How to Use Hail Data to Find Roofing Jobs: A Targeting Playbook

Hail falls on a county, the storm-chaser trucks roll in, and within 72 hours half the roofing companies in a 60-mile radius are fighting over the same three subdivisions. Meanwhile, two streets over, a 22-year-old roof that took 1.5-inch stones sits untouched for eight months because nobody bothered to look at the actual storm track. That gap between where crews go and where the damage actually landed is the whole game. Hail data, used correctly, closes it.

Most contractors treat hail data as a yes/no alert: "did it hail in my market, yes or no, send everybody." That is the amateur version, and it is why so many storm-restoration efforts burn payroll on doors that were never going to convert. The pros treat hail data as a targeting layer — a way to rank streets, size crews, time the knock, and walk into an inspection already knowing what the stone size should have done to that specific roof covering.

Below is the operating manual a sales manager or owner can actually run: where the data comes from, how to read it without fooling yourself, how to turn a swath polygon into a knock list, how to stack it with roof age and material so you stop guessing, and the compliance lines you do not cross. There are worked examples with real numbers, the workflow pros use after a storm, and the mistakes that quietly kill close rates.

What "hail data" actually means (and the four flavors you'll use)

When someone says "hail data," they could mean any of four very different things. Knowing which one you're holding matters, because each has a different error profile and a different job in your sales motion.

1. Radar-derived hail estimates (MESH and friends)

This is the workhorse. Weather radar doesn't see hailstones directly; it infers them. The most common product is MESH — Maximum Estimated Size of Hail — derived from radar reflectivity and the height of the storm's freezing level. NOAA's MRMS (Multi-Radar/Multi-Sensor) system publishes gridded MESH, and most commercial hail-report vendors build their swath maps on top of it or something very similar.

What MESH gives you: a grid (often roughly 1 km, refined to finer in commercial products) where each cell carries an estimated maximum stone size for the storm. Stitch the daily maxima together and you get a hail swath — the colored polygon you see on vendor maps, usually banded by size (e.g., 0.75", 1", 1.25", 1.5", 2"+).

What MESH is not: ground truth. It's a model. It tends to over-predict in some setups and under-predict in others, and a single radar cell covers a lot of rooftops that did not all get hit identically.

2. Ground reports (spotter / Storm Prediction Center)

The NWS Storm Prediction Center compiles human-reported hail — spotters, trained observers, the public — into a storm-reports database, sized in inches ("quarter," "golf ball," "baseball" in the old shorthand). These are sparse and biased toward populated areas and daylight, but they're real observations. Use them to sanity-check the radar swath: if MESH says 1.75" over a town and three spotters reported 1.5"–2", you trust the swath more.

3. Insurance / verification-grade reports

Commercial vendors (the hail-report companies adjusters rely on) sell a per-address PDF: estimated max hail size at that coordinate, date, and sometimes wind. Adjusters use these. You should know what your local carriers tend to reference, because a roof you call "hail damaged" gets a lot easier to document when an independent report says 1.5" stones hit that exact lat/long on that date. This is verification, not lead-gen — but it shapes which storms are worth your route.

4. Wind and the storm's full hazard profile

Hail rarely travels alone. The same supercell drops hail and a wind field; metal flashing, ridge caps, and shingles can fail from wind even where the stone size was modest. Treating a storm as "hail only" leaves jobs on the table. Pull wind gust estimates for the same footprint.

The takeaway: never run your whole campaign off one number. The radar swath tells you where to look, ground reports tell you whether to believe the swath, and the roof itself tells you the truth. Your job is to make those three agree before you spend a crew.

The physics you need so you don't chase phantom storms

You don't need a meteorology degree, but you do need a working mental model of what size hail does to what roof, because it determines whether a swath is worth chasing at all.

Stone size vs. damage threshold by covering

Damage is a function of stone size, density, fall speed, impact angle, and the roof covering's condition and age. A brittle 20-year-old three-tab and a fresh impact-rated architectural shingle respond very differently to the same 1-inch stone. Rough, practitioner-level thresholds:

• Under ~0.75" (pea to penny): Usually cosmetic at most on asphalt; rarely a functional claim on its own. Soft metals (gutters, vents, AC fins) may dent and serve as collateral indicators.
• ~1.0" (quarter): The widely cited rough threshold where functional damage to aging asphalt becomes plausible. Soft-metal denting common. Worth a look, especially on older roofs.
• 1.25"–1.5" (half-dollar to golf ball): Functional bruising and granule loss become likely on asphalt; mat fracturing shows up. This is the meat of most profitable storm work.
• 1.75"+ (golf ball to baseball): Widespread functional damage probable across most asphalt regardless of age; tile cracking; serious soft-metal and even structural denting.

