“The campaign that efficiently identifies and turns out its own voters wins — consistently, and at lower cost — compared to the campaign that broadcasts to everyone and persuades no one in particular.”
— Bull Moose Strategy LLCExecutive Summary
Down-ballot races are decided by hundreds or low thousands of votes. In races at that scale, the most consequential strategic decision is not which message to run or which digital platform to buy. It is whether the campaign identifies its own voters and then actually gets them to the polls.
That is the ground game. And the evidence on what works in the ground game — accumulated over two decades of randomized field experiments across hundreds of campaigns — is clear enough to drive specific, concrete decisions about where every campaign dollar goes.
This paper distills that evidence into a framework for down-ballot R/I candidates: the contact-method hierarchy with relative effectiveness and cost-per-vote estimates, the voter universe targeting that makes any tactic efficient, the volunteer structure required to execute at local scale, the 72-hour final push that closes competitive races, and how precision voter intelligence makes every element of the ground game more productive.
The thesis is simple: targeting tells you who to contact; the evidence tells you how to turn them out; the 72-hour push executes it. Cost-per-vote is the decision metric. Everything else is secondary.
The Turnout Principle: Why GOTV Is Not a Persuasion Problem
Most first-time down-ballot candidates make a foundational error: they try to persuade everyone. They campaign broadly, knock every door in the precinct, and hope that wider reach produces a winning margin. This approach loses more races than any opponent ever does.
The political science research on voter contact is unambiguous. Most voters’ behavior is not meaningfully changed by most campaign contact. Their minds are made up, they are not going to vote, or they are so marginally engaged that no brief interaction will move them. The exception is personal, high-quality contact with voters who were already predisposed to support the candidate but might not show up on election day. Those contacts move turnout. Generic, impersonal broadcast to everyone moves almost nothing.
This distinction — between turnout activation and persuasion — is the most important strategic concept in down-ballot field operations. A voter who supports the candidate but stays home is identical, in outcome terms, to a voter who goes to the polls and votes for the opponent. The campaign’s job is to make sure the first type of voter does not stay home. That is a turnout problem, not a persuasion problem.
The two populations that matter for GOTV: First, registered voters predisposed to support you who have inconsistent voting histories — they have voted in some elections but skip lower-turnout local cycles. These are your highest-value targets. They already agree with you; they just need a reason to show up. Second, genuine soft persuadables who could go either way. These exist in every race, but there are fewer of them than campaigns assume. Saturate your base activation universe before shifting resources to persuasion outreach.
The Local Math: Why This Matters More Down-Ballot
In a county commissioner primary with 7,500 voters casting ballots, a campaign that identifies 3,500 likely supporters in the voter file and runs a disciplined GOTV program that moves 15% of those supporters from “likely to stay home” to “voted” generates 525 additional votes. If the race is close, that margin wins.
That same campaign could spend the same dollars on a radio buy reaching 40,000 people in the county — most of them not registered, wrong precinct, or not in any moment of political receptivity. The radio buy generates no measurable turnout lift among identified supporters. The GOTV program wins the race.
The local math principle is exact: in most down-ballot races, the number of potential additional votes from quality GOTV contact within an existing identified-supporter universe exceeds the number of genuinely persuadable voters available to swing. Prioritize turnout over persuasion. Saturate the universe before expanding it.
Before any tactical planning, run this calculation:
- What vote total wins? Research historical results and model the expected turnout for this cycle.
- How many identified supporters are in your voter file universe? How many voted in the last comparable election?
- The gap between those two numbers is your GOTV target. Every tactic should be evaluated by how efficiently it closes that specific gap.
The Contact-Method Hierarchy: What the Evidence Shows
The field experiment research produces a consistent, replicable ranking of voter contact methods by their effect on turnout. These are not strategic preferences. They are measured outcomes from controlled trials across hundreds of campaigns and geographies.
Door-to-Door Canvassing — The Proven Standard
Personal, face-to-face canvassing is the most consistently effective GOTV method in the political science literature. Rigorous field experiments have found that door-to-door canvassing generates roughly one additional vote for every 14 quality contacts made — with the range across studies falling between 1 in 10 and 1 in 18 depending on conversation quality, universe targeting, and timing.
