Chapter 21 — Campaign Operations: How Modern Campaigns Are Actually Run

"I spent more time asking strangers for money than I spent legislating, debating, talking to constituents, or thinking. Anybody who tells you otherwise is either a liar or has never run for federal office." — Tom Allen, former U.S. Representative (D-ME), reflecting on his 2008 Senate campaign in Dangerous Convictions (2013)

"If you want to win, you need three things: a candidate, a plan, and the discipline not to deviate from the plan when the candidate gets bored. The third one is the hardest." — Mike Murphy, Republican strategist, in Hacks on Tap podcast (2022)

Most American Government textbooks describe elections at the level of theory — voter coalitions, the median voter, Duverger's Law — and at the level of outcome — who won, by how much, in which states. They almost never describe the level in between: the actual operation of a modern federal campaign. The 80-hour weeks, the call sheets, the volunteer turnover, the data ops, the panic at the 11 PM tracking-poll release, the field office where the air conditioning is broken in July and the Wi-Fi is slow in October.

This chapter is about that middle level. It is unusual for a textbook because almost no professor of American Government has run a campaign, and almost no campaign professional writes textbooks. The two worlds rarely touch. But they should, because campaign operations shape political outcomes in ways that voter-behavior models do not capture. Two candidates with identical positions, identical districts, and identical opponents can produce wildly different vote totals based on the quality of their operation. Knowing how those operations work — what the people inside them actually do all day — is part of understanding American politics.

21.1 The Candidate Decision: What It Actually Takes to Run

Before there is a campaign, there is a person deciding to run. That decision looks easier from the outside than it does from the inside, and the gap between perception and reality is wide enough to surprise students who have never been close to it.

The personal financial cost

Running for federal office in 2026 is not free for the candidate. Federal law prohibits using campaign funds for personal expenses (FEC regulations 11 CFR § 113.1), so the candidate's living costs during the campaign mostly come out of personal savings or are absorbed by a spouse's continued income. House candidates who keep their day jobs (or run from a position that allows flexibility, like a state legislator's part-time role) can manage. Candidates who must leave their jobs to run — for an open House seat or for a competitive Senate race — face a year or more of reduced or zero personal income.

Beyond income, there are direct costs the candidate cannot bill to the campaign: clothing for events, child care during fundraising calls, travel to events that double as campaign and personal (a candidate's child's wedding visited by donors). The line is policed by FEC opinion letters and occasionally by enforcement actions, and most candidates err on the side of paying out of pocket.

Steve Israel, who chaired the Democratic Congressional Campaign Committee from 2011 to 2015, wrote in Big Guns (2017) and elsewhere that he routinely told potential candidates: "If you don't have $100,000 in personal savings or a spouse who can carry the household for two years, you cannot afford to run for an open House seat. I'm sorry, but the system is what the system is." Republican recruiters say similar things in private. The candidate-pool consequence — that running for federal office is increasingly limited to people with substantial personal wealth, a wealthy spouse, or a willingness to take on debt — is one reason Congress underrepresents low-income Americans even more than it underrepresents racial or ethnic minorities (Carnes 2018).

The opportunity cost

The bigger cost is usually not financial but professional. A serious campaign for federal office is a 70-to-90-hour-a-week job for 12 to 24 months. Existing careers do not pause for that. A doctor who takes 18 months off to run for the House loses patient relationships that may not return. A small-business owner has to delegate or sell. A junior partner at a law firm passes up the partnership track. An academic loses tenure-clock years.

For incumbents, the opportunity cost is muted by the fact that campaigning is effectively part of the incumbent's job — the constituent meetings and town halls overlap. For challengers, the opportunity cost is total. This asymmetry is one mechanism behind the well-documented incumbency advantage in U.S. House elections (currently 95%+ reelection rates for sitting members, FEC and Cook Political Report data, 2024 cycle). It is harder to leave a real career to run than it is to keep doing the campaign-and-job hybrid most members of Congress already do.

The "invisible primary" and party screening

By the time a candidate publicly announces, a significant winnowing has already happened. Both parties run formal candidate-recruitment operations: the Democratic Congressional Campaign Committee (DCCC) and the National Republican Congressional Committee (NRCC) for the House; the Democratic Senatorial Campaign Committee (DSCC) and the National Republican Senatorial Committee (NRSC) for the Senate. These committees identify potentially competitive seats, recruit candidates they believe can win, and steer party donors toward favored recruits.

Political scientists call the period before formal primary contests the "invisible primary" (Cohen, Karol, Noel, and Zaller 2008, The Party Decides). During this phase, party leaders, major donors, allied interest groups, and prominent endorsers indicate their preferences. A candidate who fails to win major endorsements, fails to raise enough money to demonstrate viability, or fails to accumulate enough state-legislator and county-chair support typically drops out before the first primary debate.

The process looks different in the two parties since roughly 2016. The Republican invisible primary has been disrupted by the Trump-era reorganization, where the most consequential endorsement is not the Wall Street Journal editorial board or the National Review masthead but a Trump-issued endorsement (often via Truth Social). On the Democratic side, the invisible primary has remained closer to the Party Decides model, though challenged by movement-aligned candidates (Justice Democrats endorsements, Working Families Party slate) operating outside the traditional party-leader channel.

The signal of major endorsements

When students see a list of endorsements in a news article and assume it is decorative, they are mistaken. Major endorsements function as signals — to donors, to other politicians considering whether to enter the race, and to media outlets deciding which candidates to cover. The signal is most informative when the endorser has a record of picking winners and a reputation for not endorsing lightly. Endorsements from sitting members of Congress in the same delegation, from former presidents of the same party, from major unions or trade associations aligned with the party, and from the leading state-level officials carry the most weight. Endorsements from celebrities and from organizations that endorse everyone in the party carry the least.

