How to plan a get-out-the-vote program that moves turnout, based on what the field experiments show

The get-out-the-vote tactics with the strongest evidence are personal ones. In six-city field experiments in 2001 local elections, a face-to-face conversation at the door raised turnout among voters reached by about 7.1 percentage points, roughly one additional vote for every 14 conversations in that study, and mail showing voters their own and their neighbors' turnout history raised turnout by 8.1 percentage points. Mass, impersonal tactics like broadcast texts and email typically move turnout by less than one percentage point, and robocalls show little or no effect. A GOTV plan is the work of pointing your most personal contact at the supporters least likely to vote, early enough to bank ballots.

What a GOTV program is, and what it is not

GOTV is about turnout, not persuasion. You are not changing anyone's mind. You are making sure people who already support your candidate or cause actually cast a ballot.

That distinction shapes the budget. By the final weeks of a general election most minds are made up: a review of 49 field experiments by Kalla and Broockman found the average persuasive effect of campaign contact in general elections was essentially zero. Turnout effects, by contrast, show up again and again in randomized trials. So the GOTV phase puts its money and volunteer hours into the people already with you, where contact reliably converts into votes.

A complete GOTV program has six parts: a target universe, a ballot plan for every voter (early, mail or Election Day), contact channels, staffing, scripts and a final-days schedule. The rest of this guide takes them in order.

Build your GOTV universes

The universe is the list of voters you will chase. Get it wrong and every other number in the plan is wasted.

Start with the voter file and score each voter on two things: likelihood of supporting you, and likelihood of voting. Your GOTV universe is the box where support is high and turnout is uncertain. Voters who cast a ballot in every election need a reminder at most. Voters who never vote and are not with you are not a GOTV target.

Size the universe against your vote goal. If you need 40,000 votes to win and can count on 32,000 from reliable voters, your GOTV program exists to produce the other 8,000 from the core universe.

Early vote and vote-by-mail chase

Election Day is no longer one day. In the 2024 presidential election, the Census Bureau found that 39.6% of voters cast ballots in person on Election Day, 30.7% voted early in person and 29.0% voted by mail. A plan that starts on the final weekend has missed most of the vote in many states.

Run the chase in this order:

1. Ballot requests. Where voters must request a mail ballot, push your supporters to request one, and track who did.
2. Ballot tracking. Pull the state or county ballot file daily once voting starts and sort every supporter into one of three groups: returned, accepted, or flagged for a cure. Supporters whose ballots are accepted come off your turnout lists immediately, so volunteers spend their time on people who have not voted. Supporters whose returned ballots are flagged move to the cure team, not off the list.
3. Unreturned ballots. Supporters holding a mail ballot get repeated, specific reminders: where the drop box is, what the postmark rule is, whether a witness or notary signature is required.
4. Ballot curing. Many states notify voters when a mail ballot is rejected for a missing or mismatched signature and give them a short window to fix it. Have a team ready to call and visit those voters the day the list updates.
5. Early in-person vote. Move people with transportation or schedule problems to early voting, when lines are shorter and you can still recover from a problem.

Every voter you bank early is a voter you do not have to chase in the last 72 hours.

What get-out-the-vote tactics are most successful?

The best evidence comes from randomized field experiments, many run by Alan Gerber and Donald Green at Yale and summarized in their book Get Out the Vote. The rule that runs through two decades of studies: the more personal the contact, the bigger the effect.

Read the table with three cautions. First, effects are measured on different bases (people reached versus people assigned), so they are not perfectly comparable. Second, cost-per-vote figures from the 2000s used wages of $10 to $16 an hour; at today's canvasser pay of roughly $22 an hour on Indeed, multiply accordingly. Third, social pressure mail works but can irritate voters. Many campaigns use a softer version that thanks people for past voting rather than threatening to publicize their record.

Our ranking for a typical campaign: doors first, relational asks second, live calls with a plan-making script third, mail and texts as reinforcement, robocalls never as a primary tactic.