These are plausibility thresholds, not guarantees. A 1.25" swath over a neighborhood of 5-year-old impact-rated shingles may produce very few legitimate functional claims. The same swath over 18–25 year old three-tab is a target-rich street.

Why age multiplies everything

A roof's hail vulnerability climbs with age as the asphalt loses oils, granules embed loosely, and the mat gets brittle. The same stone that bruises but doesn't fracture a 6-year-old roof can fracture the mat on a 20-year-old one. This is the single most important reason to fuse hail data with roof age: the swath tells you the load; the age tells you the resistance. Damage lives where high load meets low resistance.

Directionality and the "soft metal tells the truth" rule

Hail usually comes in at an angle, driven by wind, so damage concentrates on the storm-facing slopes (often west/southwest in classic setups, but check the storm). Inspect the windward slopes hardest. And when an asphalt roof is ambiguous, the soft metals don't lie: gutters, downspouts, gutter aprons, valley metal, vents, and the AC condenser fins record directional impacts that confirm a hail event even when the shingles are arguable.

Where to get the data: free vs. paid, and what each is good for

You can run a credible targeting operation on free public data. You'll run a faster, defensible one with paid tools. Here's the honest split.

Free / public sources

• NOAA MRMS / MESH — gridded radar hail estimates; the basis for most swaths.
• NWS SPC storm reports — daily, searchable, downloadable point reports of hail size and wind.
• NWS Storm Events Database — historical archive going back decades; great for seasonality analysis and for documenting that a date had hail in a county.
• NCEI (National Centers for Environmental Information) — the long-term climate archive behind a lot of this.
• Local NWS Weather Forecast Office pages — local storm summaries, sometimes with refined size estimates.

Free data is enough to answer: did a real hail event hit my market, roughly where, and roughly how big? What it won't do cheaply is hand you a clean per-address swath joined to parcels and ranked into a route. That's the labor the paid tools sell.

Paid / commercial sources

• Hail-report / swath vendors — interactive swath maps, per-address verification reports, sometimes notification when your defined territory gets hit. This is what most storm-restoration teams pay for.
• Per-address verification reports — the documentation layer adjusters recognize.
• Property / parcel data and aerial-measurement tools — to turn a polygon into addresses, owners, and roof square footage for estimating.

The pattern that works: use free data to validate the storm and free yourself from vendor hype, then use a paid swath + parcel join to move fast on the streets you've decided are real.

From swath to street list: the targeting workflow

Here's the part most "how to use hail data" articles skip. A swath polygon is not a lead list. Turning it into a ranked knock list is a repeatable process. Run it the same way every storm.

Step 1 — Confirm the event is real and worth working

Before you mobilize anything:

1. Pull the radar swath (MESH) for the date.
2. Cross-check ground reports from SPC for the same footprint. Do the reported sizes roughly match the swath bands?
3. Check wind in the same footprint.
4. Decide your minimum workable size for this market and season. For a market dominated by older asphalt, 1.0"–1.25" may be workable; for newer stock, you may set 1.5"+.

If the swath, the ground reports, and the wind data don't roughly agree that something real happened at or above your threshold, don't burn payroll. Over-predicting MESH has launched a thousand wasted canvasses.

Step 2 — Clip the swath to your service area and your threshold

Intersect the swath polygon with: (a) your licensed/serviceable geography, and (b) your minimum size band. You now have a much smaller working polygon — the only ground worth your crews this week.

Step 3 — Join the polygon to parcels/addresses

Overlay the clipped swath on parcel/property data to produce an actual list of addresses inside the affected size bands. This is where a property-data layer or aerial tool earns its keep. Output columns you want per address:

• Address, owner-occupied vs. rental flag (owner-occupied converts better for retail restoration)
• Estimated max hail size at that point
• Roof square footage (rough, from aerial)
• Roof covering type if available (asphalt vs. tile vs. metal changes the pitch)
• Roof age estimate — covered in its own section below; this is the multiplier

Step 4 — Rank, don't just list

Now score each address. A simple, effective scoring model:

Priority score = (hail-size factor) × (roof-age factor) × (occupancy factor)

Worked example of factors you can start with and tune:

A 19-year-old owner-occupied roof under a 1.6" band scores 1.5 × 2.0 × 1.2 = 3.6. A 5-year-old rental under a 1.1" band scores 1.0 × 0.5 × 0.8 = 0.4. You knock the 3.6 first. Same storm, nine-times difference in priority. That's the entire value of fusing the layers instead of blasting the polygon.