A practical illustration: in a 2024 House primary in a competitive Pennsylvania district, the early endorsement of the AFL-CIO state federation, a state senator from the largest county in the district, and the previous nominee for the seat (who had narrowly lost the general two cycles earlier) produced a signaling cascade that closed the race well before primary day. Two other potential candidates declined to enter; donors directed their checks to the endorsed candidate; local media coverage treated the endorsed candidate as the presumptive nominee. The candidate's individual qualities mattered, but the endorsement signal — the visible coordination among the people whose opinions other party actors trusted — produced a rapid winnowing of the field.

The signaling logic also explains why some endorsements are sought even when the endorser has no formal political power. A respected former federal prosecutor's endorsement of a candidate in a competitive Senate primary signals to other party actors that the candidate has been vetted and is unlikely to produce embarrassing news. A prominent military veteran's endorsement signals that the candidate is acceptable to a constituency that values that affiliation. The endorsement is doing two things simultaneously: it is communicating information (the endorser's private assessment of the candidate's quality) and it is solving a coordination problem (telling other potential candidates and donors which candidate the party network is converging around).

21.2 Campaign Infrastructure: Who Does What

A modern competitive federal campaign is a small-to-medium business that springs up in a year or two and dissolves on Election Day. The org chart is similar across parties and across cycles, with predictable variation by race size.

The campaign manager: the CEO

The campaign manager runs the operation. In a competitive House race, the manager is typically a paid professional with one or more prior cycles of experience, often someone who came up through party committee staff jobs. Their compensation in 2024–26 ranges from roughly $7,000 to $15,000 per month for a competitive House race, $15,000 to $30,000 per month for a Senate race, and substantially more for presidential principals.

The manager owns budget, staffing, and strategic decisions. The candidate sets vision and makes final calls on the biggest questions; the manager decides everything else, including most of the operational decisions that determine whether the vision is executed. A common saying among campaign professionals: "The candidate has a campaign. The campaign has a campaign. They are not the same campaign, and the campaign manager's job is to keep the gap closed."

Communications director, press secretary, comms team

The communications director shapes the campaign's public message and manages the press relationship. The press secretary, often reporting to the comms director on larger campaigns, handles day-to-day reporter inquiries. On a small House campaign, one person does both. On a Senate or presidential general, the comms shop has 5 to 30 people, including specialty hires for ethnic-media outreach, surrogate booking, video production, and social-media management.

The defining skill is rapid response. A typical Tuesday in October on a competitive Senate race: 60 reporter inquiries, four interview requests with deadlines that day, two opposition research dumps to evaluate, one debate-eve briefing book to finalize, and three campaign events to staff. The comms director who can keep the trains running while protecting the candidate's time is worth their compensation.

Field director / political director

Field directors run voter contact — the door-knocking, phone-banking, and on-the-ground volunteer operation. Political directors handle relationships with allied organizations, local elected officials, and constituency groups. On smaller campaigns, one person does both. On larger campaigns, the field director might supervise 5 to 20 regional field directors who each run an office, and the regional field directors might supervise 50 or more part-time canvassers and hundreds of volunteers.

The field operation is often the largest single line in the campaign budget after media, and it is consistently the part of the operation under the most pressure to demonstrate measurable impact. We will return to the empirical evidence on field operations in Section 21.6.

Finance director and finance committee

The finance director runs fundraising. Their job is to keep the candidate's call time productive, to staff finance events, to track donor relationships in the campaign's customer-relationship-management system (typically NGP for Democrats, Anedot or NationBuilder for Republicans, with significant variation), and to make sure FEC compliance is clean.

The finance committee is a separate body — typically 30 to 200 supporters who have committed to raising specific dollar targets from their own networks. Finance-committee membership is an important informal institution: it functions as a credentialing system among donors, and it produces relationships between the candidate and the most consequential donors that often persist long after the campaign.

Pollster, media consultant, digital consultant

Three roles are typically filled by outside vendors rather than campaign staff:

The pollster runs the campaign's internal polling — benchmark polls at the start, tracking polls in the final 60 days, message-testing polls before major ad releases. Major Democratic pollsters in the 2020s include Anzalone Liszt Grove Research, Garin-Hart-Yang, Global Strategy Group, and others; major Republican pollsters include Public Opinion Strategies, Tarrance Group, Fabrizio Lee, and others. A serious House race pollster contract runs $50,000 to $200,000 per cycle; Senate and presidential contracts run into the millions.

The media consultant produces and places the TV and radio ads that consume 50% to 70% of a competitive campaign's budget. The big media-consulting firms (Devine Mulvey Longabaugh on the D side; Strategic Media Services and OnMessage Inc. on the R side, among others) take a commission on ad placements, traditionally 15% but increasingly negotiated lower as digital ad placement has become a separate specialty.

The digital consultant handles online ad placement, email fundraising, SMS, social media buys, and increasingly the campaign website and CRM integration. Major Democratic digital firms include Bully Pulpit Interactive and Mothership Strategies; Republican firms include Targeted Victory and Push Digital. Digital consulting is the fastest-growing line in campaign budgets, roughly tripling as a share of total spending between 2012 and 2024 (Wesleyan Media Project tracking).

Opposition researcher

Opposition research — "oppo" — is usually handled by outside firms that build deep dossiers on the opposing candidate and on potential surrogates. The work includes voting-record analysis, public-records pulls, deep dives into prior business dealings, social-media archives, and interviews with former colleagues and adversaries. Major firms include American Bridge 21st Century (D-aligned) and America Rising (R-aligned), plus boutique shops on both sides.

A typical "book" on a competitive opponent runs 200 to 500 pages of footnoted findings, indexed by topic, kept in a secure document-management system. The candidate's own campaign also commissions a "self-research" book on its own candidate — to identify vulnerabilities before the opponent does and to prepare responses.

Staff size by race

Rough staff sizes for the 2024–26 cycles:

• A competitive House race: 10 to 50 paid staff, peaking in the final 60 days.
• A competitive Senate race in a medium state: 80 to 200 paid staff.
• A competitive Senate race in a large state (Texas, Florida, Pennsylvania): 150 to 400 paid staff.
• A presidential primary campaign at peak: 300 to 800 paid staff.
• A presidential general-election campaign: 1,000 to 4,000 paid staff at peak, plus tens of thousands of volunteers and an integrated operation with the national party committee.