Staffing and volume math

Work backwards from votes, and plan against a range of outcomes rather than a single promised conversion rate. The 2001 six-city studies are a historical benchmark from specific low-turnout local elections; your electorate, election type and program will differ. Say your GOTV program needs to produce 1,000 extra votes.

Two rules govern the math. Count unique voters reached, not total conversations: a second or third conversation with the same supporter reinforces the first, but it is not a fresh chance to produce another vote. And size the universe so it holds far more supporters than the number of conversations you need, because contact rates mean you will reach only a share of it. Planning guides put suburban contact rates at about a quarter to half of doors, or roughly 5 to 8 conversations per hour.

Spread over the final three weeks, the benchmark scenario alone is roughly 28 to 44 shifts a day, before any repeat passes. Budget toward the moderate or conservative row unless you have tested results from your own electorate. If you do not have that many volunteers, you either bring in a paid canvassing team (see our canvassing cost guide for market rates), shrink the universe to the voters with the most upside, or both.

For staffing, plan on one field organizer for every 50 to 75 regular volunteers and one shift lead per staging location per shift. Those are the ratios we hold to on our own field programs, and we tighten them in the last week, when you are running far more shifts with less experienced people.

GOTV scripts that work

Short scripts beat long ones in the final weeks. The research points to a few elements worth including:

• A plan, not just a reminder. Ask when they will vote, where, and how they will get there. Plan-making is the element that moved turnout in the Nickerson and Rogers study.
• The specific information they need. Early voting site and hours, mail ballot return rules, ID requirements.
• A positive social norm. In two experiments, Gerber and Rogers (2009) found that messages stressing high expected turnout raised intent to vote more than messages lamenting low turnout, especially among occasional voters. Use a high-turnout message only when it is true and you can point to evidence for it, such as early-vote numbers running ahead of the last comparable election. Never say "nobody votes."
• A commitment. Ask for a yes and write it down. Call back the yeses in the final days.

A sample door script: "Hi, I'm Jordan, a volunteer with [campaign]. Early voting is open at the library through Saturday. Have you voted yet? Great, when do you think you'll go? Will you be going from home or work? Can I count on you?" Partisan campaigns add the candidate's name and one reason. A nonpartisan 501(c)(3) program leaves both out.

The last 72 hours: a GOTV plan

The final three days are about execution, not new ideas. Everything should already be printed, cut and loaded.

Two rules for the final days. Never send volunteers to a door you could have removed from the list, and never drop a supporter whose returned ballot still needs a cure. And keep a rapid response number for voters who hit a problem at the polls.

Poll monitoring basics

Poll watchers and challengers are regulated state by state. The NCSL summary of state rules covers qualifications, training and what watchers can and cannot do, and the National Association of Secretaries of State keeps a comparable chart. In most states, watchers must be appointed in advance by a party or candidate, and the number per polling place is limited, as the Brennan Center notes.

For a campaign or nonpartisan group, the practical job is narrow. Train monitors on your state's rules, have them observe without interfering, write down problems with time and detail, and report them to a hotline staffed by someone who can reach election officials or counsel. A monitor who argues with poll workers hurts voters. One who documents a broken machine at 7:15 a.m. and gets it reported helps them.

Relational organizing: GOTV through people voters know

Relational organizing asks volunteers to turn out people they already know: family, friends, coworkers, congregation members. It is built on the same finding as door knocking. Personal contact works, and contact from someone you trust works even better.

The evidence behind it is some of the strongest in the field. In a randomized evaluation of the Turnout Nation captain model in 2019 municipal elections, voters assigned to captains in their own social networks turned out at rates 13.2 percentage points higher than the control group, which the authors called the largest intent-to-treat effect in two decades of GOTV experiments. A study of friend-to-friend texts through the Outvote app in the 2018 midterms found large effects among contacted friends, though with a wide confidence interval. Execution decides the size of the effect. A lighter-touch 2022 Georgia runoff program run by Vote Rev measured a 0.51-point effect, which is why the structure below matters.