The personal element is not incidental. It is the mechanism. A real human being, at someone’s doorstep, making a personal ask to participate in a specific local election, activates social norms around civic participation that impersonal contact cannot replicate. Voters who receive a genuine personal visit feel a social obligation to follow through on their stated intention to vote that persists through election day.
The quality caveat is critical: sloppy canvassing can approach net zero impact. A canvasser who reads from a clipboard, fails to engage, and leaves without completing the conversation generates no meaningful turnout lift. Training, scripts, and conversation quality matter as much as door count. The goal is not doors knocked. The goal is quality contacts that activate the voter’s intention to participate.
Volunteer Phone Banking — Genuine Conversation Required
Phone banking with genuine, two-way conversation — not scripted recitation — shows consistent but smaller effects than in-person canvassing. Field experiments find volunteer phone banking generates roughly one additional vote per 35–50 conversations.
The key word is genuine. Phone banking that reads from a script and hangs up when the voter hesitates produces effects similar to direct mail — modest at best. Phone banking where the volunteer truly engages, asks about the voter’s concerns, and creates a real moment of connection approaches (though does not equal) the impact of a door visit.
Direct Mail — Social Norms Over Generic GOTV
Direct mail consistently shows small but measurable turnout effects — approximately one additional vote per 200–400 pieces mailed, depending on content type and universe precision. The critical finding: social norms mail dramatically outperforms generic GOTV mail.
The social norms effect is one of the most robust and replicable findings in the field experiment literature. Voters who receive a message stating that a high percentage of their neighbors voted in the last election turn out at measurably higher rates than voters who receive generic encouragement. “In this township, 74% of registered Republicans voted in the last primary” is a more effective GOTV message than “Your vote matters.” The mechanism is social proof: voters are motivated to conform to a perceived community norm of participation.
Robocalls — The Evidence Is Unambiguous
Multiple controlled studies have found that robocalls produce effectively zero measurable turnout lift. They are broadly disliked, widely ignored, and generate no personal obligation for the voter to act. The exception — logistical reminders about polling locations or ballot return deadlines — shows modest utility. As a persuasion or activation tool, robocalls are essentially inert.
Any robocall budget in a down-ballot campaign should be redirected to canvassing volunteer recruitment, additional mail pieces, or the 72-hour push logistics described below.
| Method | Turnout Effect | Est. Cost/Vote | Priority |
|---|---|---|---|
| Candidate personal canvassing | Highest | $15–35 | Highest |
| Volunteer door-to-door canvassing | High | $20–50 | High |
| Volunteer phone banking (genuine conversation) | Moderate-High | $35–70 | High |
| Social norms direct mail (targeted universe) | Moderate | $50–100 | Medium-High |
| Standard direct mail (targeted universe) | Low-Moderate | $100–150 | Medium |
| Professional phone banking | Low | $100–200 | Voter ID only |
| Robocalls | Negligible | Not measurable | Last resort |
Building the Right GOTV Universe
No GOTV tactic produces its expected return unless it is deployed against the right universe. Canvassing the wrong doors generates zero turnout lift regardless of volunteer quality. Calling the wrong voters generates pushback and wasted time. The universe is the foundation on which everything else rests.
The universe definition: the specific, named voters who have a meaningful probability of supporting your candidate and are at risk of not voting without contact. These are your targets. Everyone else is not.
Universe Construction Criteria
Define your GOTV universe using a combination of the following data layers:
Registration and party data. In states with party registration, registered voters of the relevant party are your base filter for a primary. For general elections, combine registration status with geography and voting history as the starting frame.
Voting history. Voters who have voted in at least some elections but missed the last comparable local cycle are your highest-value base activation targets. Reliable primary voters who have never missed an election don’t need GOTV resources — they will show up. The inconsistent voters who might skip this one are where contact hours pay off.
Support scores. Modeled support probability — which voters in the universe are most likely to support the candidate based on demographic patterns, geographic patterns, and public data signals — concentrates the highest-intensity contact on the voters most worth reaching. High support probability plus low-to-moderate turnout propensity is the gold tier. Prioritize those voters for canvassing and volunteer phone calls. Lower support probability segments can receive mail if resources allow.