These are paid staff. Volunteer counts are an order of magnitude higher.

A budget snapshot

The composition of a competitive 2024 House campaign budget, drawn from FEC quarterly filings of several races and reconciled against published consulting-firm estimates, looks roughly like this:

The categories are not perfectly clean — direct mail can be classified as paid media or as field, depending on how the campaign codes it — but the dominance of paid media is the most consistent finding across competitive races. A campaign that allocates 30% to field and 35% to paid media, like several insurgent progressive primary campaigns in 2018–2020, looks structurally different from a campaign that allocates 60% to paid media and 8% to field.

Who actually runs the operation

A useful exercise for understanding modern campaigns is to imagine the daily flow of decisions a campaign manager makes between 7 AM and 11 PM. Some examples drawn from interviews with campaign professionals:

• 7:15 AM: Read overnight tracking poll. Brief the candidate at 8 if there is a meaningful movement.
• 9:00 AM: Approve the day's earned-media plan. Decide whether to push back on a story that ran in a regional newspaper.
• 10:30 AM: Approve a digital ad creative for placement that afternoon. Resolve a dispute between the digital consultant and the comms director about tone.
• 12:00 PM: Lunch with a major donor whom the candidate cannot meet today.
• 2:00 PM: Sign off on a direct-mail piece. Sign off on the script for a phone-bank shift starting that evening.
• 4:00 PM: Resolve a personnel issue between the field director and a regional field director who has been missing his contact goals.
• 6:00 PM: Attend the campaign's evening event with the candidate. Manage the press list and the surrogate roster.
• 9:00 PM: Review the day's fundraising totals. Compare to plan. Decide whether to add a call-time block to the candidate's schedule for the next morning.
• 10:30 PM: One-hour debrief with the senior team about the day. Identify three priorities for tomorrow.

A campaign manager who is doing the job well is making 80 to 200 small decisions per day, in addition to the large strategic ones. The candidate is making fewer decisions than the manager but is making them under more public scrutiny.

21.3 The Fundraising Operation

For most candidates, fundraising consumes more time than any other single activity. The numbers are blunt enough that they shock students hearing them for the first time.

Call time

"Call time" is the campaign term for blocks of time during which the candidate sits in a quiet room with a list of prospective donors and a phone, and dials. In a competitive race, candidates spend 30% to 50% of their working hours on call time — three to five hours per day, four or five days per week, for 12 to 18 months.

Tom Allen, a six-term Democrat from Maine who ran for Senate in 2008, wrote bluntly that fundraising "consumed me." Steve Israel published a piece in The New York Times Magazine in 2016 describing his own routine as DCCC chair: "I would sit in a small windowless cubicle in the basement of the DCCC, looking at a script with the names of donors I'd spoken to dozens of times, the amount they'd given, the amount we wanted them to give, and the children's names of those donors so I could ask after them. I cannot remember a single legislative meeting I had with a non-donor constituent during those years that lasted as long as the average call." Tim Roemer, a former Indiana Democrat, gave similar testimony in his interviews with the Brennan Center. Republican members in private interviews describe similar routines.

The cumulative effect of this allocation of time is that members of Congress spend a substantial share of their working hours interacting with the small slice of the American public — fewer than 1% — that gives money to political campaigns. Whether this translates into policy bias is contested, and we will examine the evidence in Chapter 34. The behavioral fact is not contested.

Donor cultivation

Major donors — those who give the maximum allowed by federal law, currently $3,500 per election to a candidate, plus the maximum to a joint fundraising committee — are cultivated relationally over months or years. The cultivation process is not transactional in any single moment. A finance director might bring a new prospect into the candidate's orbit at a small dinner, fly the candidate to the prospect's city for a private meeting, invite the prospect to a campaign retreat, and only six months later ask for a major check. Repeat donors expect calls of thanks, holiday cards, and access to the candidate during legislative debates of interest to them. Major donors often form personal friendships with the candidates they support. These relationships sometimes persist for decades, across multiple campaigns, and into the donor's role as a board member of allied nonprofits.

Online fundraising: ActBlue and WinRed

The biggest change in campaign finance since 2008 has been the rise of online small-dollar fundraising. Two platforms dominate.

ActBlue launched in 2004 as a Democratic-aligned 501(c)(4) that processes online contributions for Democratic candidates and progressive organizations. By the 2024 cycle, ActBlue processed roughly $4.1 billion in contributions across the cycle, with the median contribution around $25 (ActBlue annual transparency reports; FEC data).

WinRed launched in 2019 as the Republican-aligned counterpart, jointly endorsed by the RNC, NRSC, and NRCC. WinRed processed roughly $2.3 billion in 2024, with similar median-contribution patterns (WinRed reports; FEC data).

The platforms changed the economics of running for federal office. A candidate who can produce a viral fundraising moment — a debate clip, a moment of confrontation with the opposing party, a tragic news event affecting the candidate's district — can raise hundreds of thousands of dollars within hours, from tens of thousands of donors, most contributing $5 to $50. This has shifted the incentive structure: candidates who can credibly produce such moments are advantaged in the modern fundraising environment, regardless of seniority or institutional backing. The implications for both parties' candidate pools are still being debated.

Joint fundraising committees and federal limits

Federal contribution limits for the 2025–26 cycle (FEC, indexed for inflation):

• To a federal candidate: $3,500 per election (primary, runoff, and general count separately, so a maximum of $7,000 to $10,500 to a single candidate across the cycle).
• To a national party committee: $44,300 per year.
• To a state, district, or local party committee: $10,000 per year (per committee), with sub-limits for separate accounts.
• To a multi-candidate PAC: $5,000 per year.

Joint fundraising committees (JFCs) allow donors to write a single large check that is then split among multiple committees, each receiving up to the maximum allowed by law for that committee. A donor can legally give a single JFC check of more than $800,000, with the proceeds split across the candidate's campaign committee, the national party, several state party committees, and allied PACs. The Hillary Victory Fund (2016) and Trump Victory (2016, 2020, 2024) are well-known examples on each side.

