How to Identify Hail-Damaged Neighborhoods (Without Wasting a Week of Knocks)

Every storm-restoration crew has lived the same bad week. A line of storms rolls through on a Tuesday night. Wednesday morning the office is buzzing, the hail map app shows a fat purple blob over half the county, and three sales reps load up and start driving. By Friday they have knocked 400 doors, written four inspections, and signed nobody, because the blob covered 600 square miles and the hail that actually broke roofs fell on about nine streets.

The blob is not the answer. A hail report says hail happened somewhere inside a polygon. It does not say which subdivision took 1.75-inch stones at a steep angle for eleven minutes versus which one got a five-minute dusting of pea-sized ice that bounced off and melted by sunrise. The gap between those two outcomes is the entire difference between a profitable storm and a wasted tank of gas.

Identifying hail-damaged neighborhoods is a narrowing problem. You start with a region the size of a county and you have to end with a list of streets ranked by how likely each roof is to fail an adjuster's inspection. Everything below is the method for doing that narrowing fast, with the public data that already exists, plus the ground-truthing that separates a real damage call from a hopeful one. None of it requires you to buy leads. It requires you to read the storm correctly and then confirm with your own eyes.

Why the hail-map blob lies to you

The radar products and consumer hail maps you have seen are built mostly from MESH — Maximum Estimated Size of Hail. MESH is a radar-derived estimate. A dual-polarization radar measures reflectivity and the shape of particles aloft, runs it through an algorithm, and spits out an estimated maximum stone size for each pixel. It is genuinely useful. It is also an estimate of what was in the cloud, not a measurement of what hit the ground.

Four things break the link between the radar pixel and the roof:

• Melt. A 1.5-inch stone aloft can arrive as a 1-inch stone at the ground in warm air, or shatter, or partially melt. The radar saw the big version.
• Wind drift. Hail does not fall straight down through a sheared storm. A strong updraft and crosswind can carry stones a quarter mile or more downwind of where the radar core sat. The damage stripe is offset from the reflectivity core, and the offset direction is consistent within a storm.
• Beam geometry. Radar beams climb with distance from the site. Far from the radar, the beam is sampling ice 10,000-plus feet up, and the relationship to ground impact gets loose. Close to the radar there are cone-of-silence gaps overhead.
• Duration and density. MESH reports a maximum stone size. It says nothing about how many stones fell or for how long. A roof's fate depends heavily on stone density and dwell time, not only on the single biggest rock.

So the purple blob is a starting hypothesis, not a target list. The job is to take that hypothesis and tighten it using independent sources until you have street-level confidence.

What "damage" even means before you knock

Keep one number in your head as the rough threshold: functional hail damage to asphalt shingles generally starts showing up around the 1-inch stone mark, and gets serious past 1.25-1.5 inches, depending on shingle age, type, and impact angle. Pea-sized (0.25") and even dime-sized (0.7") hail rarely produces the bruising and granule loss an adjuster will pay for, on a roof that was sound to begin with. Quarter-sized (1") is the honest "start paying attention" line. Golf-ball (1.75") and up will damage almost anything in its path.

This matters because half the neighborhoods inside the blob got sub-threshold hail. If you can rule those out from the desk, you just saved your crew two days.

The narrowing funnel, top to bottom

Here is the whole workflow as a funnel. The rest of the piece walks each stage in detail.

1. Region — Which counties got reported hail at all? (NWS / SPC reports)
2. Corridor — Where inside the county did the strongest core track? (radar archive + MESH)
3. Swath — Where did the ground swath actually fall, accounting for wind drift? (spotter reports, social, your own driving)
4. Subdivision — Which neighborhoods in that swath have roofs old or vulnerable enough to fail? (housing age + roof condition)
5. Street and door — Which specific roofs are most likely due, ranked? (ground-truth + per-roof modeling)

Most contractors do step 1, glance at step 2, and skip straight to knocking. The money is in steps 3 through 5.

Stage 1: Pin the region with official storm reports

Start with the free public record, because it is authoritative and it is what adjusters lean on too.

NWS Local Storm Reports (LSRs) and the SPC Storm Reports page are your first stop. The Storm Prediction Center publishes a daily map and CSV of every hail, wind, and tornado report phoned in or relayed to the local forecast office. Each hail report carries a size (in hundredths of an inch — a "100" is a 1.00-inch stone), a time, a lat/long, and often a short remark. The next-day archive is what you want; the day-of map updates live.

What to pull for each storm:

• Every hail report and its size, sorted largest first.
• The time sequence — reports in time order trace the storm's path direction, which tells you which way to expect wind drift.
• The county warning text from the Iowa Environmental Mesonet (IEM) archive, which mirrors NWS warnings and LSRs and lets you query by date and county for free.