Step 5 — Sequence routes by score density

Don't just sort the spreadsheet — map it. Cluster the high scorers geographically so a canvasser walks the densest pocket of 2.5+ scores in a tight loop instead of zig-zagging. Score density per block is what makes a canvasser productive. A block with eight 16+ year roofs under a 1.5" band is a morning's worth of high-probability conversations.

Step 6 — Time the knock

There's a window. Knock too early and homeowners haven't registered anything happened and you look like an ambulance chaser. Knock too late and three competitors and the carrier's own process have already shaped the conversation. The practical window for retail storm restoration is generally the first one to three weeks, with the strongest results often after the first newscast about the storm but before saturation. Use the storm date from your data as day zero and plan the canvass calendar from it.

Fusing hail data with roof age and material — the step pros actually win on

If you remember one thing: hail data alone tells you where the load hit. It cannot tell you which roofs were vulnerable. Vulnerability is mostly age and covering. The contractors who out-convert everyone are the ones who knock the aging, storm-hit roofs first and skip the new ones a competitor wastes a week on.

The problem has always been getting roof age at scale. You can't pull a permit for every address before you knock, and "drive by and guess from the curb" doesn't scale past a few blocks.

This is where a per-roof intelligence layer changes the workflow. RoofPredict scores roofs house-by-house: it estimates a roof-age range per address from aerial imagery and models storm impact per individual roof rather than smearing one swath band across a whole zip code. So instead of a flat polygon, you get a ranked picture of which roofs are actually due — the ones aging out on their own, and the ones a given storm most plausibly wore out, address by address.

How that drops into the workflow above: it does Step 3 and Step 4 for you with real per-roof inputs. Rather than guessing the age factor, you start from an estimated age range; rather than one swath band for the block, you get storm modeling resolved to each roof. Your canvassers walk a list already sorted by the roofs most likely to be both old enough and hit hard enough to have a legitimate, documentable condition.

Honest limits, because they matter:

• Roof age is a range, not a birth certificate. Aerial-derived age is an estimate — "roughly 15–20 years," not "installed March 2007." Use it to prioritize, then confirm at the inspection. Don't ever represent an estimated range as a known install date to a homeowner or carrier.
• Storm modeling gives odds, not proof. Per-roof storm physics tells you which roofs were most likely affected and how hard. It raises or lowers a roof's priority. It is not evidence of damage and must never be presented as proof that a specific roof is damaged. The only proof is the documented inspection.
• It ranks doors; it does not replace the ladder. The tool decides where to spend the first hour. A qualified inspector still has to get on the roof and document actual conditions.

Used that way — to rank, not to claim — fusing per-roof age and storm modeling with your hail swath is the difference between a canvasser having forty good conversations a day and forty doors slammed. RoofPredict is a targeting and routing layer, not a lead-buying service and not a claims tool: you still document conditions, the insurer still decides coverage, and the homeowner still owns the claim.

A full worked example: one storm, start to finish

Make it concrete. Numbers are illustrative but the method is exactly what you'd run.

The storm. A late-April supercell tracks across the north side of a metro you serve. MESH shows a swath: a 0.75" outer fringe, a wide 1.25" core, and an embedded 1.75"–2.0" streak about 14 miles long and 2 miles wide running west-to-east through three suburbs.

Step 1 — Validate. SPC ground reports for that evening: two "golf ball" (1.75") reports and one "half dollar" (1.25") near the towns under the core. Wind gusts estimated 60–70 mph along the same line. Agreement is good. This is a real, workable event. Minimum workable size for this older-stock market: 1.25".

Step 2 — Clip. You serve two of the three suburbs. Intersect the swath with your service area and the 1.25"+ bands. Working polygon: roughly 9 square miles, the part of the core and the 1.75"+ streak inside your territory.

Step 3 — Join parcels. The clipped polygon contains ~3,100 single-family parcels. Property data gives owner-occupancy and aerial square footage. A per-roof layer estimates age ranges and models the storm per roof.

