Measuring What Matters: How Center for Common Ground Proved Their Postcards Move Voters

Most campaigns never find out whether their work moved the needle. They send postcards, make calls, knock doors, and then wait for election night to see if their candidate won. If the candidate wins, they assume their work helped. If not, they try harder next time. Either way, they’re guessing.

Center for Common Ground doesn’t guess. They measure. And their results from the 2024 general election are the strongest case I’ve seen that postcards actually move voters to the polls.

The Numbers

CFCG ran a postcard campaign targeting Black voters in Virginia and Georgia ahead of the 2024 general election. After the election, they matched their postcarded voters against voter file data to see who actually turned out. Then they compared those turnout rates to similar voters who didn’t receive postcards.

Virginia: 63,389 voters who received postcards were matched against 64,961 who didn’t. Among postcarded voters, 70.9% turned out. Among the comparison group, 62.8%. That’s an 8.2 percentage point difference.

Georgia: 857,474 postcarded voters versus 241,883 in the comparison group. Turnout was 64.8% for postcarded voters and 61.2% for the rest. A 3.6 percentage point lift.

To put those numbers in context: in the low-propensity voter research I covered last year, randomized field experiments showed that each piece of direct mail raises turnout by about half a percentage point. CFCG’s methodology is different (more on that below), but even accounting for that, the gap is striking.

How They Measured It

The key is the comparison group. You can’t just send postcards to everyone and then point at the overall turnout rate. You need to know what would have happened without the postcards.

CFCG used a matched comparison approach. They took their list of postcarded voters and found a similar group of voters who didn’t receive postcards from the same demographic. Then they pulled turnout data from the voter file after the election and compared the two groups.

This isn’t a randomized controlled trial. There are reasons to be cautious about the comparison: voters who received postcards may differ from those who didn’t in ways beyond the postcard itself. But it’s far better than no measurement at all, and the size of the effect in Virginia (8.2 points) is hard to explain away with selection bias alone.

CFCG doesn’t limit their measurement to postcards, either. They track each tactic separately, and the differences are revealing. In 2022 Petersburg, VA, voicemails produced a 5.9% turnout increase among Black voters. Live phone conversations in the same city, same election: 17.3%. In 2024 Mecklenburg, NC, canvassing produced a 2% early voting increase. Among postcard recipients invited to town halls, 90% showed up.

Each tactic gets its own measurement — that’s how you learn that a live phone call is worth three voicemails.

How to Set Up Your Own Measurement

You don’t need CFCG’s scale to do this. Here’s the basic approach.

Before your campaign: Hold back a random subset of your target list. If you have 10,000 addresses, send postcards to 8,000 and keep 2,000 as your comparison group. This is the part most organizations skip, and it’s the part that makes real measurement possible. It feels wrong to not contact people you could reach. Do it anyway. The knowledge you gain makes every future campaign better.

After the election: Pull voter file data for both groups. Calculate the turnout rate for each. The difference is your estimated effect.

Be honest about the limits. A holdout group gives you cleaner data than a matched comparison because you’re deciding who gets contacted upfront, not sorting it out after the fact. But even a matched comparison beats guessing. Start with what you can do.

Track across tactics. If you’re running postcards, phone banking, and canvassing, measure each one separately. CFCG’s data shows live phone conversations (17.3%) dramatically outperformed voicemails (5.9%) in the same city during the same election. You can’t learn that without measuring each tactic on its own.

Why This Matters for 2026

We’re eight months from the midterms. If you’re planning a postcard or letter campaign, build measurement into it now. Not after you’ve sent everything. Now. Decide your holdout group before you print a single card.

The organizations that measure their impact don’t just run better campaigns. They get better over time, because they know what’s actually working instead of going on gut feeling. And they can answer the question every funder and every potential volunteer eventually asks: how do you know this works?

CFCG’s full impact data is public. That transparency is itself a recruiting tool. When a potential volunteer asks “does this really make a difference?”, CFCG can show them exactly how much of a difference, down to the percentage point.

That’s the kind of answer you want to be able to give.