Most of the country has no early vote data yet, and that is the first thing campaigns need to internalize before October 1. Only four states, Idaho, Minnesota, South Dakota and Virginia, had opened early in-person voting before this month, with Virginia voters already casting ballots at the Arlington County Courthouse on September 18 (Yahoo News). The rest of the map opens mid-to-late October, running out to Kentucky's October 29 start (CEIR). A campaign that panics or celebrates off week-one numbers in most states is reacting to a number that does not exist yet.
The calendar is not uniform, and that is the first trap
North Carolina's absentee ballots go out starting September 4, but in-person early voting does not open until October 15 and runs through October 31, with same-day registration available in that window (NCSBE). Georgia's local officials do not even have to start mailing ballots until between October 5 and 9, which means Georgia's earliest usable statewide signal is a mid-October event, not an early-October one (CEIR). Nevada runs October 17 to 30. Illinois opens September 24 and runs to November 2. Iowa and Tennessee both start October 14 (NewsNation).
Field teams comparing raw vote totals across these states on the same calendar day are comparing states that are not in the same phase of the process. Worse, within a single state, local officials often have discretion over exact start dates, meaning county-to-county totals on a given day are not directly comparable either (CEIR). Build the tracker around each state's own calendar, not a national one.
The number that matters is a gap, not a count
Once returns start landing, the instinct is to watch total ballots cast or the raw partisan split. Both mislead. Population size scales the count, and a state where Democratic-registered voters historically bank early while Republicans show up on Election Day, common through much of the Sun Belt, will produce an apparent "D lead" every single cycle regardless of outcome. That lead is structural, not predictive.
The number that matters is the change in that gap against the same state's 2022 baseline, a midterm comparison, not 2024's presidential-turnout numbers. If the D-to-R early return ratio in a state is running tighter than 2022 or wider, that is a real signal. If it simply matches 2022, nothing has moved. Track the delta, not the level.
This is not a fringe read. The University of Florida Election Lab's Michael McDonald has argued in past cycles that early-vote party composition is a weak standalone predictor precisely because it misses Election Day same-party turnout. CEIR's own research backs the caution from the processing side: states vary in how fast they pre-process and report, so an apparent early surge in one state can just mean that state counts faster, not that turnout is higher.
Twelve states plus DC will still be counting after Election Day
Twelve states and DC accept postmarked-by-Election-Day mail ballots that arrive after November 3, with deadlines running out to November 23 in Washington state. Twenty-five states cut off receipt on Election Day itself (CEIR). A campaign in one of the twelve-plus-DC states that treats its Election Night total as final, for spend decisions, for concession framing, for anything, is working off an incomplete number by design, not by accident. Know which category your state falls into before October, not after.
So What
Build the tracker off each state's own election calendar, not a shared October 1 start date. Pull the exact early-in-person and mail deadlines for every state or district you're running in and set your first data-collection date to match. A tracker built for a national mid-October start will sit empty in Virginia and miss the window entirely in Kentucky.
Set the 2022 baseline now, before returns start. Get the final 2022 partisan early/mail split for your state from your data vendor or state party file. Without that number in hand on day one, you have no way to tell a meaningful shift from routine noise once returns land.
Brief your surrogates and volunteers on what a "D lead" in early returns actually means before it shows up. In Sun Belt states especially, an early partisan gap is a near-certainty every cycle. A GOTV team that hasn't been told this in advance will read it as bad news and either lose morale or, worse, start reallocating spend off a false signal.
Do not adjust spend off week-one volume in any state that processes and reports on a fast timeline. A high early count in a same-day-processing state is often a speed artifact, not an enthusiasm signal. Wait for the party-composition delta against baseline before moving money, and confirm any number your team wants to act on against a named source before it changes a plan.
Source notes: Primary sourcing is the Center for Election Innovation & Research's September 2026 state-by-state calendar (updated September 28), supplemented by state election office pages (North Carolina, Virginia) and news aggregation (Yahoo News, NewsNation) for individual state dates. No 2026 vendor or state-party return data exists yet at time of writing; the McDonald reference on early-vote predictiveness reflects his argument in past cycles and needs a 2026-specific citation before publication.
If your race sits in a state with an early mid-October start, especially Georgia or North Carolina, and you want the baseline-comparison tracker built before returns start, talk to DHP.