A practical read: if you see a clean line of hail reports running southwest-to-northeast with sizes climbing from 1" to 2.5" and back down, you are looking at a discrete supercell that laid a long, narrow swath. If you see scattered 1"-1.25" reports splattered over a wide area with no line, that was a messy multicell or a squall line that dropped marginal hail broadly. The first one is a tight, high-value target. The second one is a low-yield grind.

The reports are sparse — that is the catch

LSRs depend on a human seeing the hail and someone reporting it. Rural areas and overnight storms are massively under-reported. A blank spot on the report map is not proof of no hail; it often means nobody was awake to measure it. So treat the reports as confirmed positives, never as confirmed negatives. You will fill the gaps in the next stages.

Decode the report sizes fast

NWS hail reports come in hundredths of an inch, and reps misread them constantly. Keep this reference taped to the dash so you can sort a report list at a glance:

When you sort a storm's reports by this scale, the picture sharpens immediately. A storm whose biggest report is a 75 is a drive-by; a storm with a string of 150s and a 250 is a season-maker. Work the second one and let a competitor burn fuel on the first.

Tell the storm type from the report pattern

The shape of the report cloud tells you what kind of storm you are chasing, and that changes everything about how you canvass.

• Discrete supercell. A clean line of reports, sizes ramping up then down, oriented along the storm's track. This is the dream: a narrow, intense swath maybe a half mile to two miles wide and many miles long. High damage density, sharp edges, easy to map. Prioritize these.
• Squall line / QLCS. A long band of mostly marginal 1" to 1.25" reports spread broadly with embedded stronger cells. Damage is patchy. You have to hunt for the embedded cores; the rest is low yield.
• Multicell cluster. Scattered reports of varying size with no clear line. Hail fell in clumps. Ground-truthing is mandatory because the desk data will not resolve where the clumps landed.
• Slow / training storms. Reports concentrated in a small area over a long time window. Low storm motion means long dwell time, which means even modest stone sizes can do real damage because the same roofs got pelted repeatedly. Do not dismiss a slow 1.25" storm; dwell time compensates for size.

Name the storm type before you load the trucks. It tells you whether you are looking for one tight stripe or a scatter of clumps, and it sets your expectations for damage density.

Stage 2: Trace the corridor with radar

Now you want the storm's actual track, not merely the dots people called in. Free and low-cost tools:

• NOAA NCEI radar archive (NEXRAD Level II / Level III). You can pull the historical radar for the exact date and step through the storm frame by frame. Watch the reflectivity core and, if you have a viewer that shows it, the MESH / hail size product and correlation coefficient (a dual-pol field where a sharp drop inside a high-reflectivity core is a strong hail signature).
• Consumer hail-map services that package MESH into address-level maps. Fine as a quick visual, but remember what MESH is and is not.

What you are extracting at this stage:

• Core track centerline. Draw the line the strongest reflectivity followed across the county.
• Storm motion vector. Direction and speed. A fast-moving cell (40+ mph) spreads its energy and shortens dwell time per roof; a slow or training cell hammers the same roofs.
• Inferred drift direction. Hail gets thrown to the right and downwind of a sheared updraft. Combine storm motion with the fact that the heaviest ground accumulation usually sits slightly off the reflectivity centerline, on the downwind side.

Mark the corridor as a band maybe a mile or two wide along that centerline, nudged downwind. That band is your stage-2 hypothesis. You have gone from a county to a corridor.

Reading the dual-pol hail signature

If your radar viewer exposes dual-polarization products, you have a sharper read than reflectivity alone. Two fields earn their keep:

• Correlation coefficient (CC). This measures how uniform the particles in a pixel are. Rain is uniform, so CC stays high. Large hail is irregular and tumbling, so CC drops. A pocket of low CC sitting inside a core of very high reflectivity is a classic large-hail signature. That pocket is where the big stones were.
• Differential reflectivity (ZDR). Rain drops are flattened, giving high ZDR; large hail tumbles and looks round, giving ZDR near zero. A core with sky-high reflectivity but ZDR near zero is hail rather than heavy rain. The combination — high reflectivity, low CC, near-zero ZDR — is the fingerprint of a significant-hail core.

Step the storm through frame by frame and watch where that fingerprint sits and how it tracks. The ground path of the worst hail follows that signature, offset downwind. This is the same physics adjusters and meteorologists lean on, so reading it yourself puts you on equal footing when you discuss a storm.