Step 4 — Score. Apply the scoring model. Results bucket out roughly:

Step 5 — Sequence. The 520 top scorers cluster heavily in two subdivisions under the 1.75"+ streak. Those become Canvass Zone A for day one and two. The 1.5–2.9 band becomes Zone B for the back half of week one.

Step 6 — Time it. Storm was Tuesday night. Local news covered it Wednesday. You start knocking Zone A Thursday morning — homeowners aware, market not yet saturated. You hold the under-1.5 band as a lower-priority sweep or skip it.

The outcome logic. Instead of putting six canvassers randomly across 3,100 doors, you put them on the 520 highest-probability roofs first. If even a modest fraction of those old-and-hit roofs convert to inspections and a portion of those to signed jobs, your cost-per-acquired-job on this storm drops sharply versus the spray-the-polygon approach — because you stopped paying canvassers to knock 5-year-old roofs that radar happened to color in.

Reading hail size shorthand and the equipment that records it

Your team needs a shared language for stone size, because "big hail" means nothing on a route sheet. The classic NWS reference scale, smallest to largest, is the one adjusters and homeowners both half-remember, so use it:

Train every canvasser and inspector to translate radar bands into this scale on sight, because homeowners describe what they saw in these terms ("it was golf-ball size in the yard"). When a homeowner's recollection matches your swath band, that's a second confirmation that the roof saw workable hail — and it's a natural, honest way to open the inspection conversation without overselling.

A note on the evidence hail leaves and how to capture it. Hail itself melts, so the proof lives on the property: granule piles in gutters and at downspout outlets, spatter marks (clean spots where oxidation was knocked off) on painted surfaces, dents on the AC condenser fins and its top cover, bruised soft-metal vents, and the actual bruises on shingles you find with a flat hand. Issue inspectors a chalk circle, a measuring guide, and a camera with a scale reference. Document the directional pattern — which slopes took the most hits — because it should line up with the storm's wind direction. When your photo set shows directionally consistent impacts across both the roof and the collateral soft metals, you have an honest, defensible condition report. When it doesn't line up, you have an honest "no workable damage" answer, which protects your reputation just as much.

Pulling MESH and SPC data yourself: a hands-on mini-tutorial

You don't have to depend entirely on a vendor dashboard. Here's how to validate a storm with public tools so you always have an independent read.

SPC storm reports (fastest free check). The Storm Prediction Center posts a daily storm reports page. Find the date, open the hail reports, and you get a list of point observations with size (in hundredths of an inch — "100" means 1.00", "175" means 1.75"), time, location, and remarks. Plot those points against the towns you serve. Three or more reports of 1.25"+ clustered over your market is a strong "this is real" signal. Sparse or no reports doesn't always mean no hail — reporting is biased to populated, daytime areas — but a cluster of large reports is hard to argue with.

Storm Events Database / NCEI (history and documentation). When you need to establish that a county had a hail event on a given date — for storm-worthiness analysis or to understand a roof's exposure history — the Storm Events Database lets you query by state, county, date range, and event type. Export the hail events and you have a defensible record of when and how big.

MRMS / MESH (the swath itself). The raw gridded MESH product is more technical; most contractors consume it through a commercial swath tool that renders it as a map. If you go to the source, know that you're looking at a maximum-estimated-size grid, and that the same caveats apply: it's modeled, calibration varies with distance from radar, and a cell averages over many roofs.

A simple validation routine you can run in 20 minutes per storm:

1. Open the SPC reports for the date. Note the largest hail reports near your market and roughly where they cluster.
2. Open your swath tool (or MRMS). Confirm the swath's high bands sit over the same towns as the ground reports.
3. Pull wind reports/gusts for the same footprint.
4. Write one sentence in your storm log: "Date X, [towns], MESH up to 1.75", three ground reports 1.25"–2.0", wind to 65 mph — WORKABLE, min size 1.25"." Or "marginal, single over-predicted cell, no ground confirmation — HOLD."

That one-sentence log entry, kept for every event, becomes a season-long record of which storms were real and how your reads compared to reality. Over a year it makes your team calibrate their own market.

Staffing and crew math: matching headcount to the swath

Hail data also sizes the operation, not only the route. Once you have a scored address list, you can do real capacity planning instead of guessing.