JFCs are legal under the McCutcheon v. FEC (2014) line of decisions, which struck down aggregate contribution limits. They concentrate fundraising attention on the small number of donors capable of writing six- and seven-figure checks, while small-dollar online fundraising democratizes the donor base in the opposite direction. Both trends are happening simultaneously, and the modern campaign uses both.

A useful illustration of how the math works: suppose a presidential nominee, the national party committee, and 35 state party committees form a joint fundraising committee. A donor giving the maximum allowed to each can write a single JFC check that splits as follows: $6,600 to the candidate's primary and general, $44,300 to the national committee, and $10,000 to each of 35 state parties — a total exceeding $400,000 from a single donor in a single check. After 2014, the McCutcheon decision allowed an individual donor to give the maximum to as many committees as they wish in a cycle, opening the door for these large JFC checks. Critics call the practice a workaround of contribution limits; defenders argue it streamlines compliance for donors who would otherwise write 37 separate checks. Both observations are descriptively true.

PACs, super PACs, and dark money — brief preview

A standard political action committee (PAC) is a registered committee that raises money to support candidates and parties, subject to federal contribution limits ($5,000 from any individual per year). A super PAC, authorized by the post-Citizens United (2010) and SpeechNow.org v. FEC (2010) decisions, can accept unlimited contributions and make unlimited independent expenditures, but cannot coordinate with candidates' campaigns. "Dark money" refers to political spending by 501(c)(4) social-welfare organizations and certain other entities that are not required to disclose donors. Chapter 34 covers all three in detail.

Bundlers

A bundler is a donor who solicits checks from others and bundles them on behalf of a candidate. Federal law requires disclosure of bundlers only in certain limited circumstances (lobbyist bundlers, per the Honest Leadership and Open Government Act of 2007). The Obama campaign in 2008 voluntarily disclosed all bundlers raising $50,000 or more, creating what became known as the "Obama bundler list." Subsequent campaigns of both parties have selectively disclosed bundlers; the practice has receded in the post-2016 era but remains an informal credentialing system among major donors.

21.4 Voter Contact: How Campaigns Reach Voters

A modern federal campaign uses five primary channels of voter contact. Each has a distinctive cost structure, persuasive effect, and operational requirement.

Door-to-door canvassing

A canvasser walks a list of pre-identified addresses, knocks on the door, and either has a brief conversation with the voter (if they are home) or leaves campaign literature (if they are not). Canvassers can be paid (typically $15–$25 per hour for part-time work in 2024–26 prevailing rates) or volunteer.

The empirical evidence on canvassing comes from a long series of randomized field experiments led by Donald Green, Alan Gerber, and others starting in the late 1990s. The summary finding (consolidated in Green and Gerber's Get Out the Vote, now in its 4th edition, 2019): in-person canvassing produces a statistically detectable but modest increase in turnout among contacted voters — roughly 7 to 14 percentage points among those reached, but only 1 to 3 percentage points across all addresses on a list, because most addresses do not produce a contact.

A more recent and more disorienting finding comes from David Broockman and Joshua Kalla's research (Kalla and Broockman 2018, American Political Science Review). Using 49 field experiments, they found that in-person door-knocking, phone calls, and other voter-contact tactics produce essentially zero average effect on candidate choice during the heat of a general election. The "minimal effects" finding does not say campaigns do not matter — they argue campaigns matter at the margin, and contact still affects turnout — but it does say that in the late stages of a campaign, when both sides are contacting voters intensively and the airwaves are saturated, marginal contact rarely changes minds.

Broockman and Kalla have separately documented one technique that does change minds: "deep canvassing," in which a canvasser engages a voter in a 10-to-20-minute conversation that draws out the voter's own values, listens, and offers a reframing of the policy issue. Their 2016 paper in Science, on a deep-canvassing intervention to reduce transphobia, found durable effects three months later — a result almost unprecedented in the persuasion literature. Subsequent replications have found the effect to be real but harder to scale than the initial paper suggested. Deep canvassing is now used by some advocacy organizations and a few candidate campaigns, but it is operationally expensive (each conversation takes a long time, and the canvasser training is intensive).

Phone banking

Phone banking ranges from the most basic (volunteers calling from a list and reading a script) to the most sophisticated (predictive dialers that connect a caller only after a live person picks up, integrated with the campaign's voter file). Most modern phone banking is volunteer-staffed; paid phone banking has moved largely to peer-to-peer texting because of cost and because younger voters answer phones less frequently than older voters.

Phone-bank effects on turnout are smaller than in-person canvassing effects but cheaper per contact. The marginal cost-effectiveness depends on the local labor market and the price of digital alternatives.

Texting

Peer-to-peer (P2P) texting platforms became dominant after 2018. A volunteer or paid staffer using a P2P platform sends a text from a personal-looking number, with the message addressed to the voter by name. If the voter responds, the staffer can have a real conversation. Under the federal Telephone Consumer Protection Act (TCPA) and FCC implementation, automated dialers ("autodialers") are restricted, but messages sent by a human operator pressing send are permitted. The legal landscape has shifted with FCC rulings and Facebook v. Duguid (2021), which narrowed the definition of an autodialer.

Texts have higher open rates than emails and are cheaper than phone calls. Their persuasive effect is modest, but they are highly effective for tactical communications: reminding registered voters where to vote, confirming volunteer shifts, and chasing absentee-ballot returns.

Direct mail

Despite the rise of digital, direct mail remains effective for older voters and for highly targeted persuasion attempts. Voter-file analytics allow campaigns to send specific pieces of mail to specific voter profiles — a piece on Social Security to seniors, a piece on student debt to recent college graduates, a piece on agricultural policy to farm-county voters. A typical direct-mail piece costs $0.50 to $1.00 per household to design, print, and deliver. Effects on turnout and candidate choice are real but small.

Digital ads

Digital ads — on Meta platforms (Facebook, Instagram), Google's properties (YouTube, Search, Display), and emerging platforms (TikTok, with restrictions) — allow targeting by demographics, geography, voter-file matches, and behavioral signals. Cookie-based targeting has narrowed since 2020 because of platform-level privacy changes (Apple's App Tracking Transparency, Google's Privacy Sandbox), but voter-file matching to platform user accounts remains an effective targeting channel.

