How a local election campaign tracks voters without spreadsheets
Field data needs record-level concurrent writes, an audit trail, offline capture and role-scoped access. A shared spreadsheet provides none of the four. Keep the support scale to five points, and settle the legal and retention questions before collecting anything.
Every local campaign starts the same way. Someone exports the electoral roll — the official register of who is entitled to vote in each area — puts it in a spreadsheet, shares it with the team, and for about two weeks it works.
Then a ward organiser phones to say the numbers for his area have changed overnight and he does not know why. Somebody opened an old copy and saved over the live one. Nobody can tell which version is right, because there is no record of who changed what.
This is not a discipline problem. Spreadsheets are the wrong shape for the job, and it is worth being precise about why.
Four things field data needs that a spreadsheet does not provide
Concurrent writes without overwriting
Twenty workers are entering data at the same time, mostly on phones, mostly on patchy mobile data. A spreadsheet resolves conflicts at the file or cell level with last-write-wins. A canvassing system needs to resolve at the record level: two people updating two different households in the same street must both succeed, always.
An audit trail
When a ward reports 400 confirmed supporters and the candidate's own instinct says 150, somebody has to find out which is true. That requires knowing who entered each record and when. Without it you are choosing between two numbers with no way to test either.
This is not primarily about catching dishonesty, though inflated canvass returns are a known problem in campaigns that pay per contact. It is about being able to audit at all.
Offline capture
Canvassing happens in stairwells, in basements, in villages with one bar of signal. A tool that needs connectivity at the moment of entry produces workers who write on paper and transcribe later — and transcription is where accuracy tends to suffer, because it happens hours later from memory.
Role-based visibility
A ward worker needs their own streets. A ward organiser needs their ward. The campaign manager needs everything. A shared spreadsheet gives everyone everything, which is both a security problem and a usability one — nobody wants to scroll 40,000 rows to find their 300.
What a canvassing record should actually contain
The temptation is to capture everything. Resist it: every field you add is a field a tired volunteer has to fill at a doorstep, and long forms produce blank fields and invented answers.
The core set:
- Household identifier — the address, tied to the roll where possible.
- Who was spoken to — a named voter, or "resident, name not given".
- Support level — the single most important field. See below.
- Date and worker — automatic, never typed.
- Contactability — mobile number, if offered. Optional, and it should stay optional.
- One free-text note — capped short. "Works nights, don't call before 2pm" is worth more than three structured fields.
Things that look useful and usually are not: detailed issue-by-issue preferences, long demographic forms, anything the voter would find intrusive at the door. They slow the conversation and the data is rarely acted on.
The support scale, and why it should be short
Support level is the field everything else depends on. The standard shape is a five-point ordinal scale — ordered categories rather than measured numbers — plus a state for “no conversation yet”:
| Level | Meaning | Election-day action |
|---|---|---|
| Strong support | Committed, will vote for you | Turnout target — make sure they vote |
| Lean support | Probably, not certain | Second contact before polling day |
| Undecided | Genuinely open | Persuasion target |
| Lean against | Probably opposed | Leave alone |
| Strong against | Committed elsewhere | Leave alone — contact helps the opponent |
| Not contacted | No conversation yet | Canvassing target |
Five points, not ten. A ten-point scale sounds more precise and produces less reliable data, because two workers will not apply "6" and "7" the same way. Coarse data applied consistently beats fine data applied randomly.
The last row of that table is the one campaigns get wrong most often. Contacting a committed opponent reminds them to vote. The canvass is not just about finding your voters — it is about knowing whose door to stop knocking on.
Why this structure pays off on one specific day
All of it exists for polling day. The turnout operation is simple to state and impossible without clean data:
- Take everyone marked strong or lean support.
- Through the day, record who has voted — from polling-station observers or sign-in sheets.
- Subtract. What remains is your list of supporters who have not yet voted.
- Send workers to exactly those addresses, prioritised by how close they are.
- Repeat every two hours, with the list shrinking each time.
In a close local race this is frequently the difference between winning and losing, and it is entirely mechanical. But it only works if the support data is trustworthy and the "has voted" updates can flow in from multiple polling stations at once without conflicting — which brings you back to the four requirements at the top.
With a spreadsheet, step 3 happens once, at about 2pm, badly.
Handling the data responsibly
A campaign database is a file of named individuals and their political opinions. Political opinion is treated as a special category of personal data under many data-protection regimes, including the GDPR, and handling it usually carries stricter conditions than ordinary contact data.
Before you collect anything, check what your own jurisdiction requires. Election law and data-protection law both apply, they differ substantially between countries, and they change. Rules on what you may do with an electoral roll, whether you need consent, how you must respond to access or deletion requests, and what you may do with the data after polling day are all jurisdiction-specific. Nothing on this page is legal advice, and a campaign handling voter data at any scale should take advice locally rather than rely on a general article.
Independent of what the law requires in your case, these are reasonable baselines:
- Collect only what you will act on. If you will never run a phone bank, do not collect phone numbers. Every extra field is additional risk with no operational return.
- Scope access by role. A ward worker should never be able to export the full roll.
- Revoke access when workers leave. Campaigns have high turnover and this step is almost always forgotten.
- Decide retention and deletion before you collect. Write down how long each kind of record is kept, what happens to it after polling day, and who is responsible for carrying that out. Deciding this after the campaign, when volunteers have dispersed and copies are on personal phones, is how campaign data leaks.
- Never merge in data the voter did not knowingly give you for this purpose.
The practical argument sits alongside the ethical one. Campaign data leaks are damaging in a way that is hard to recover from mid-race, and the worker whose phone is lost on a bus is a far more likely source than any attacker.
The short version
A field campaign needs record-level concurrent writes, an audit trail, offline capture and role-scoped access. A spreadsheet provides none of the four. Everything else — dashboards, heat maps, budget tracking — is useful, but those four are what decide whether the number on polling morning is real.
Votix — campaign management, panel and field app
Voter records, household surveys, field worker assignments, analytics, budgets and election-day operations in one connected system. Web panel for the office, Android app for the field. Demo and pricing by enquiry.
See Votix →