. Add in another 150,000 black votes from a surge in interest (not to mention a competitive top of the ticket with Jason Carter running for governor) and another 50,000 votes from Latinos, Asians, and "young urban whites," and they're over the 1.4 million mark.

Literally everything else in the document is about making this plan come true. It's pages and pages of identifying constituency groups and figuring out who good surrogates will be and so on -- but it all flows from the campaign's expectation of who's actually going to come to the polls.

There are lots of other details worth pulling out of the document, but we're going to stop at two. The first is this diagram, created by bluelabs.

This looks complicated, but it isn't. In order for the campaign to hit its number targets, it needs to both predict and track support among voters. The campaign plan articulates how to do that -- polling to identify areas of support and doing phone sweeps meant to identify broad swaths of the population. What bluelabs is describing is a more sophisticated tool for evaluating how people feel about the campaign.

But what's interesting about the graph is the three buckets into which it puts people. There are the people who the campaign should ignore, those who aren't likely to support Nunn, who aren't circled. There's the group that needs to hear the campaign's message, those people in the middle range who are circled. Pitched with the campaign's message, they'll go vote for Nunn, or so the theory goes. Then there's the group at right, the people who strongly support Nunn. That group needs a different message: get to the polls and maybe give money.

You can think of this graph as one side of a two-dimensional chart. Instead of the vertical axis presented above, think about how regularly people vote -- information that's usually easy to ascertain (although Nunn's team points out that Georgia's voter history data is iffy). Maybe you have a person in that far right spike of potential donors, but who never votes. The campaign's goal is to get that person to the polls -- or, if they literally never vote, to give up on them with a sigh. Or maybe you have someone who always votes but fits in that middle pool. Those people need to hear Nunn's message, and soon.

And then there's the group of people who vote a lot but fall on the left side of the bluelabs spectrum. Those are the people that Nunn's team hopes decide not to recreate the surge of 2010.

One more thing. Nunn's team explains the pitfalls of modern campaign communications.

The political press is not inclined to cover a candidate repeating their message. In fact, many reporters see their job as getting the candidate to "reveal" what their "true" inclinations and orientation may lay or to cause a gaffe. Any deviation from that message will be newsworthy to them. They also understand that effective candidates and campaigns stick to their message, and will see a deviation in message as an erred campaign or candidate.

Jerk reporters always trying to "trick" candidates into "not repeating campaign gibberish"! Just imagine what a field day of gotchas the press corps would have if they stumbled onto an accidentally leaked campaign plan.