Digital ad spending grew from roughly 1% of total campaign spending in 2008 to roughly 25% in 2024 (Wesleyan Media Project, Borrell Associates). The shift is more pronounced in primaries (where small audiences and narrow demographic targets favor digital) than in general elections (where TV's reach advantage among older voters keeps it dominant in spending share).

Channel allocation in practice

Modern voter-contact plans usually allocate channels based on the segment of voter being reached. A simplified version of how a 2024 House campaign might think about a 200,000-voter "persuasion universe":

• Top 30,000 (highest persuadability score, regular voters): Multi-touch program — at least two pieces of mail, one canvasser visit (or two attempts), one phone call, three to six digital ad impressions. Estimated cost per voter: $8–$15.
• Next 70,000 (medium persuadability, regular voters): Two pieces of mail, digital ads, possibly a phone call. Estimated cost per voter: $3–$6.
• Next 100,000 (lower persuadability or irregular voters): Digital ads only, with light direct mail to specific subgroups. Estimated cost per voter: $1–$2.

The dollar gap between high-priority and low-priority voters is wide on purpose. A persuadable, high-turnout voter receives 10 to 30 times the contact a low-priority voter receives. Whether this is the optimal allocation has been debated since the late 1990s, with the empirical literature generally supporting concentration of resources but the exact thresholds disputed. The campaign that "pulls in" — concentrates resources on a narrower universe — and the campaign that "pushes out" — spreads resources across a broader universe — represent different theories of how marginal voters move, and neither has decisively won the argument.

TCPA and digital regulation

The Telephone Consumer Protection Act (TCPA) of 1991, the CAN-SPAM Act of 2003, and various FCC regulations create a complicated legal regime governing campaign communications. Robocalls without recipient consent are generally prohibited; recorded political messages can be delivered to landlines but not to cell phones. SMS messages sent by automated dialers require prior consent; messages sent by a human operator pressing send do not. Email solicitations require an unsubscribe option. Digital ads on social-media platforms must include disclosure of who paid for the ad — though enforcement has been uneven and platform-by-platform.

The Facebook v. Duguid (2021) decision narrowed the definition of "automatic telephone dialing system" under TCPA, which had the effect of expanding the legality of certain peer-to-peer texting platforms. State-level laws sometimes layer additional restrictions on top of federal rules. Modern campaigns retain compliance counsel to navigate this regime, and FEC and FCC complaints filed against campaigns are routine.

21.5 Voter Data and Analytics

Behind every voter-contact decision is a database. Modern campaigns are data operations, and the data infrastructure has consolidated around a small number of vendors.

Voter files

Each state's secretary of state (or equivalent state office) maintains a registered-voter list. The information available varies by state. Some states record party affiliation; others do not. Most record vote history (whether you voted in each recent election, but not how you voted). Address, registration date, and date of birth are usually included. These files are available, for a fee or for free depending on the state, to candidates, parties, and certain other purchasers.

Vendor enhancement

Vendors take state voter files, append commercial data (homeownership, magazine subscriptions, vehicle registrations, charitable-giving histories), and merge across states to produce national files. The dominant vendors:

• NGP VAN is the Democratic-aligned voter-file infrastructure. Every Democratic candidate of any size uses VAN at some level, integrated with NGP's fundraising and CRM tools. VAN was built largely on the Catalist data infrastructure, which itself was a post-2004 Democratic project to build a permanent voter-data asset (Issenberg, The Victory Lab, 2012).
• Data Trust and i360 are the Republican-aligned counterparts. Data Trust is owned by an RNC-aligned trust; i360 is associated with the Koch network.
• TargetSmart (D-aligned) and L2 (cross-partisan) are independent commercial vendors.

Modeling

On top of the enhanced voter file, campaigns layer statistical models. A typical 2024 model set assigns each voter several scores:

• Turnout score: the predicted probability that the voter casts a ballot in this election (0 to 100).
• Partisan score: conditional on voting, the predicted probability the voter chooses the candidate's party (0 to 100).
• Persuadability score: the predicted probability that the voter is persuadable (likely to vote, but not committed to either side).
• Issue scores: the predicted importance of specific issues to the voter (immigration salience, abortion salience, climate salience, etc.).

Campaign managers allocate contact resources based on these scores. "Persuasion universe" = high turnout score + middle partisan score + high persuadability score. "Mobilization universe" = high partisan score + middling turnout score (you need to get this person to the polls). "Save-it-for-Election-Day universe" = high partisan score + high turnout score (don't waste contact resources on voters who are already with you and going to vote anyway).

GOTV: Get Out The Vote

The final 72 to 96 hours of a campaign run on GOTV machinery: door-knocks to mobilization-universe voters, calls and texts confirming voting plans, transportation services for voters who lack a way to the polls, and rapid responses to voting-place issues (long lines, machine breakdowns) reported by poll observers. The pace is brutal. Campaign offices that have been running steadily at 60 hours per week shift to 90 hours per week.

Modeling pitfalls and the limits of analytics

Modern modeling is impressive but bounded. Two systematic limits matter for understanding political outcomes:

Calibration drift. A model trained on prior cycles' data assumes voters in this cycle behave like voters in prior cycles. When the cycle is unusual — a particularly polarized presidential race, a new issue salience, a candidate who reorganizes the existing coalitions — the model's predictions decay. The Democratic models that performed well in 2008 and 2012 underperformed in 2016 in part because the working-class white voters in the upper Midwest who were predicted to be reliable Democratic voters did not behave that way. Republican models in 2018 and 2020 had similar surprises with college-educated suburban voters in some states. No model fully recovers from a sudden coalition shift; the campaign discovers the shift in the polling data and adjusts the targeting, but always with a lag.

