Campaign field guide
Scheduling the field program: canvasses, phone banks, early voting, GOTV
A field director's working guide to the four scheduling problems every downballot campaign hits between now and November, and how to run each one without finding a gap at 6am.
Canvass shift scheduling
A canvass runs on three pieces: a launch location, a walk list, and volunteers in pairs. The scheduling failure mode is uneven turf: eight volunteers at the Saturday launch in the candidate's home ward and zero in the ward that decides the race. Canvassing software earns its keep by making the imbalance visible before Saturday.
The working pattern: create canvass shifts as templates per launch site, two-to-four-hour blocks, weekends heavy. Assign by ward first, not by name, because a canvasser who lives on the turf knocks more doors per hour and gets fewer doors slammed. In Capstone Volunteer the fill panel already ranks by lives-in-ward, so the ward-first discipline happens by default. Log the no-shows after every launch, because canvass reliability in September predicts election day reliability in November.
Walk lists themselves stay in whatever tool cuts them. The schedule that gets bodies to the launch on time is the part campaigns lose track of, and the part this system holds.
Phone bank scheduling
Phone banks are the easiest shifts to fill and the easiest to no-show, because staying home to make calls feels optional in a way a vote center does not. Treat them as real shifts: fixed windows, named assignments, attendance logged. A Tuesday phone bank with twelve names and five callers is a recruitment problem you can fix in October. The same discovery in the GOTV week is not fixable.
Weeknight evenings, 6pm to 8:30pm, two-hour blocks. Schedule callers in the same grid as everything else so the conflict detection catches the volunteer who signed up to canvass and phone bank at the same hour. The follow-ups-due queue covers confirmation calls the day before, which is the single biggest lever on phone bank attendance.
Early voting site coverage
Early voting is a marathon disguised as a sprint: in DC, 25 early-voting sites across a 7-day window, all needing greeter coverage during voting hours. That is hundreds of shifts before election day arrives, and it is where volunteer fatigue starts showing up in the no-show numbers.
The pattern that works: shorter shifts than election day, four hours maximum, and the same volunteer at the same site across the week so faces become familiar. Coverage tolerance can be looser than election day, and the point of the daily report during this window is triage: which sites can run thin today so that the reliable people stay fresh for Tuesday. The 7am report makes that call a two-minute read instead of a nightly reconciliation.
Election day and GOTV
Get out the vote week is when every scheduling debt comes due. On the seeded DC campaign, election day meant 74 vote centers from 6am, three shift blocks each, staffed from the same 268-person pool that just worked a 7-day early-voting window. There is no slack and no next week.
Three rules hold the day together. First, overstaff the morning: the 6am shift has the highest no-show rate and the least recovery time, so schedule it at 120 percent. Second, hold a ward-based reserve: five to ten proven-reliable volunteers per region with no assignment, released by the gap panel as holes appear. Third, run the night-before check from the gap report, not from memory: every gap visible at 9pm on Monday is fillable, and the same gap at 6am on Tuesday is a hole in the program. The night-before gap report is the reason this product exists.
The reserve pattern is where reliability history pays off. The fill panel ranks the reserve by who has covered each center before, so the volunteer sent to Coolidge High School at 7am has stood at that table before and needs no orientation.
The calendar between now and November
August and September are recruitment and canvass launches. October is phone banks, confirmations, and the early-voting schedule build. The first week of November is execution. Every one of those phases writes into the same grid, so the volunteer pool, the reliability history, and the coverage picture compound instead of restarting with each phase. A campaign that starts the grid in August walks into GOTV week with three months of no-show data behind every fill decision. The field organizer carrying that calendar has a guide of their own.