Work backward from doors. Suppose a disciplined canvasser knocks roughly 60–80 doors in a productive afternoon shift and reaches a conversation at maybe a third of them. Of those conversations, a fraction agree to an inspection, and a fraction of inspections with legitimate documented conditions convert to a signed job. The exact rates are yours to measure, but the structure lets you plan:

• If your top-priority bucket (score 3.0+) has 520 addresses and a canvasser can meaningfully cover ~70 doors a shift, that's roughly 7–8 canvasser-shifts to work the A-list once. Two canvassers for four days, or four for two days.
• Your inspection crews need to keep pace with set inspections, or your funnel clogs and the timing window closes on appointments you already earned.
• Your production/build capacity has to absorb the signed jobs before the next storm stacks on top.

The point: a scored, ranked list from hail data lets you decide how many people to put where, and it tells you when to bring in extra labor for a big swath versus when a single storm doesn't justify pulling crews off existing work. Spraying an unscored polygon gives you none of that — you find out you were over- or under-staffed only after the money's spent.

A practical staffing tiering by swath size:

Tracking conversion so the data actually improves your aim

Targeting with hail data only compounds if you measure it. The metrics that matter, by stage:

• Doors knocked per canvasser-shift — productivity baseline.
• Contact rate — conversations per door. Low contact rate means timing or list quality is off.
• Inspection-set rate — inspections booked per conversation. This is where roof age and storm targeting pay off: a well-targeted list books more inspections per conversation because you're talking to the right homeowners.
• Documented-condition rate — inspections that find legitimate workable conditions. If this is low on a "hot" swath, either the swath over-predicted or you targeted too-new roofs.
• Signed-job rate — and ultimately cost per acquired job, which is the number that proves or disproves your whole approach.

Tag every lead with its source storm, swath band, and roof-age band so you can later answer: did the 1.75"+ / 16-yr+ bucket really convert better than the 1.25" / 9–15-yr bucket? Almost always it does, and the magnitude tells you how aggressively to weight age and size in your scoring model next storm. This is how a one-size-fits-all scoring table becomes your scoring table, tuned to your market's roofing stock and your carriers' behavior.

A simple post-storm scorecard you can fill in:

Run it every storm. After a season you'll know exactly which storm profiles and which roof-age/size combinations are worth mobilizing for — and which marginal events you should let competitors waste their payroll on.

Reading the swath without fooling yourself: calibration and error

The fastest way to waste a storm is to believe the swath too literally. A few discipline points:

MESH over- and under-prediction

Radar hail estimation has known biases that vary by storm structure, radar distance, and beam height. Far from the radar, the beam is higher and may overshoot low-level melting, skewing estimates. Always treat the size band as approximate. A "1.5" cell" might have produced anywhere in a window around that — which is exactly why ground reports and the actual soft-metal evidence on site are your reality check.

One cell covers many roofs

A single grid cell can blanket dozens of homes that did not all receive identical stones — micro-variations in the storm, tree cover, and slope orientation matter. The swath is a probability surface, not a parcel-by-parcel verdict. This is precisely why per-roof modeling and an on-roof inspection still matter even inside a hot swath.

Recency and "is this storm fresh or stacked?"

Markets get hit repeatedly. A roof under today's 1.25" swath may also have sat under a 1.75" swath 14 months ago. Pull the storm history for the address, not only today's event. A roof with multiple recent hail exposures and high age is a much stronger candidate than a one-storm roof of the same age. The historical archives (Storm Events Database, NCEI) make this checkable.

Don't confuse "in the polygon" with "damaged"

Being inside a colored band means the roof plausibly saw that size. It does not mean it's damaged, and it absolutely does not mean a claim is owed. Functional damage is a roof-by-roof, inspection-by-inspection finding. Your data ranks who to inspect. The inspection determines condition. Keep that line bright — it protects you legally and reputationally.

Building a year-round pipeline, not only a post-storm scramble

The best storm operations don't start canvassing when hail falls — they start prepared and never fully stop, because steady age-out demand fills the gaps between storms.

Pre-storm: own your territory's baseline

• Map your market's roof-age distribution. Know which subdivisions are crossing the 15-year line. Those are aging-out targets regardless of weather, and they're the streets where the next storm will produce the most legitimate damage. A per-roof age layer makes this a map instead of a guess.
• Pre-stage canvass zones. Have your high-age clusters already drawn so that when a storm hits, you intersect the swath with a list you already trust.
• Set notification triggers. Configure alerts for hail above your threshold in your defined territory so you're not refreshing radar manually.