Coverage gaps. Voter files cover registered voters reasonably well in most states, but some constituencies — recently moved voters, voters in states with same-day registration, voters who have changed their registration since the file's last refresh — are imperfectly represented. A campaign that mobilizes only the people on its file misses everyone the file does not capture. Same-day registration states (Maine, Minnesota, Wisconsin, and others) are particular cases where the GOTV operation has to run partly off the file and partly off real-time intake on Election Day.

Privacy and ethics. As voter modeling has become more sophisticated, questions about the appropriate use of personal data have intensified. The Cambridge Analytica revelations of 2018 — in which a firm associated with the 2016 Trump campaign acquired Facebook user data without proper consent — produced a wave of regulatory and platform-level responses, including Meta's restriction of certain audience-targeting tools for political advertisers. The European Union's GDPR (which applies to data on European users) has indirectly affected U.S. campaign data practices. Modern campaigns generally adhere to the platform-level policies, but the underlying question — how much personal data should be available to political operations, and under what consent regime — remains unresolved at the federal level in the United States.

21.6 Field Organizing and the "Obama 2008 Model"

The 2008 Obama campaign produced a field operation that became, in subsequent years, both genuinely admired and partly mythologized. David Plouffe, the campaign manager, described the model in The Audacity to Win (2009). Jim Messina, manager of the 2012 reelection, refined and scaled it. Sasha Issenberg, in The Victory Lab (2012), embedded with the operation and produced what remains the best journalistic account.

The model in summary: many small field offices opened months before the election, staffed by paid organizers in their twenties; each organizer responsible for recruiting and managing a network of volunteer "neighborhood team leaders" who in turn recruited their own teams; relentless attention to data — every door-knock, phone call, and contact recorded in VAN; daily reports flowing from the field offices to state-level field directors to the national team in Chicago; a culture of measurable goals and accountability for hitting them.

The empirical question is how much marginal vote-share the field operation produced. Here the literature is more skeptical than the journalistic mythology. Several political scientists who studied the 2008 and 2012 campaigns concluded that field operations of the Obama scale produced perhaps 1 to 3 percentage points of marginal turnout among contacted voters — meaningful at the margin in a close race, but not transformational, and below what the campaign's own estimates suggested. Field operations remain valuable, but the post-2008 idea that a sufficiently massive field operation could overwhelm a fundraising or message disadvantage has not survived empirical scrutiny.

The "ground game" (field) versus "air war" (paid media) tradeoff is a real campaign-management question. Different campaigns and different consultants have different views on the optimal mix. As of the 2024 cycle, both parties run sophisticated field operations, and the partisan asymmetry that briefly favored Democrats has narrowed substantially.

What field actually does

A useful frame for understanding the marginal value of field is to distinguish three things field operations do:

Mobilization. Field operations get partisans to the polls who would otherwise stay home. The empirical evidence (Green and Gerber, Get Out the Vote) supports a real but modest effect — a few percentage points among contacted voters, somewhat smaller across the targeted universe.

Persuasion. Field operations attempt to persuade undecided voters. The Kalla and Broockman 2018 finding suggests that during the heat of a campaign, in-person canvassing produces close to zero average effect on candidate choice. Outside of campaign season, with deep-canvassing techniques, the effect is larger but the operation is more expensive.

Visibility and morale. Field operations have the additional, harder-to-measure effect of producing visible campaign presence — yard signs, canvasser activity, local-media coverage of a campaign office opening — that signals competitiveness to other voters and to the press. Whether this produces measurable vote share is unclear; whether it produces measurable donor confidence and volunteer recruitment is more credible.

The realistic answer, based on the published evidence, is that field operations are worth investing in for mobilization purposes and for the visibility-and-morale effect, but campaigns should not expect them to produce outsized persuasion effects in a saturated media environment. A 2024 House campaign that opened ten field offices and ran a field operation through the cycle is doing something useful, but the marginal vote produced by the tenth field office is probably smaller than the marginal vote produced by the same dollars spent on targeted digital ads or direct mail.

Republican field investment since 2016

The narrowing of the partisan asymmetry in field operations is a notable development. Through the 2008 and 2012 cycles, Democratic field organization was substantially more developed than Republican field organization. By 2016, the Trump campaign relied heavily on the RNC's separate field operation rather than building an extensive campaign-side field operation. By 2024, the Republican Party had invested significantly in field infrastructure — including the Election Integrity Network, the RNC's Bank Your Vote early-voting outreach, and the integration of conservative grassroots organizations like Turning Point USA and faith-based outreach networks. These operations are different in tone and methodology from the Democratic model — relying more on existing community networks (churches, sportsmen's clubs, small-business associations) than on paid organizers building neighborhood teams from scratch — but they constitute a real field presence that the party did not have a decade ago.

21.7 Communications Strategy

A campaign's communications operation does five things simultaneously: it sets the message, places paid media, manages earned media, runs digital and social channels, and handles rapid response.

Message discipline

A modern campaign repeats two or three messages over and over until the candidate is sick of saying them and the staff is sick of writing them. This is by design. Voters who pay close attention to politics are a small minority; most voters absorb campaign messages in fragments, often through media coverage rather than directly. Repetition is what makes a message reach a voter who is paying minimal attention.

The discipline of message has costs. Candidates who deviate from the script — whether because they get bored, because they want to engage on a different issue, or because they think they are being clever — tend to underperform. The campaign professionals' frequent complaint about candidates is that they want to talk about things their pollster knows are not winning issues for them.

Earned media

"Earned media" is news coverage the campaign does not pay for. Press conferences, debate moments, surrogate appearances, and viral social-media moments all count. The challenge in the modern environment is breaking through. Local television news has shrunk dramatically since 2008, and national cable news consumes a small share of voter attention compared to social media and short-form video. A 2024 campaign that gets a 30-second clip into the rotation on local TV news and Facebook is reaching a different audience than a campaign that gets the same 30 seconds onto TikTok.

Social media

Each platform reaches different demographics. TikTok skews younger; Facebook skews older; X (formerly Twitter) is heavily skewed toward journalists and political junkies, who then translate what they see there into broader media coverage. Instagram reaches a different demographic still. A modern campaign maintains a presence on all major platforms, with a small team of platform-native producers. Increasingly, video-first content (TikTok, Reels, YouTube Shorts) is consuming more of campaign digital attention than text or image posts.