Note the storm's history over the area

While you are in the radar archive, check whether the cell intensified or weakened as it crossed your service area. A cell that was a marginal blob upstream and then exploded right over a subdivision means the worst damage is concentrated at the far end of its track, not the near end. Conversely a cell that was a monster upstream and was collapsing by the time it reached town may have dropped its real hail one county over. The track is a story with a beginning, middle, and end; knowing which act played out over your neighborhoods keeps you from over- or under-estimating the swath.

Worked example: offset is real

Say the radar core tracked due east along the north edge of a town, and storm motion was 35 mph to the east with strong southerly inflow. The biggest LSR (a 2" report) came in from a spot about half a mile south of the core centerline. That is your drift signal. Shift your whole search band south by roughly that half mile. Crews that knock the streets directly under the radar core — instead of the offset band — routinely find weaker damage and can't figure out why. The hail walked sideways on the way down.

Stage 3: Ground-truth the swath

The corridor is still a desk hypothesis. Stage 3 is where you confirm where ice actually hit pavement. This is the highest-leverage hour of the whole process and most reps skip it.

Sources that tell you where it really fell

• Spotter and CoCoRaHS reports. CoCoRaHS is a volunteer precipitation network; some observers log hail size and notes. Combined with trained-spotter LSRs, these are real ground observations.
• Local social media. Search the storm date plus the town name on neighborhood apps and local Facebook groups. People post photos of hail next to a quarter, a golf ball, a deck railing full of ice, dented car hoods, shredded garden plants. A photo of stones piled in a gutter on Elm Street is worth more than any radar pixel. Note the address or cross streets in the post.
• Collateral damage on the drive. When you do roll out, read the soft targets first — they damage at lower thresholds than roofs and they are visible from the truck:
  • Metal: dented gutters, downspouts, garage doors, A/C condenser fins, mailboxes, road signs. Round dents from above = hail. Window screens with tears or spatter marks.
  • Soft surfaces: shredded leaves and stripped tree bark on the storm-facing side, mangled hostas and annuals, holes punched in vinyl siding or window wraps.
  • Spatter marks: the clean circular splash marks hail leaves on wood decks, painted surfaces, and the oxidized film on metal flashing and electrical boxes. Fresh spatter has a directional bias — it tells you the impact angle, which tells you which roof slopes took the worst of it.

The directional fingerprint

Hail almost never falls straight down in a real storm. It comes in at an angle, driven by wind, so it hammers one or two slopes of every roof and barely touches the others. Read the spatter and the soft-metal dents to find the dominant impact direction for the neighborhood — say, hail came from the west-southwest. Now you know that on every house, the west and south slopes are the ones to inspect hardest, and a roof that looks clean from the street (east side) may be wrecked on the back. This single insight prevents the most common false negative in the trade: clearing a roof because the visible slope looked fine.

A 20-minute windshield survey beats a day of guessing

Drive a few transects across your corridor band, perpendicular to the storm track, slow, with a passenger logging collateral. You are looking for the edges. Hail swaths have surprisingly sharp boundaries — you can cross from "every gutter dented" to "nothing" in three or four blocks. Pin those edges. The interior of that envelope, biased to the high-damage side, is your real target neighborhood. You have now gone from a corridor to a confirmed swath with mapped edges.

A collateral scoring sheet for the passenger

Give whoever rides shotgun a simple tally sheet so the windshield survey produces data, not vibes. At each cross-street, have them score what they see, then you have a numeric gradient across the swath instead of a memory:

Total each block. A block scoring 10-plus across multiple houses is core swath. A block scoring 2-3 is an edge or miss. When you map the totals, the high-damage spine of the swath draws itself, and that spine — not the radar centerline — is where the trucks go first.

Drive the transects in the right order

Drive across the suspected swath first to find the edges and the spine, then drive along the spine to find where it is widest and most intense. Two perpendicular passes give you the full geometry in under half an hour. Resist the urge to start knocking on the first damaged street you hit; you do not yet know whether it is the edge or the core, and you want your reps starting at the core.

Stage 4: Inside the swath, find the roofs that will actually fail

Here is the part nobody tells the new rep: the same hail produces totally different outcomes on two roofs next to each other, because the roofs were not the same age or quality going in. A 3-year-old architectural shingle laughs off hail that destroys a 22-year-old 3-tab two doors down. So within the swath, you re-rank by roof vulnerability.

Vulnerability drivers, roughly in order of impact:

1. Shingle age. Older asphalt has lost flexibility and granules; it bruises and fractures at lower impact energy. This is the single biggest factor.
2. Shingle type and weight. Thin 3-tab fails before heavier laminated/architectural. Older mat technology fails before modern SBS-modified "impact-resistant" Class 4 product.
3. Prior wear. A roof already curling, blistering, or balding on the granules is primed to fail.
4. Slope orientation relative to the storm's impact angle (from Stage 3).
5. Exposure — no tree cover on the storm-facing side means full impact energy.