Geographic concentration. In local races, universe targeting typically maps cleanly to 8–15 precincts where supporter concentration and untapped voter potential are highest. These are the canvass precincts. Do not try to cover every precinct at equal intensity with limited volunteers. Cover the right ones at high intensity.
Universe Tiers
Divide the universe into contact tiers that match resource allocation to voter value:
Tier A: High support probability + low-to-moderate turnout propensity. Receives canvassing (multiple contacts), volunteer phone banking, at least two mail pieces. This is where the race is won or lost.
Tier B: High support probability + moderate-to-high turnout propensity. Receives canvassing (at least one contact), one mail piece. They will likely show up; confirm it and move on.
Tier C: Moderate support probability + low turnout propensity. Receives mail, phone banking if capacity remains after saturating Tiers A and B. No canvassing unless the higher-priority tiers are exhausted.
When campaign capacity is limited, cutting the universe is more efficient than diluting it. A Tier A universe canvassed twice beats a full county voter list canvassed once at a fraction of the intensity.
The Candidate at the Door
No campaign resource replicates the turnout impact of the candidate personally knocking doors. Voters who receive a personal visit from the candidate themselves turn out at higher rates and are more likely to vote for that candidate than voters who receive any other form of contact. The research on this is consistent across jurisdictions and races.
The mechanism is intuitive: running for local office and then personally visiting voters in the neighborhoods you want to represent signals authenticity and genuine connection in a way that is difficult to fake and impossible to replicate at scale with paid staff. A volunteer showing up for a candidate is meaningful. The candidate showing up for themselves is something different entirely.
In practical terms, a down-ballot candidate running a disciplined door-knocking program — weekend mornings, weekday evenings, targeting the Tier A universe precincts in their walk list — can knock 3,000–5,000 doors personally over the final 8 weeks of the campaign. In a race decided by a few hundred votes, that is the race.
The door conversation is not a speech. It is a brief, genuine exchange: introduce yourself, one locally-grounded sentence about what you’re running on, a direct ask for support, and then — crucially — listening. The voter who is listened to feels more connected to the candidate than the voter who was talked at. Let them talk.
The 72-Hour Final Push
The final 72 hours before the polls close are the highest-leverage GOTV window in the entire campaign. Voters who were soft earlier become more receptive to contact. Infrequent voters make final decisions about whether they will make the trip to the polls. The absentee chase enters its last window. And the concentrated energy of a fully deployed volunteer operation generates a social momentum that reinforces each individual voter contact.
Structured, disciplined final-72-hours ground programs have been proven at scale across Republican campaigns at every level, adapted from the systematic precinct-captain-led GOTV operations developed in the early 2000s and since refined across dozens of election cycles. The core framework translates directly to down-ballot races — it just requires size-scaling to local geography and volunteer capacity.
The 72-Hour Operations Sequence
Sunday (72 hours out): Full volunteer deployment. Every available canvasser in the field, focused on Tier A precincts only. Absentee chase calls to all voters who requested a mail ballot but have not yet returned it. Final mail piece timed to arrive today or tomorrow.
Monday (48 hours out): Second canvass pass through Tier A precincts — anyone not reached Sunday gets a second attempt. Volunteer phone banks running from noon through 8pm. Candidate at doors through 7:30pm. All election day volunteer assignments confirmed by text.
Election day: Volunteer greeters at key polling locations in target precincts (where legally permitted). Last-chance phone calls from noon through close to identified supporters who have not yet voted. Data tracking: which supporters have voted, which are outstanding. The goal is zero identified Tier A supporters who are still available to turn out left uncontacted by 7pm.
The Sunday–Monday window is underused in most local campaigns. Campaigns routinely hold back volunteer capacity for election day and underuse the Sunday–Monday push. The evidence supports the opposite priority: election day is often too late for the marginal contact that moves a low-propensity voter. The weekend before is when that contact matters most. Concentrate intensity then — not day-of.
Volunteer Deployment Structure
The 72-hour push requires every volunteer in the campaign’s network to show up at once. That means recruitment and confirmation beginning six weeks out, not six days out. Specific territory assignments for every canvasser, generated fresh from the non-contacted universe before each shift. Real-time field communication between the field director and all team leads via group text or app. End-of-shift debrief: which precincts are saturated, which need more coverage tomorrow.