During the season: standing operating procedure

Write the six-step workflow above into an SOP your team runs identically every event. Speed plus discipline beats either alone. The teams that fumble storms usually fumble because every event is improvised.

Between storms: work the age-out backlog

When no storm has hit in weeks, you still have the aging roofs. The same per-roof age data that prioritizes storm work also feeds a steady "your roof is at the age where it's worth a free inspection" retail motion — honest, value-first, no weather event required. This is what keeps crews busy in a quiet month and keeps you off the boom-bust roller coaster that wrecks storm-only companies.

Roof material and regional factors that change your thresholds

The same swath means different things in different markets, because roofing stock varies by region, era, and price point. Tune your targeting to what's actually on the roofs you serve.

• Asphalt three-tab (older, lower-cost): The most hail-vulnerable common covering, especially past 15 years. Lower your workable-size threshold here; a 1.0"–1.25" event over an old three-tab neighborhood is genuinely productive.
• Architectural / laminate asphalt: Thicker, more impact-tolerant than three-tab but still vulnerable with age. Your bread and butter in most suburban markets.
• Impact-rated (Class 4) shingles: Designed to resist hail. A neighborhood of recent Class 4 roofs under a 1.25" band may produce very few functional claims — deprioritize unless stones were large. Knowing where builders used Class 4 stock saves you wasted canvassing.
• Tile (concrete/clay): Common in parts of the Southwest and South. Cracks and shatters under larger hail (often 1.75"+), and damage can be subtle from the ground. Different inspection skill set; size threshold runs higher.
• Metal: Often cosmetic denting rather than functional failure from hail; the conversation is different and frequently a harder sell on functional grounds.
• Wood shake: Splits and impact fractures; aging shake is vulnerable, inspection is specialized.

Regional realities matter too. The hail belt running up the central U.S. — broadly the Plains and into the upper Midwest and parts of the Southeast — gets the most frequent severe hail, which means more competition and more carrier scrutiny. In those markets your targeting discipline is your edge, because everyone has the same swath; the contractor who knocks the right age-and-size combination first wins. In lower-frequency hail markets, a single significant storm is a bigger event relative to baseline demand, and being fast and well-targeted matters even more because the opportunity won't repeat for a while.

Roof pitch, tree cover, and orientation also modulate per-roof outcomes inside one swath. Steep, storm-facing slopes take more direct impacts; heavily tree-shaded roofs may be partially protected; complex roofs with lots of valleys and penetrations have more soft-metal collateral to read. This micro-variation is exactly why a flat swath band oversimplifies and why per-roof modeling plus an actual inspection remain necessary even inside a hot zone.

From inspection to honest handoff: keeping the line clean

Once your targeting lands a crew on a legitimately old, legitimately hit roof, the job is to document — cleanly, honestly, and in a way that respects who decides what.

A defensible inspection package contains: dated, scaled photos of representative shingle impacts; the directional pattern across slopes; soft-metal collateral (gutters, vents, AC, flashing); spatter and granule-loss evidence; and a written summary of observed conditions tied to the storm date your data flagged. You are recording what you see. You are measuring and estimating scope. You are not adjudicating a claim.

The handoff language matters as much as the photos. Good practice sounds like: "Here's what we documented on your roof. Your roof appears to be roughly in the 15–20 year range, and a storm with workable hail passed over your address on [date]. These are the conditions we found. The decision on coverage is your insurer's, the claim is yours, and we'll document and estimate the repair scope to support whatever you decide." That framing is accurate, it builds trust, and it keeps you on the right side of every regulator: the roofer documents conditions and estimates, the insurer decides coverage, the homeowner owns the claim. Targeting got you to the right door; honesty closes it without exposing you.

Compliance, ethics, and the lines you do not cross

Storm restoration has a reputation problem because a slice of the industry has crossed lines. Targeting with hail data is completely legitimate. How you talk about it is where companies get into trouble. Keep these straight.

Don't promise outcomes you don't control

You document conditions and write estimates. The insurer decides coverage. The homeowner owns the claim. Never tell a homeowner their claim will be approved, never promise to "get them a new roof," and never offer to absorb, waive, rebate, or "eat" a deductible — that last one is illegal in many states and is insurance fraud in spirit and often in law. Several states have explicit statutes about deductible handling; know yours.

