A hurricane makes landfall on a Tuesday. By Thursday, restoration search volume in the affected metros is running at five to seven times baseline, payouts spike, and every contractor within two hundred miles is fielding calls.
Most operators handle this badly, in one of two directions: they drop calls they could have serviced, or they book work they cannot staff. Both are expensive, and the second is worse.
Routing has to read live capacity
The default setup in home services is a static waterfall — calls go to branch A, overflow to branch B, then C. It works fine at baseline and fails completely under surge, because branch A stacks up while branch C sits idle.
Capacity-aware routing reads live crew availability and sends the call to whoever can actually take it. During a six-times surge, that difference is the entire story.
Define service areas at ZIP level
Metro-level service areas are a convenient fiction that costs money during surges, because "the Tampa metro" is ninety minutes across in traffic and considerably worse after a storm.
Draw service areas at ZIP level, per crew. Overflow should then go to the nearest available crew rather than to the next branch alphabetically, which is what most waterfalls actually do.
Separate emergency from scheduled intent
An IVR gate distinguishing "water is coming in right now" from "I want a roof inspection sometime" is thirty seconds of work and reorders your entire queue correctly.
Emergency intent should preempt. Scheduled intent can wait, and callers with scheduled intent are markedly more tolerant of a callback — as long as someone tells them that is what is happening.
Set caps that lift deliberately
Concurrency caps exist so you do not pay for calls that ring out. During a declared weather event, those caps should lift — but by a decision, not automatically.
Build the lift as a switch someone flips when the extra crews are actually confirmed. Caps that auto-lift on volume are just a slower way of overselling.
Arrange overflow before you need it
The time to negotiate a relationship with a partner two states away is in February, not on Thursday afternoon. Overflow buyers, contracted answering services, and mutual-aid arrangements with non-competing operators all need to be papered in advance.
The restoration group that handled a six-times surge best in our network had all three arranged before the season started, and used all three.
After-hours matters more than it seems
Storm calls do not respect business hours. A contracted answering service that captures the call, sets expectations, and warm-transfers at open converts a large share of calls that would otherwise go to whoever answered at 11pm.
The margin case
None of this is about capturing more volume. It is about capturing the volume you can service and declining the rest cleanly.
An operator who books three weeks of work they cannot staff spends the following month on refunds, reviews, and rescheduling — and the reputational cost lands right before the next season.