Rapid response

The 24-to-48-hour news cycle requires immediate answer to attacks, breaking news, and unforced errors. A modern rapid-response shop has staffers monitoring news 16 to 20 hours per day in the final 60 days, with pre-written response packages for predictable lines of attack and the authority to push out short-form responses (a tweet, a press release, a clipped video) within minutes.

A canonical case study in rapid response is the Howard Dean "scream" of 2004, which went from a private moment in a concession speech to a saturating cable-news story within 36 hours. By 2008, campaigns had built infrastructure to manage similar moments — pre-staffed war rooms, designated surrogates ready to defend or contextualize a candidate's comment, and digital teams to push counter-content into social-media feeds. By 2024, the rapid-response timeline had compressed further: a candidate's debate misstatement can produce a 30-second TikTok clip viewed five million times within hours, before the campaign has a chance to brief surrogates or respond. The pace creates tension between accuracy (taking time to verify before responding) and speed (responding before the story sets in the media bloodstream). Rapid-response shops resolve the tension differently in different campaigns; some err on the side of fast and cleaning up later, others err on the side of careful and accepting that the first 12 hours will be unfavorable.

Surrogate operations

Surrogates — politicians, advocates, and public figures willing to speak on behalf of the candidate — are an undervalued part of the modern communications operation. A surrogate program assigns specific surrogates to specific issues (a former military officer for veterans' issues, a small-business owner for economic policy, a former federal prosecutor for law-enforcement issues), books them onto television and radio appearances, and provides them with talking points calibrated to the campaign's message of the day. The candidate's time is finite; the surrogate operation extends the campaign's voice into media slots the candidate cannot fill personally. A well-run presidential general-election operation might have 50 to 200 active surrogates, each with a defined role.

21.8 Debate Preparation

The preparation for a single major debate consumes weeks of campaign time and tens of thousands of dollars in staff and travel.

Murder boards and mock debates

A "murder board" is a session in which campaign staff, allied operatives, and outside experts simulate every possible attack on the candidate. The candidate sits at a podium, the team peppers them with hostile questions, and the candidate works through their answers — repeatedly, until the answers are tight, on-message, and short enough to fit a debate's response window.

A mock debate is a more elaborate simulation. A surrogate plays the opposing candidate, sometimes after weeks of researching the opponent's debate style. The mock debates are scored, and the candidate works through their performance with the prep team. Major presidential mock debates have featured high-profile stand-ins: Robert Reich played Walter Mondale's debate-prep partner; Rob Portman has played Democratic opponents in prep for multiple Republican nominees; Philippe Reines played Trump for Hillary Clinton in 2016 (and the photographs of Reines in costume became a small cultural moment).

The contemporary debate landscape

Primary debates have evolved substantially since 2016, when Donald Trump's stage presence on the Republican primary stage broke many of the unspoken conventions about candidate behavior and forced both parties to reconsider the format. The COVID era (2020) introduced experiments with virtual or hybrid formats, then was abandoned. The 2024 general-election cycle saw the Biden-Trump CNN debate in June (no audience, muted microphones) — itself a departure from the Commission on Presidential Debates' traditional format.

The Commission on Presidential Debates, the bipartisan nonprofit that organized general-election debates from 1988 to 2020, was bypassed entirely in 2024 in favor of network-hosted debates, raising questions about whether the Commission will continue in future cycles. Down-ballot debates remain organized by various sponsors, with no single national authority.

21.9 Negative Campaigning

The empirical question is settled: negative ads work, especially comparative ads ("Senator X voted to do Y; my opponent supports the same thing"). Decades of research consolidated in the Wesleyan Media Project and earlier work by Lynn Vavreck, John Geer, and others find that negative ads change voter assessments of candidates more reliably than positive ads do.

The ethics question is contested. Defenders of negative campaigning argue that voters need to know the candidate's actual record, including the parts the candidate would prefer not to discuss; that calling out negative information about an opponent is a form of accountability journalism; and that a campaign that "stays positive" while ignoring an opponent's record is doing voters a disservice. Critics argue that negative campaigning depresses turnout (the "demobilization" hypothesis, contested in the political-science literature), degrades the political culture, and crowds out substantive policy discussion.

Tactically, modern negative campaigning uses three patterns:

• Framed contrasts. "I support X. My opponent voted against X. Voters can decide which approach is right for our state." This frames the negative claim as a comparison rather than an attack, and it is usually well within the bounds of accepted political speech.
• Direct attack ads. "My opponent did Y, which is wrong." These are the classic negative ads. They work, but they carry the risk of "backfire" — the candidate appearing too negative — and most campaigns deploy them through the campaign's own committee only when the attack is fully verified.
• Third-party hits. Outside groups (super PACs, party committees, allied 501(c)(4)s) run the harshest attacks while the candidate remains "positive" in their own ads. This is the dominant pattern in the post-Citizens United era. The candidate publicly stays above the fray; the outside group does the work; the strategic effect is the same as if the candidate had run the attack ads themselves.

Steel-manning the disagreement

The argument over negative campaigning is not a partisan argument — both parties run negative ads, both parties' candidates have at various times campaigned positively and at other times campaigned aggressively against opponents — but it is a real argument. The strongest version of each side:

The defense of negative campaigning rests on the claim that voters need to know the negative information, that "going negative" is in many cases simply describing the opponent's actual record, that voters reward candidates who show willingness to fight rather than candidates who appear meek, and that the alternative — a campaign in which neither side discusses the other's record — would be worse for democratic accountability. On this view, the political-science finding that negative ads work is a reflection of voters' real desire for information, not a corruption of the system.

The critique of negative campaigning rests on the claim that negative ads exaggerate or selectively quote the opponent's record, that the demobilization effect (whether or not it is decisive) discourages exactly the marginal voters democracy most needs to engage, that the saturation of negative messaging crowds out substantive policy discussion, and that the long-term effect on political culture — voters cynical about all politicians, candidates incentivized to define themselves by what they oppose rather than what they propose — is corrosive. On this view, negative ads "work" in a narrow sense (they shift vote shares) but harm the underlying system.