Don't represent estimates as facts

• Roof age from imagery is a range/estimate, not a documented install date. Say "appears to be roughly 15–20 years old"; confirm with the homeowner or a permit if it matters.
• A hail swath or per-roof storm model is an odds/likelihood statement, not proof of damage. Never show a homeowner a swath map and say "this proves your roof is damaged." It proves a storm passed; the inspection proves condition.

Stay clean on contracts and contingencies

Know your state's rules on contingency agreements tied to insurance proceeds, right-of-rescission windows, and what must be disclosed. The FTC's Cooling-Off Rule and many state home-solicitation laws give homeowners cancellation rights on door-to-door sales — disclose them.

Don't fabricate or "create" damage

Obvious, but it has to be said: documenting pre-existing or storm-caused conditions honestly is the job. Manufacturing damage, or exaggerating cosmetic marks into functional claims, is fraud and ends careers. Good targeting means you don't need to — you're knocking roofs that genuinely were old and genuinely got hit.

Safety is part of compliance

Every roof your team inspects is a fall hazard. OSHA fall-protection requirements apply to your inspectors and crews. A storm surge in workload is exactly when corners get cut and people get hurt. Bake fall protection and a heat plan (storm season is hot) into the SOP, not as an afterthought.

What pros get wrong (the expensive mistakes)

A field guide to the errors that quietly destroy storm ROI:

1. Chasing the whole polygon. Treating every address in the swath as equal. Fix: score and rank by age × size × occupancy.
2. Ignoring roof age entirely. Knocking 4-year-old roofs because radar colored them in. Fix: fuse an age layer; deprioritize new stock.
3. Believing MESH literally. Mobilizing on a single over-predicted cell with no ground-report or wind agreement. Fix: triangulate before you spend payroll.
4. Skipping wind. Calling a storm "no hail, skip it" when a 70-mph wind field tore ridge caps and flashing. Fix: always pull wind for the same footprint.
5. Bad timing. Knocking three weeks late into a saturated market, or so early homeowners think you're crazy. Fix: date-zero off the storm, canvass in the 1–3 week window.
6. Forgetting storm history. Treating today's 1.2" event as the roof's only exposure. Fix: pull the address's hail history; stacked exposures + age = strong candidates.
7. Overselling the data to homeowners. Showing a swath as "proof." Fix: data ranks who to inspect; the inspection is the truth, and you say so.
8. No off-storm motion. Going dark for two months between events. Fix: work the age-out backlog year-round.
9. Ignoring soft-metal evidence. Arguing about shingles while the dented gutters and bent fins sit right there telling the real story. Fix: train inspectors to read collateral first.
10. Crossing compliance lines. Deductible offers, coverage promises, fabricated damage. Fix: train the whole team on the do-not-say list; it protects the brand and keeps you legal.

A 10-point post-storm checklist you can hand to a manager

Print this. Run it every event.

1. Pull MESH swath for the date; note max-size bands.
2. Cross-check SPC ground reports for the same footprint — sizes roughly agree?
3. Pull wind field for the same footprint.
4. Set/confirm minimum workable hail size for this market and season.
5. Clip the swath to service area + minimum size band.
6. Join the clipped polygon to parcels; pull owner-occupancy, square footage, roof age range.
7. Score every address: hail-size × roof-age × occupancy. Sort descending.
8. Map score density; draw Zone A (top scorers, clustered) and Zone B.
9. Set the canvass calendar from storm date-zero; start Zone A inside the window.
10. Brief crews on safety (fall protection, heat) and the compliance do-not-say list before anyone knocks.

Putting it together

Hail data is not a lead list and it is not proof of damage. It is a targeting layer. The contractors who win storms are the ones who treat it that way: validate the event against multiple sources, clip it to what's actually workable, and — this is the part that separates the top crews — fuse it with roof age and material so they knock the old, hit, owner-occupied roofs first and let competitors waste a week on the 5-year-olds inside the same polygon.

Get the targeting right and everything downstream gets easier. Your canvassers have better conversations because they're at the right doors. Your inspectors find real conditions because vulnerability and load actually lined up. Your cost-per-job falls because you stopped paying people to knock roofs that were never candidates. And you stay clean — documenting conditions, letting the insurer decide coverage, letting the homeowner own the claim — because honest targeting means you never have to stretch the truth to make a number.

The swath shows you where the storm went. Roof age shows you what it could hurt. Put those two together, rank the doors, time the knock, and you've turned a weather event into a route worth running.

FAQ

What size hail actually causes roof damage I can work?