Every contact made in the 72-hour window is recorded and entered into the universe data the same day. The non-contacted list shrinks with each shift. The goal on election eve is a universe where every reachable Tier A voter has been contacted at least once — and the unreachable ones have a phone call queued for election day.
Early Voting and Absentee Chase
In most states, 20–45% of votes in local elections are cast before election day — by mail, absentee request, or in-person early voting. This collapses the old model of election-day-only GOTV into a multi-week window where contact timing against individual voters matters as much as contact method.
The absentee chase program is straightforward in concept and operationally demanding in execution: track, by name, which voters in your GOTV universe have requested a mail ballot and which have returned it. Voters who have requested but not yet returned a ballot are your chase targets. They have already signaled intent to vote. Your job is to get them to complete the transaction before the return deadline.
Chase tactics in the final 10 days: phone calls and texts to outstanding-ballot voters, door contacts for those within canvass range in the final 2–3 days, and email or text reminders with return instructions. Update the outstanding list daily. The “requested, not returned” column of your absentee tracker is your daily priority target for the final two weeks.
In-person early voting should be treated as an accelerated version of election day: encourage identified Tier A supporters to vote early, track their completion, and redirect those confirmed votes off the GOTV target list. A voter who has already cast a ballot is a locked vote — stop spending contact resources on them and shift to the outstanding targets.
The Targeting Layer: Where Voter Intelligence Makes the Difference
Every component of the ground game framework described in this paper is more productive when the underlying voter targeting is precise. The contact-method hierarchy tells you how to turn voters out. Targeting tells you which voters are worth turning out and in what order.
The connection between these two layers is the core reason why data-driven voter intelligence is not a luxury add-on for campaigns with big budgets. It is the mechanism that makes the evidence-based GOTV model efficient enough to work at the county level.
Consider the practical difference:
A campaign with a full county voter list and unlimited enthusiasm can knock every registered Republican’s door in a township. It will use three times the volunteer hours to generate half the turnout lift of a campaign that started from a scored, tiered universe of 800 high-probability supporters with inconsistent voting histories in six target precincts — and knocked those 800 doors twice.
The difference is not canvassing skill or volunteer quality. It is universe precision. The second campaign put its contact resources against the voters where each contact produces measurable turnout impact. The first campaign spread the same resources across the full voter file, diluting the effect to near zero on most of its contacts.
BMS Voter Intel — the voter intelligence infrastructure developed by Bull Moose Strategy for down-ballot R/I campaigns — applies exactly this logic: a scored, segmented voter universe built from multiple public data sources, designed specifically to support the kind of tiered targeting that makes evidence-based GOTV efficient at the county and township scale. The targeting layer is what makes the ground game pay off.
This paper complements our published Voter Intel whitepaper, which covers the data architecture and universe-building methodology in detail. The field operations framework here assumes that universe has been correctly built. Without it, any GOTV program is working at a significant efficiency disadvantage.
Conclusion
The evidence on what moves voters to the polls is not ambiguous. Personal contact works. Quality conversation works. Social norms messaging works. Broad, impersonal broadcast — robocalls, generic mail, unfocused reach buys — largely does not. And the campaign that saturates a correctly targeted universe of its own supporters with high-quality contact will, in most down-ballot races, outperform a campaign that spends the same dollars reaching more people less effectively.
The local math makes this more decisive, not less. A county commissioner race decided by 400 votes is a race where a disciplined GOTV program in the right precincts is the margin. The candidate who knocked 4,000 doors personally, whose volunteers canvassed the Tier A universe twice, whose absentee chase confirmed 600 outstanding ballots, and whose 72-hour push left no reachable supporter un-contacted — that candidate wins competitive down-ballot races.
That campaign does not require a large staff or a major party apparatus. It requires a correctly targeted universe, a trained volunteer team, a candidate willing to knock doors until the night before the election, and a 72-hour push plan executed with discipline. Those are resources available to any serious down-ballot R/I candidate.
Targeting tells you who to contact. The evidence tells you how to turn them out. The 72-hour push executes it. Cost-per-vote keeps every decision honest. That is the ground game.