Identifying voter turnout to drive your campaign or GOTV efforts

What the service does

Usually you give VoteCheck a list of people. VoteCheck matches it against the public record of who has participated in a given election and reports back: how many of them have voted, how that is moving day to day, how it compares to the county and the region around you, who they are demographically, and where the nearest polling places are.

You do not have to bring a list. If you are working a district rather than a roster, we can build the universe from the voter file itself — by district, precinct, city or turf — and track its turnout the same way, day by day. That is the same public record every campaign already works from; you are simply not the one assembling it.

It is a turnout measurement and civic engagement tool. It is built so that a school district, a hospital system, a union local, a membership association or a campaign can run a real get‑out‑the‑vote program across everyone on its list and actually know whether it worked. What comes back depends on which of the two doors you came in — that is the next section.

Two doors

There are two ways in, and they are not interchangeable: you are either a campaign working a voter universe, or an employer measuring your own people. Different upload, different storage, different product.

The employer door has two settings. Every employer account starts anonymized. An employer that needs to know which of its people to follow up with can sign an affidavit and switch — but that is a deliberate second step, not a different product and not a different door.

One thing sits between the two doors: the voter record itself — public county data, improved over time by what customers on both sides contribute to it.

Why the default is anonymized. You hold the key that turns a row token back into a person, so a report split by whether someone voted is a targeting list with one step removed — which is why that split does not exist at all on the default setting, rather than merely being switched off. Lifting it is precisely what the affidavit is for: it moves that decision, and the responsibility for it, onto a named signer.

The first question everybody asks

It is worth being precise about why, because the answer is not that the information is secret. In Texas, whether a registered voter participated in an election is public record — the registrar reports it to the Secretary of State after every election. How they voted is not, and never leaves the ballot box.

So the difference is not confidentiality — it is leverage. A campaign looking at a voter file is looking at strangers it can only ask for a vote. An organization looking at its own roster, split into those who voted and those who did not, is holding a list of identified individuals over whom it already has power, describing an act of political participation.

Where that organization is an employer, Texas law treats acting on it with unusual severity: subjecting an employee to a loss of wages or another benefit of employment in retaliation for how they voted — or for declining to say — is a third‑degree felony. The same asymmetry, in milder form, holds for a union and its members, a district and its staff, or a congregation and its people.

Nobody buying this service intends that. But a capability that exists can be requested, exported, forwarded, logged, discovered and subpoenaed. The list would sit in your systems for years, and every future dispute with every person on it would have to be litigated against its existence. We remove the exposure by not building the capability — which is also why we can give you a great deal more than a service that kept it would dare to.

What we take, what stays yours, and what we refuse

Four fields identify a person well enough to match them against the voter file. You may send more, and if it tells us something about that person as a voter we will keep it and improve their record with it. What we have no use for is the shape of your organization — who reports to whom, who sits next to whom, what anyone is paid.

Contact details that arrive with a roster are kept on the voter’s record, tagged with your organization as their source, and used to make that record more complete. They are never used by us to reach your people, and the voter file cannot be searched by the organization that contributed to it — so no one can ask it who belongs to you. All outreach is yours, sent through your own channels, to everyone on your list.

How a roster becomes a measurement

This sequence runs once per upload, in order. Steps three and four are the ones that make everything downstream safe.

1. 01The file is read and checked Required fields present, prohibited fields rejected, and a hard ceiling on how many rows may fail before the whole upload is refused rather than silently half‑loaded.
2. 02Each person becomes a token An HMAC‑SHA256 of their name and birth year under a salt unique to your campaign, rendered as VC4F91A0C3E8B2. The same person re‑uploaded next month produces the same token, so nobody is ever counted twice.
3. 03You are given the only key A crosswalk of token ↔ name is generated for you to download and keep. It is the sole means of turning our output back into your people.
4. 04We destroy our copy The names are purged from our store and the crosswalk expires after seven days. From that point the correlation exists only on your side of the boundary.
5. 05Tokens are matched to the voter file Confident matches settle. Everything else is held open, never deleted — so a fuller roster uploaded later can resolve it without creating a duplicate person. On the first real roster we processed, 48.5% of rows were too ambiguous to settle on names alone; discarding them would have quietly overstated turnout.
6. 06Attributes are captured in aggregate For settled matches only: age band, sex, precinct, and party lean inferred from primary participation. Never written back against a name, because by now there is no name to write it against.

