What Is a Voter File?

Every serious campaign runs on a voter file. The yard signs, the door knocks, the mail pieces, the digital ads, the phone banks — all of it is allocated against a list of registered voters and what the campaign believes about each of them. If the file is wrong, every other line item is wasted.

Despite that, most first-time candidates — and a surprising number of veterans — could not tell you what is actually in the voter file their campaign relies on. This post fixes that.

The voter file is a public record, with structure

In Arizona, the Secretary of State maintains the statewide Voter Registration Database (VRDB). The current active file contains roughly 4.96 million records — one row per registered voter. Each record carries a defined set of fields:

• Identity: name, date of birth, residential address, mailing address.
• Registration status: active, inactive, cancelled, suspense.
• Party: Democratic, Republican, Libertarian, No Party Designated (PND, ~1.5M of the file), and a handful of small parties.
• Geographic precinct, ward, legislative district (LD), congressional district (CD), supervisor district, and other overlapping district memberships.
• Vote history: which elections this voter participated in, and through what method (early ballot, in-person, provisional).

What the file does not contain is also worth knowing: no phone numbers, no email addresses, no income, no race, no party support score, no turnout score. Those are derived — either by the campaign, by a vendor, or by appending a commercial dataset.

The file is structure. The intelligence is what you build on top.

The raw VRDB is a starting point, not a finished product. Useful campaigns extend it with derived columns:

• Propensity scores: the probability a voter turns out in this kind of election (presidential, midterm, off-year, special, primary).
• Persuasion scores: the probability a voter is persuadable on a given issue or candidate.
• Cohort tags: drop-off voters (registered, voted in 2024 general but not 2022 midterm), surge voters (newly registered since the last cycle), high-propensity ballot-chase universes for AEVL.
• Household structure: single-voter households, multi-voter households, mixed-party households.
• Geographic rollups: precinct-level turnout, LD-level partisan composition, congressional district profiles.

BlueList builds these layers on top of the Arizona VRDB and surfaces them as district-level intelligence — not as voter contact lists. We do not resell the voter file. We do not produce per-voter contact universes for clients. What we publish is aggregate analytics on districts, cohorts, and turnout patterns, derived from the same underlying record set, with the per-voter rows kept in our compute layer and never returned to clients.

Why “just buy a list” is the wrong question

New candidates often arrive asking for a list of likely-supportive voters in their district. That request is reasonable but premature. Before the list there are two questions a campaign has to answer first:

• What does the district actually look like? Party composition, age structure, turnout in comparable elections, the gap between registration and actual votes cast in the last two cycles. Without this, a list of names is just a list of names — you do not know whether to spend on persuasion, registration, or turnout.
• What can you afford to do with the list once you have it? Doors, mail, calls, and digital each have a different cost per contact and a different conversion rate. The right list size depends on the budget and the conversion math, not the other way around.

The District Intelligence Report exists to answer the first question. Any reputable consultant or coordinated-campaign program can help with the second. Both have to happen before the list matters.

What this means in practice

For a state legislative or down-ballot campaign in Arizona, the practical sequence is:

• 1. Pull the district profile. Understand party composition, turnout history, and the segment of voters who decide close races (typically PND, drop-off, and newly registered).
• 2. Decide the strategy. Persuasion-heavy if the district is genuinely competitive. Turnout-heavy if your party has a registration advantage but a turnout deficit. Registration-heavy if the underlying universe is too small to win without growth.
• 3. Then build the universes. The voter file segment that fits the strategy, sized to the budget, allocated across channels.

The campaigns that lose by 200 votes in an LD that broke the wrong way usually skipped step one.