How to make it work:

• Ask each volunteer to build a list of 10 to 30 people they will personally turn out, then match that list to the voter file.
• Give volunteers a simple app or spreadsheet, a reminder schedule and the exact information their people need.
• Treat relational lists like any other universe. Check them against the returned-ballot file and chase who has not voted.
• Start early. Relational lists take weeks to build and are hard to stand up in the final ten days.

Frequently asked questions

What get out the vote tactics are most successful?

Personal, face-to-face contact has the strongest record. In six-city field experiments in 2001 local elections, door-to-door canvassing raised turnout by about 7.1 points among voters reached, roughly one extra vote per 14 conversations in that study, and relational organizing through people voters know has produced some of the largest effects ever measured. Live phone calls that help voters make a plan also work. Mass texts, email and robocalls add little on their own.

How do you plan a get out the vote campaign?

Set a vote goal, build a universe of supporters with uncertain turnout, decide the ballot plan for each (early, mail or Election Day), then assign contact channels and staffing to reach them several times. Track ballots daily, remove supporters whose ballots have been accepted, and route rejected or flagged ballots to a cure team. Write the final-72-hours schedule weeks in advance.

How can you increase voter turnout?

Contact people personally, help them make a concrete plan, and give them the exact information they need about where and when to vote. Reach people early so they can vote early or by mail, and follow up with anyone who has not voted. Relational outreach from friends and family is one of the strongest tools available.

Are text messages effective for GOTV?

Mass reminder texts help a little. A pooled analysis of four Danish experiments found about a 0.74-point increase in turnout. Texts are cheap and useful for logistics (polling place, hours, ballot deadlines), but they should support personal contact, not replace it. Friend-to-friend texts appear to be much more effective than texts from strangers.

Is get out the vote partisan?

It can be either. Campaigns, parties and PACs run partisan GOTV aimed at their supporters. Nonprofits, including 501(c)(3) charities, can run nonpartisan GOTV as long as it does not favor or oppose any candidate or party in who it targets or what it says.

When should GOTV start?

Start the ballot plan as soon as mail ballots can be requested and early voting dates are known, which in many states is six or more weeks before Election Day. Relational lists and volunteer recruitment should start even earlier. The heaviest contact comes in the final three weeks, peaking in the last 72 hours.

What is relational organizing?

Relational organizing is outreach through people's existing relationships. Volunteers commit to turning out friends, family and others they know, usually tracked through an app or list matched to the voter file. Randomized studies have found some of the largest turnout effects ever measured, with the biggest gains in programs that recruit committed captains and track their lists closely.

What is a GOTV universe?

It is the list of voters a campaign will contact to turn out. It usually includes supporters with low or middle likelihood of voting, new registrants and voters who have requested but not returned mail ballots. Reliable voters and non-supporters are generally left out.

How Leo Operations approaches this

Leo Operations is a national field, media and creative firm, and GOTV is core to the field work we do, alongside paid canvassing, voter registration and petition drives. Our founder, Rodriquez Wright II, has worked campaigns since 2012. He ran regional GOTV for Hillary Clinton's 2016 New Hampshire primary campaign and served as Tennessee State Organizing Director for Mike Bloomberg's 2020 presidential campaign. Our field programs have knocked more than 650,000 doors across 25 states.

We build GOTV from the vote goal backward. First comes a universe your team can actually contact and a ballot plan for every voter in it: early, mail or Election Day. Then we set staffing to the shift math, load daily returned-ballot updates so nobody wastes a knock, and write the last-72-hours schedule weeks before anyone needs it. Doors and relational contact sit at the center of the program because that is where the evidence points, with texts, calls and mail reinforcing them on a set schedule.

If you are planning a turnout program, email us your race, election date and budget, and we will tell you honestly what it takes to hit your vote goal.