What you actually receive

Removing the individual list does not thin the product out. It is what lets us be generous with everything else.

The dashboard

Example · real figures from a live roster

The morning report

Example · real figures from a live roster, entity name shown

On a phone

Example · the same report at 414pt

Why the export has no status column

Because its complement is trivial. Hand someone a list of who voted and they can subtract it from the roster they already own to get the people who did not. A “non‑voter” list and a “voter” list are the same artefact viewed from either end, and either one turns an all‑staff reminder into targeted contact based on an individual's participation. So there is one list, it contains everyone, and it carries no status to filter on.

The thresholds that protect the individual

Aggregates can leak. A department of three, broken out by turnout, names people as surely as a list does. Every figure VoteCheck publishes passes a floor first.

The same discipline governs the wording of your daily report. When your group's turnout sits below the county baseline, the percentage is withheld from the copy we hand you and the message leads with a count or the deadline instead. Publishing a low participation rate to a group tells them that not voting is normal here, and that is a measurable depressant. The figure is still on your dashboard — it is your number. It is simply not the sentence we suggest you send.

Where this sits in the Texas Election Code

The design above was drawn against these provisions. This is a description of engineering decisions and the constraints they answer — it is not legal advice, and it does not substitute for your own counsel's reading.

There is a straightforward way to describe the boundary to your own board: VoteCheck reads public participation data and returns statistics. It does not return records about identified individuals, because it does not retain the identities required to make one.

What happens when you sign up

Five steps. Nothing runs and nothing is billed until you have seen the price and said yes.

1. 01You create the campaign and upload your roster Four fields per person: last name, first name, birth year and ZIP. You choose the election to track. The file is checked and matched on the spot, so the next step uses your real numbers rather than an estimate.
2. 02We price it and email you a proposal Priced off the roster you just uploaded, itemised, with the annual alternative shown if it would suit you better. Nothing has started and nothing has been billed.
3. 03You accept, and we invoice One link in that email. Accepting confirms the figures so we can raise an invoice — it does not take payment. Purchase orders accepted.
4. 04We approve the campaign A person here checks it and turns it on. Until then it is not visible to your administrators and no reports go anywhere.
5. 05Reports every morning at 7am Through early voting and Election Day, with a final wrap-up the morning after. Your dashboard is live throughout.

What it costs

Priced per cycle — from load, about two weeks before early voting opens, through the day after Election Day. The two doors are priced on different things, because they are buying different things: a campaign pays for the size of the universe it chooses, a private group pays for the size of the roster it already has.

Campaigns — per registered voter

You name the race — a city council seat, a school district, a state house or congressional district, a whole county — and the registered voters in that jurisdiction set the price. One rate, whatever the size: five cents a voter, per election cycle.

The count is the one the county certifies for that entity at each election, so it is a number you can check. Every cycle includes the universe build, the daily roster match through early voting, strike lists cut each morning, and the full voter detail for that universe. There is no separate onboarding fee for a campaign.

Annual — parties, PACs and anyone running more than one race

A county party works the March primary, the May runoff, the May uniform date and November. Paid by the cycle that is four separate universes and four separate invoices. The annual plan replaces both.

You name a footprint — the largest area you will work in the year, usually a county or a congressional district. For 2.5× that footprint's cycle price you get every race inside it, all year: primaries, runoffs, the May uniform date, November, and any special called in between. Sub‑universes cost nothing extra, so a county party can pull a precinct, a city, a school district or a state house district out of its footprint as often as it likes.

It pays for itself on the third race. Two races in a year and paying by the cycle is cheaper — we will say so rather than sell you the subscription. Runoffs are already included, not billed at half.

Estimate a campaign universe

Estimated. The figure that counts is the one in the proposal we send once the universe is defined — priced off the file itself, and nothing is billed until you accept it.

Private groups — per person on your roster

The band is set by how many people are on your roster.

Every cycle includes the roster upload and match, the daily morning report through early voting and Election Day, ready-to-send copy for everyone on your list, turnout broken down by unit, the final wrap-up, and email and phone support while the cycle runs.

Estimate a private group

Estimated. The figure that counts is the one in the proposal we email you once your roster is uploaded — that one is priced off the file itself, and nothing is billed until you accept it.

Add-ons

Runoffs for existing customers are billed at half the cycle price. Per-cycle fees are invoiced at roster load; annual plans up front or by semester. Purchase orders accepted.