2026-05-08

Halloween is an early dinner rush your phone can't absorb

Halloween pulls dinner two hours forward, closes streets your drivers use, and lands on a thin crew. Here is what happens to the phone and how to plan for it.

At 5:40 on October 31 the phone rings eleven times in nine minutes. Four of those callers want delivery to a street that is closed for a neighborhood parade. Two are asking whether you are even open. One is a parent placing a $180 order for a house full of kids who are walking out the door at 6:15. Your line cook, who usually grabs the phone, is plating. The driver isn't back. The host is at the door with a walk-in party of six.

Halloween doesn't add volume so much as it compresses it. Roughly the same number of orders you'd take on a decent Friday arrive inside a ninety-minute window that opens before your normal dinner peak and closes hard once the kids are out.

The curve moves forward about two hours

Families eat before trick-or-treating. That single fact reshapes the night. Your 6 to 8 p.m. dinner block becomes a 4:30 to 6:30 block, and the tail after 7:30 is thinner than a normal weeknight, sometimes dramatically so.

Most kitchens are still in transition at 4:30. Prep is finishing, the dinner crew is arriving, and the person who ends up answering the phone at that hour is whoever is closest to it. So the busiest phone hour of the night lands on your least phone-ready hour of staffing, and the calls that ring out are the ones you never find out about. That asymmetry is the whole argument in the real cost of a missed restaurant phone call.

There's a second effect worth planning for. Because the tail dies early, the crew you scheduled to close is standing around at 8:15 on a night you paid overtime for. Shifting an hour of labor from the back of the night to the front usually costs nothing and buys you a person on the phone during the only hour that matters.

If you have never measured it, the front half of Halloween evening is a good place to start. Count total inbound calls against answered calls for 4 to 7 p.m. and compare the ratio to a normal Friday.

Delivery breaks for reasons that have nothing to do with your kitchen

The food can be perfect and the night can still fall apart in the car.

Streets get closed. Blocks get roped off for parties. Sidewalks fill with people who are not looking for cars, so a driver moves at walking pace through exactly the residential neighborhoods that generate the most orders. A run that takes twelve minutes in March takes twenty-eight, and the driver comes back later than the next three tickets assumed.

The fix is boring and it works: shrink the zone for one night. Pull out the neighborhoods with parades or heavy foot traffic, extend your quoted times across the board, and be honest on the phone about both. Configuring that as a temporary rule instead of a verbal instruction to whoever is answering is the difference between a policy and a hope, which is what delivery zone rules for phone orders is about.

An honest 55-minute quote loses a few orders. A cheerful 30-minute quote that turns into 55 produces a call back, a refund conversation, and a review.

Also decide in advance what happens when a driver can't reach an address. On a night with closed streets it will happen more than once, and the default outcome is a driver sitting in a car calling a customer who isn't picking up while three other orders go cold in the bag. A meet-at-the-corner rule, agreed at order time for the affected blocks, is uglier than a doorstep handoff and much better than the alternative.

You will be short a person, and you already know which one

Halloween is a night people ask off. Parents want to take their own kids out. Younger staff have plans. Whoever you have is likely to be a slightly thinner crew than a normal Friday, running a compressed rush.

The role that gets sacrificed first is the phone, because it's the only job with no one standing in front of it. Nobody skips a table to answer a phone; they skip the phone to serve a table. That is the correct instinct in the moment and it costs you the whole off-premise side of the night. The structural version of this problem sits in restaurant staffing shortage phone coverage.

If you're running a voice agent, this is the night it earns its keep, because the calls it handles are exactly the ones your crew was never going to reach. If you're not, the realistic alternative is deciding in advance who owns the phone from 4 to 7 and taking them off the floor for it. Half measures produce a ringing phone.

The kitchen ceiling is the real constraint

At some point on Halloween evening your kitchen hits its capacity and the honest thing to do is stop promising times you can't hit.

A phone that keeps taking orders at the same quoted time regardless of how deep the queue is will bury your line. Throttling is the mechanism that stops it: as the queue builds, quoted times stretch automatically, and at a defined ceiling new pickup slots move out rather than stacking. The mechanics are in order throttling and kitchen capacity.

This is worth setting before the day, not during it. In the middle of a rush nobody is going to calmly adjust a quote policy.

What to do the week before

Set your Halloween hours in whatever system answers your phone, including the early close if you're taking one, so callers asking "are you open tonight" get a correct answer without a person. That's the ordinary use of holiday hours overrides, and it applies to any single-day schedule change.

Then decide three things and write them down where the people working that night can see them. What time does the delivery zone shrink and to what. What is the quoted pickup time at the peak, and at what queue depth does it stretch. Who owns the phone if you are answering it by hand.

Also take large orders in advance and take them seriously. A house party ordering for twenty is a catering order wearing a takeout costume, and it should be prepped on a schedule with a name, a hard time, and payment already handled rather than fired when it happens to reach the top of the queue.

The number to pull on November 1

Look at your call log for October 31, specifically 4 p.m. to 7 p.m., and compare answered calls to total calls in that window against a normal Friday's same three hours.

If the answered share dropped meaningfully, you didn't have a Halloween problem. You have a compression problem that shows up on every spike night you have: the first real snow, the local team's playoff run, the Friday before Thanksgiving. The same pattern turns up in snow day delivery surge calls, and the fix is identical each time: know the window, staff the phone specifically for it, quote honestly, and cap what the kitchen will accept.

One night a year isn't worth rebuilding your operation for. Six or seven nights a year, all failing the same way, is a different arithmetic, and October 31 is the cheapest place to run the test.

More on seasonal playbooks

All seasonal playbooks articles

Frequently asked questions

Hear it answer a real call.

Call the demo line and order like a customer would, or book time and we'll walk your team through it.