fieldwork

Staffing Crews Against a Swath

Hail seasonality is regional, so staffing crews against a swath is a calendar problem before it is a hiring problem

Staffing is a calendar problem before it is a hiring problem

Every roofing owner who has run a storm has had both bad weeks: the one where the phone rings and there is nobody to send, and the one where everybody is on payroll and nothing is happening. Both come out of the same mistake, treating staffing as a question about people when it is first a question about the calendar. Hail has a season, the season has a shape, and the shape differs depending on where you work.

The season moves north

The USDA Midwest Climate Hub's 2024 factsheet on hail is the clearest public statement of that shape. It reports that most states have their largest fraction of annual severe hail reports in May, June or July. Inside that window the gradient runs south to north. States in the south such as Missouri and Illinois see their highest monthly hail fractions during April and May, while northern states such as Wisconsin, Minnesota and the Northern Plains have higher monthly hail fractions in June, July and August. The measure is monthly hail fraction, the percentage of a state's annual hail reports falling in a given month, computed from the NCEI Storm Events Database.

That one fact has a direct staffing consequence, and it cuts both ways. A firm operating across the gradient, say from Missouri up into South Dakota or Minnesota, has a longer usable season than a firm in one metro, because the peak arrives at different times in different parts of its territory. A firm in one metro has a sharper peak, which means a narrower window to staff against and fewer weeks to recover from a slow start. Neither is better. They are different problems. The multi-state firm needs crews that travel and a system for moving them. The single metro firm needs surge capacity it can turn on and off without carrying it all year.

What a tight season looks like up close

Regional averages hide how concentrated a local season can be. Two NWS office climatologies make it concrete, and both are old enough that you should read them as shape rather than as current frequency.

NWS Lake Charles examined 404 hail reports in its county warning area over the period 1955 to 1995 and found that 67 percent of hail reports occur during the March to May time period, with May the peak month at 105 reports, correlating to 26 percent of all reports. A secondary peak occurs in the autumn months. Two thirds of a forty year record landing in three months is a very tight season. If your market looks like that, your staffing problem is almost entirely a surge problem.

NWS Tulsa's page for Fort Smith Municipal Airport covers the period of record 1972 through December 1997 and counts 50 total hail occurrences, about 2.0 hail events per year. May leads with 14 occurrences, April has 11 and March has 10, and there is a small second peak in November with 4 occurrences. Two events a year at one measuring point is a useful corrective: a single location sees hail rarely even in an active state. Your season is built out of a handful of days, which is why the response to each one matters.

Neither record extends past the 1990s. Use them for the shape of a local season in those places over those periods, not as a current count.

Where the volume is

Where to look for work is a separate question from when. The Insurance Information Institute's hail page, drawing on the SPC annual severe weather report summary, lists 5,432 hail events in the United States in 2025, up from 5,373 in 2024. The top states by hail events in 2025 were Texas with 902, then Kansas 375, Oklahoma 369, Nebraska 315, Missouri 253, Colorado 244, South Dakota 232, Tennessee 216, Illinois 167 and Arkansas 154.

Add those ten counts together and you get 3,227 of the 5,432 events, a little under 60 percent of the national total in ten states. Texas by itself carries more than twice the count of the next state on the list. If you are deciding where a traveling crew should be based, that distribution is more useful than opinion about it. Put it next to the timing from the Midwest factsheet and you get a rough itinerary: most of that volume sits in and around the plains, and across the plains the peak slides later as you go north.

Inspection capacity and production capacity peak at different times

No public dataset will tell you this part, so read it as reasoning rather than as a finding.

A swath makes two different demands on a company, and they do not arrive together. The first is inspection capacity. Someone has to get on roofs, document what is there, and talk to homeowners, and that demand spikes within days of the storm, while the event is fresh and homeowners are still deciding whom to call. The second is production capacity: crews, material, dumpsters, and the people who schedule them. That demand lags, because it waits on adjuster schedules, on approvals, and on supply.

The gap between those two peaks is the central staffing problem in storm work. Staff for the inspection peak and you carry idle production capacity for weeks, then hit a production peak you cannot cover. Staff for the production peak and you miss the window in which the work is actually sold. Most firms learn this the hard way, usually in the second or third event of a season, when the backlog from the first one collides with inspection demand from the new one.

The useful response is not a number, because there is no honest number to give here. It is a structure. Hire or subcontract the two capacities separately, measure each queue on its own, and accept that the ratio between them changes with the size of the event and with how fast local carriers are moving. The one thing you can settle in advance is which of the two is your constraint now, because that is the one you add to before the next storm.

Watches and outlooks stage crews. They do not hire them.

It helps to be precise about which weather product answers which question, because the planning horizon and the dispatch horizon are not the same length.

SPC issues Convective Outlooks for Days 1 through 3, plus a Day 4 to 8 outlook covering roughly 72 to 192 hours out. That is the planning horizon: days, not weeks. The watch lead time goal is at least 45 minutes for the first severe thunderstorm watch of the day, average lead time is about 1 hour for severe thunderstorm watches, and watches typically last 6 to 8 hours. A watch is not a warning.

Put those together and the operational picture is plain. Days out, you can decide where to have people and material. Hours out, you can decide where they sit tonight. None of it is a hiring signal, because you cannot hire in an hour. Hiring runs off the climatology. Staging runs off the forecast.

What the report record cannot tell you

One caveat, because it limits how much weight the numbers above can carry.

The SPC severe weather database is compiled from NWS Storm Data, with hail and damaging wind reports going back to 1955, and the page is explicit that reports cluster around population centers. A report exists because a person saw hail and told someone. Thinly populated counties therefore look quieter than they are. The same page notes that the NWS raised the severe hail threshold from 0.75 inch to 1.00 inch in 2010, and that 0.75 inch reports remain in the annual hail files for legacy purposes, so users may need to filter. Any long record that crosses 2010 has a break in it that has nothing to do with the weather. The current year's reports are preliminary. The Midwest factsheet makes the same point from the other direction, noting that assessing hail climatology is difficult largely because the reporting of hail has been inconsistent.

So report based climatology tells you about seasons and about regions. It does not tell you about next week, and it does not hand you a per county expectation you should bet payroll on.

Three things to settle before the season starts

Write down which of your two capacities, inspection or production, is the binding constraint today, and add to that one first.

Write down where your territory sits on the south to north gradient, and therefore whether you are staffing a long season or a sharp peak.

Write down in advance what you will do with a Day 4 to 8 outlook and what you will do with a watch, so that the staging decision is a policy rather than an argument held on the night of the storm.