The empirical literature does not resolve the normative question. Both sides describe real phenomena. A reader can hold the empirical finding (negative ads are effective) and reach either normative conclusion (and thus they should be used; or, and thus regulators should constrain them).

21.10 The Endgame: Final 72 Hours and Election Night

The last 72 hours of a campaign are a different operation from the prior 12 months. The persuasion phase is over; the mobilization phase is everything.

GOTV operations

Every minute is allocated. Volunteers are stationed at high-traffic intersections holding signs. Phone banks run from 6 AM to 9 PM. Door-knockers hit only the highest-priority addresses on the mobilization-universe list — voters who are likely supporters but not certain to vote. Text-message reminders to absentee-ballot returners and to voters who haven't yet checked in at their polling place. Rides to the polls dispatched on demand.

Election Day legal teams

Modern campaigns also field legal teams on Election Day. Lawyers and trained observers are stationed in counting centers and polling places to monitor for irregularities, file emergency motions if voting hours need to be extended due to machine breakdowns, and document any disputes that may matter for post-election challenges. The legal teams have grown larger and more professionalized in every cycle since 2000, with particular escalation since 2020.

Election Night war rooms

The candidate, key staff, and major donors gather at a single location on Election Night. A separate war room — usually at campaign headquarters or a hotel — runs the operation: monitoring vote returns by precinct, comparing to internal models, drafting both victory and concession remarks, fielding calls from media outlets requesting reactions, and managing the campaign's social-media output as results come in.

The decision to declare victory or to concede is one of the most consequential made by a campaign manager and the candidate, and it is made under pressure with incomplete information. The traditional norm — wait until the opposing candidate concedes or the result is mathematically certain — was honored in nearly every federal race for decades. The post-2020 environment has seen more candidates declining or delaying concession, and campaign managers now plan for both scenarios.

A typical night

A useful exercise: imagine the staffing of a competitive Senate race's election-night war room. The room contains, typically, the campaign manager, the political director, the data director and a deputy, the deputy campaign manager handling press, two members of the legal team, the finance director (whose evening starts with a panicked review of whether all bills are paid through Election Day), and a junior staffer whose job is to make sure the coffee never runs out. A small kitchen-table-sized board on the wall lists each county in the state and is updated by hand every 30 minutes as returns come in. The data director runs a parallel digital model that projects the eventual result based on partial returns; the model gets continuously more confident as more precincts report. By 11 PM in most races, the model has converged enough to support either a victory or a concession decision; in close races, the model is still uncertain at midnight, and the team waits.

A campaign that has been preparing for both outcomes — and serious campaigns prepare for both — has two complete remarks ready to go: a victory speech that the speechwriter has revised seven times in the previous three days, and a concession speech that the same speechwriter wrote on a separate computer to keep the team that doesn't want to think about losing from having to think about losing. When the call is made — by the candidate, in consultation with the manager and political director — the relevant draft is handed to the candidate, the room exhales, and the public-facing stage of the night begins.

21.11 Post-Campaign

After Election Day, the campaign winds down rapidly.

Debt retirement

A winning campaign ends with cash in the bank that can be transferred to a member's reelection committee, used for transition expenses, or donated to other campaigns and party committees. Winning campaigns sometimes raise additional money post-election to pay down outstanding debts (vendor invoices, deferred staff compensation) and to build a war chest for the next cycle.

A losing campaign sometimes ends with debts that take years to retire. Hillary Clinton's 2008 primary campaign carried debt for years. The Trump 2016 primary campaign likewise had outstanding obligations to vendors. FEC rules require ongoing reporting until the committee is formally terminated, and various organizations track the multi-year resolution of campaign debts.

The political-consulting pipeline

Most campaign staff exit the cycle and move on. Many move into political consulting, either by joining one of the firms that supplied vendor services to their cycle's campaigns or by starting their own. Others move into government — congressional staff jobs, executive-branch appointments, party committee staff. Others leave politics entirely; the burnout rate is real.

The relationships persist. The staff of a single competitive Senate campaign typically produces, over the next two decades, several future candidates, dozens of senior congressional and administration staffers, multiple lobbyists, and a network of political and policy professionals who continue to do business with each other long after the campaign's specific candidate has retired. This network — the "campaign veterans" pipeline — is one of the underappreciated mechanisms by which professional political communities reproduce themselves.

21.12 Why This Matters

A student finishing this chapter should be able to do something most American Government students cannot: walk into any campaign field office in October of an election year, see the people moving around with clipboards and phones and laptops, and have a working hypothesis about who each of them is, what they are doing, and why. That is not a trivial thing to be able to do. The mechanics of campaigns shape political outcomes — they shape which candidates win, which constituencies are mobilized, which messages reach voters, and which complaints reach members of Congress in their next term.

The mechanics are bipartisan. Most consultants take work from one party but the techniques cross over freely. The voter-file vendors, the fundraising platforms, the field organizing models, the digital ad platforms — all of them are used by both parties, with predictable variation in tone and target. The partisan asymmetries that briefly opened in 2008–2012 have largely closed. As of the 2024 cycle, both parties run sophisticated data, fundraising, and field operations.

Whether one finds this state of affairs reassuring (campaigns are professional, accountable, and increasingly evidence-based) or alarming (campaigns are dominated by professionals and donors at the expense of unmediated democratic participation) is a normative question. The empirical fact — that this is how modern federal campaigns are run — is what this chapter has tried to convey. The next chapter turns from the campaign to the voter, and asks what the voter is actually doing when they decide.

Sources for the empirical claims in this chapter, including FEC contribution-limit data, Kalla and Broockman field-experiment results, Wesleyan Media Project ad-spending tracking, ActBlue and WinRed annual reports, Plouffe and Issenberg accounts of 2008–2012 Democratic operations, and Parscale/RNC accounts of 2016–2024 Republican operations, are documented in further-reading.md for this chapter.