As a rough plausibility guide on asphalt shingles: under about 0.75" is usually cosmetic, around 1.0" (quarter-size) is the rough threshold where functional damage to aging roofs becomes plausible, 1.25"-1.5" is where bruising and granule loss get likely, and 1.75"+ tends to produce widespread functional damage regardless of age. These are plausibility thresholds, not guarantees. The same stone affects a brittle 20-year-old roof very differently than a fresh impact-rated one, which is why you fuse size with roof age before deciding which streets to work.

Where can I get hail data for free?

NOAA's MRMS system publishes gridded MESH (Maximum Estimated Size of Hail). The NWS Storm Prediction Center publishes daily ground-reported hail and wind, and the NWS Storm Events Database plus NCEI archive historical events going back decades. Free data is enough to confirm a real storm hit your market and roughly how big. What it won't do cheaply is hand you a clean per-address swath joined to parcels and ranked into a route, which is the labor paid swath and property-data tools sell.

Is MESH or a radar swath proof that a roof is damaged?

No. MESH is a radar-derived estimate of maximum hail size, not a measurement and not proof. A single grid cell covers many roofs that did not all receive identical stones, and radar has known over- and under-prediction biases. Being inside a colored swath band means a roof plausibly saw that size, which tells you who to inspect. Only an on-roof inspection determines actual condition. Never present a swath map to a homeowner as proof their roof is damaged.

How do I turn a hail swath into a list of addresses to knock?

Clip the swath to your service area and your minimum workable size band, then overlay it on parcel/property data to produce actual addresses with owner-occupancy, roof square footage, and a roof-age estimate. Then score each address by hail size times roof age times occupancy and sort descending. Finally, cluster the top scorers geographically so canvassers walk dense pockets of high-priority roofs instead of zig-zagging. The polygon is a starting point, not a lead list.

Why does roof age matter so much when targeting with hail data?

Hail data tells you the load that hit a roof; roof age and covering tell you the roof's resistance. Damage concentrates where high load meets low resistance. The same stone that bruises a 6-year-old roof can fracture the mat on a 20-year-old one. Knocking new roofs inside a swath wastes canvasser time, while old roofs under the same band are the high-probability conversations. Getting roof age at scale used to be the bottleneck, which is why per-roof age estimates change the workflow.

How does RoofPredict fit into a hail-targeting workflow?

RoofPredict is a targeting and routing layer that scores roofs house-by-house. It estimates a roof-age range per address from aerial imagery and models storm impact per individual roof rather than smearing one swath band across a zip code, so you get a ranked picture of which roofs are most likely due. It does the parcel-join and prioritization steps with real per-roof inputs. Honest limits: roof age is a range not an install date, storm modeling gives odds not proof, and it ranks doors rather than replacing the inspection. It is not a lead-buying service and not a claims tool.

When is the best time to knock doors after a hailstorm?

There's a window. Knock too early and homeowners haven't registered anything happened. Knock too late and competitors and the carrier's process have already shaped the conversation. The practical window for retail storm restoration is generally the first one to three weeks, often strongest after the first local news coverage of the storm but before market saturation. Use the storm date from your data as day zero and build the canvass calendar from it.

Should I ignore a storm that was wind-only with little hail?

No. Hail rarely travels alone, and the same storm's wind field can tear ridge caps, flashing, and shingles even where stone size was modest. Calling a storm 'no hail, skip it' leaves jobs on the table. Always pull the wind gust estimates for the same footprint as the hail swath and factor wind-driven damage into which streets are worth working.

What compliance lines should I never cross in storm restoration?

Never promise a claim will be approved, never promise a 'free roof,' and never offer to waive, rebate, or absorb a deductible, which is illegal in many states. Don't represent an estimated roof-age range as a documented install date, and don't present a swath or storm model as proof of damage. You document conditions and write estimates; the insurer decides coverage; the homeowner owns the claim. Also disclose cancellation rights under the FTC Cooling-Off Rule and state home-solicitation laws, and never fabricate or exaggerate damage.

How do I keep crews busy between storms?

Work the age-out backlog. The same per-roof age data that prioritizes storm targets also feeds a steady retail motion to roofs crossing the 15-year line, regardless of weather, with an honest value-first inspection offer. Mapping your market's roof-age distribution ahead of time also pre-stages your canvass zones so you can intersect the next swath with a list you already trust. This keeps you off the boom-bust cycle that wrecks storm-only companies.