Timing the Drop: When to Mail So It Lands

This is the last post in our series on the physical postcard. We have covered layout, pens and legibility, and cost. This week is timing: when the card should land, and therefore when it should go in the mail.

The most common mistake is not mailing too late. It is treating the mail date as a decision for October. The mail date is a calculation you can do now, from three inputs: how long USPS takes, when ballots arrive in your target states, and when you want the card on the kitchen counter. For the broader fall calendar, see the primary-to-general post. This post is only about the drop.

How long the mail takes

USPS’s service standard for single-piece First-Class Mail is one to five days. In 2025 USPS changed the standards in two steps. Mail originating more than 50 miles from a regional processing center got an extra day, and even local mail that stays inside one facility’s area is now standard at two to three days. The old assumption of “two or three days” is no longer safe for most routes.

Do not guess. USPS publishes a service standards map and a ZIP-to-ZIP lookup. Enter the ZIP code you will mail from and a few ZIP codes from your target list. Write down the longest answer, then add two days, because a standard is a target, not a promise, and late October is heavy mail season. For planning, assume five to seven days from drop to doorstep, and seven if you are mailing cross-country or from a rural post office.

When the card should land

The answer depends on how your target voters vote.

In vote-by-mail states, the card should land a day or two before the ballot does, so the voter expects the ballot and opens it. Colorado mails ballots 18 to 22 days before the election, Washington 18 days, Oregon and Nevada 20. For November 3, 2026, that is ballots going out between October 12 and 16. A card meant to land October 10 needs to be in the mail by October 3 to 5. That is earlier than most programs plan for, and it is why this calculation belongs in July.

Other states mail earlier: California 29 days out, Arizona 24 to 27, Michigan 45. Look up each state on your list and work backward.

In states where most people vote in person, the card should land in the final one to two weeks before election day. That is the default for turnout mail. The evidence for “later is always better” is weaker than the folk wisdom suggests, though. Panagopoulos (2011) tested nonpartisan turnout calls delivered three days, two weeks, and four weeks before a municipal election and found no significant difference overall, though low-propensity voters responded better to the late contact. Vote Forward reports that in 2020 and 2022, letters sent earlier in their window were slightly more effective than letters sent later, and in 2024 they used a mail window of October 1 through 29 rather than a single date.

The safe plan for in-person states is to land between October 20 and 27, which means a drop between October 13 and 20.

One drop, mailed centrally

A single drop is simpler. Everyone writes, you collect, it all goes out on one day, and you know when it landed. For a program with one target state, this is the right call.

Staggered drops make sense when you have targets in both vote-by-mail and in-person states, which need cards on different dates. Split the list by state, give each segment its own land date and drop date, and run them as two small campaigns. Do not stagger for its own sake. Two drops double the collection and tracking work. What you should never do is let the drop happen by accident, with cards trickling out over three weeks as volunteers get around to it.

Which is the argument for mailing centrally. Letting volunteers mail their own cards is easier on you and harder on the calendar. Cards go out when volunteers remember, from wherever they live, and you have no way to know what actually went out. Collecting cards and mailing them yourself costs a few hours and some boxes. In exchange you control the date, you can count the cards, pull the ones with a blank address or writing in the barcode strip, and drop the whole stack at one post office with a known service standard.

If volunteers are remote, have them mail their finished packet back to you (budgeted in last week’s worksheet) two weeks ahead of your drop. If that is impossible, give them a specific mail date and a reminder the day before, and accept the loss of control.

Returned mail

A postcard has no return address unless you print one. If you want to see returns, print your organization’s address small on the message side, outside the barcode strip.

Log every one that comes back. Mark the address bad in your list and do not send to it again this cycle. Returned mail is the cheapest list cleaning you will ever get, and as we wrote in the primary-to-general post, it is the main thing a mail program learns about its own data. Do not re-mail a returned address on the theory that the second card might get through. It will not, and it costs 65 cents.

What to do this week

List your target states and whether each is mostly vote-by-mail or in-person. Look up the ballot mail date for each vote-by-mail state. Run the USPS lookup from your mailing ZIP to three ZIPs in each state. Pick a land date per state, subtract delivery time plus two days, and put the drop date on the calendar. Then work backward from the drop to set the hand-in deadline for volunteers.

That is the series. If you are starting from the top, begin with the anatomy of a postcard that gets read.