Modern campaigns do not spread money evenly. They concentrate it on a small slice of voters who might actually change their minds, in a small number of places where the outcome could go either way. That is the core of how campaigns work: the decision about where to spend is a budgeting problem, not a messaging problem.
The logic is simple. A dollar spent persuading a voter who was already committed is a dollar wasted. So is a dollar spent on an opponent's strongest precinct. What is left is the persuadable middle, and campaigns hunt for it with voter files, canvass routes, and increasingly precise models.
This piece explains how that targeting happens, what door-knocking and phone work actually do, and where the math gets contested. It is a companion to our broader coverage of politics and to our explainer on what a mandate is and how to judge whether an election gave one.
What does targeting actually mean?
Targeting is the process of deciding which voters a campaign will contact, and how. The starting point is the voter file: a public-records-based list of registered voters, their party registration where the state records it, and their turnout history. Campaigns buy, merge, and enrich these files with commercially available data to build a picture of who lives where and who votes when.
From there, a campaign sorts voters into broad buckets. Base voters need reminding to vote. Opposed voters are usually left alone. The contested category is the persuadable middle, people with a mixed record or no strong record at all. Each bucket gets a different treatment, and each treatment costs a different amount per contact.
The sorting is imperfect. Files go stale, households change, and models mislabel people. Campaigns accept that error rate because the alternative, treating everyone identically, is more expensive per vote won.
Why do campaigns ignore most of the country?
Because most of the country is not in play. Where an election is decided by a wide margin, spending there buys little. Where the margin is narrow, each additional contact has a chance of moving the total. Campaigns therefore stack their budgets into the narrowest contests, which is why some states see constant advertising and others see almost none.
Readers who want the structural background can see our pieces on why control of Congress is the whole ballgame in 2026 and the 2026 gubernatorial map. The same concentration logic applies at every level, down to a state legislative district.
The concentration has a well-known cost: voters in uncompetitive places are largely ignored, and their sense of being taken for granted is a recurring criticism of the system. Campaigns know the criticism. The math pushes them anyway.
What does canvassing actually accomplish?
Canvassing is door-to-door or phone contact by campaign volunteers and staff. Its two jobs are persuasion and turnout. Persuasion conversations try to move a leaning voter. Turnout contact simply confirms a supporter will vote, or helps them make a plan to do so.
The mechanism matters more than the slogan. A short, respectful conversation that asks what a voter cares about, and connects it to one concrete issue, is the standard playbook. Scripts are tested and revised. Field directors track contact rates and conversation quality, not just doors knocked.
Canvassing is expensive in labor. That is why campaigns pair it with cheaper channels: mail, text, and digital advertising. The mix shifts as Election Day nears, with turnout work crowding out persuasion in the final stretch, because reminding a supporter to vote is more reliable than trying to convert a stranger at the last minute.
How does the persuasion math work?
Campaigns think in terms of cost per persuaded or cost per turnout, not cost per impression. If a channel reaches many people but moves few, its cost per persuaded vote is high. If a channel reaches fewer people but moves more of them, it wins budget even at a higher price per contact.
Two numbers drive the allocation. The first is the expected margin: how many votes the campaign believes it needs. The second is the estimated persuasion or turnout rate by channel and by voter segment. Multiply, compare, and allocate until the budget is spent where the marginal return is highest.
Our analysis: the honest description of this process is that it is a spreadsheet with error bars. Models are estimates, and their confidence intervals are wide. A campaign that treats its targeting model as fact, rather than as a best guess to be tested, tends to misallocate. The strongest campaigns re-test assumptions mid-cycle and move money when the evidence changes.
What are the main criticisms?
Three recur. First, privacy: the data enrichment behind targeting raises questions about how much campaigns should know. Second, distortion: concentrating on a few competitive areas gives those voters outsized attention and can skew which issues get emphasized nationally. Third, effectiveness: some analysts argue that mass persuasion is much harder than campaigns claim, and that most spending functions as turnout infrastructure rather than genuine conversion.
The counter-argument from practitioners is that even small persuasion effects matter when margins are thin, and that the alternative to modeled targeting is wasteful blanketing. Both points have merit, and the honest position is that targeting improves efficiency at the margin while its grander claims deserve skepticism.
This is also a media-literacy issue. If you want to understand why a particular ad reached you, our guide on how to read political news without getting spun covers the same skeptical habits. And for the deeper definitional background, the entry on Politics at Wikipedia describes politics as the activity of settling affairs in an organized society, including promoting political views among people, which is exactly what a campaign's targeting operation is built to do.
What should a voter take from this?
Three practical takeaways. First, if you live in a competitive area, you will be contacted heavily; if not, you may need to seek out the campaigns yourself. Second, being labeled persuadable or base is a guess about you, not a fact about you, and you can correct the record by how you engage or by updating your registration. Third, the loudest message is not the most effective one; campaigns pay for measured contact, not volume.
Understanding how campaigns work does not make the noise disappear. It makes it legible. Behind every ad break and door knock sits a budget line, a model, and a guess about you, and knowing that is the beginning of judging the message on its merits.