Reading neighborhood age from the outside

You can estimate housing age across a subdivision without knocking a single door:

• County assessor / parcel data is public in most places and lists year built per parcel. A subdivision platted in 2001 with original roofs is a far better target than one built in 2019. Many counties expose this through a free GIS portal.
• Census data and historical aerials show when tracts developed.
• Visual tells from the truck: uniform roof color and style across a block usually means the whole subdivision was roofed around the same time — either at build or after the last big storm. If a block has mostly weathered, faded shingles with a few crisp new roofs scattered in, the new ones are the houses that already claimed after a prior storm. The faded majority are your due roofs.

The "already done" tell

Walk or drive a target street and count fresh roofs. A cluster of brand-new roofs on otherwise-old houses is a fingerprint of a previous hail event that already got worked — by you or a competitor. That tells you two things: this neighborhood does take hail damage (good), and a chunk of the inventory is already too new to re-fail (so your real opportunity is the houses that were skipped, were rentals, had absentee owners, or had owners who declined last time). Those skipped houses are gold because the competition already creamed the easy ones and left them.

Stage 5: Rank the doors — and where per-roof data does the heavy lifting

By now you have a swath with mapped edges, an impact direction, and a sense of which subdivisions are old enough to be vulnerable. The last step is turning that into a ranked door list so your reps knock the highest-probability roofs first while the storm is fresh and before competitors saturate the area. Sequencing matters: in a worked storm, the difference between knocking a street on day 2 versus day 9 is enormous.

Doing the ranking by hand is slow. You are mentally joining three datasets — where the hail fell, how old each roof is, and how each roof's slopes face the impact — across thousands of parcels. That is exactly the join that per-roof storm intelligence is built to automate.

Where RoofPredict fits

RoofPredict is a tool for roofing contractors that does this narrowing at the address level. Instead of a county-sized blob, it gives you a per-roof view that combines two things:

• A roof-age range for each address, estimated from aerial imagery. Important honesty point: this is a range (for example, "roughly 14-19 years"), not an install date. Aerial estimation can tell an old, weathered roof from a fresh one and bracket it; it cannot read a permit. You still confirm on the roof.
• Storm physics modeled per roof — the hail energy and impact characteristics estimated for that specific structure, rather than a single MESH number smeared over a whole zip code.

Put together, that lets you sort a neighborhood by which roofs are most likely due — old enough to be vulnerable and in the part of the swath that took real energy — and hand reps a route that knocks those first. It is the Stage 4 and Stage 5 join done across every parcel at once, so your windshield survey becomes confirmation instead of discovery.

What it is not: it does not tell you a given roof is damaged, it does not approve or handle anything with an insurer, and a modeled storm is odds, not proof. Treat the output as a prioritized knock list — a much sharper hypothesis than the blob — that your inspection then confirms or clears. The roof still has to be documented honestly on-site, the homeowner still owns any claim, and the insurer still decides coverage. The tool gets your best people in front of the right doors faster; it does not replace the ladder.

Build the ranked list (manual version)

If you are doing this without per-roof tooling, here is the poor-man's version that still beats the blob:

1. Pull the assessor parcel layer for the swath and filter to homes built before a cutoff year (often roughly 12-15 years before today, adjusted for local roof lifespan).
2. Overlay your mapped swath edges from Stage 3.
3. Drop the parcels with obviously new roofs (from your drive notes / recent aerials).
4. Bias the remaining list toward the downwind/high-energy side of the swath and the storm-facing slopes.
5. Sort by age oldest-first; that is your knock order for the first 48 hours.

Assign streets, do not let reps free-roam

Once you have the ranked list, carve it into territories and assign them. A swath worked by reps who each pick their own streets gets double-knocked in some spots and skipped in others. Hand each rep a defined block grid with the knock order baked in, and have them mark every door as one of: inspected-damage, inspected-no-damage, not-home, or declined. That status map is what lets you run efficient second and third passes on the not-homes — which is where a surprising share of signed jobs actually come from, because the best prospect is often the one who was at work during the first sweep.

The desk-and-field tool stack

You do not need expensive software to run this method, but a few tools make it fast. Here is a practical stack, free first:

• SPC Storm Reports + NWS LSR — confirmed hail sizes and times. Free.
• Iowa Environmental Mesonet (IEM) — query NWS warnings and LSRs by date and county; archive goes back years. Free.
• NOAA NCEI radar archive — step through historical Level II/III radar to trace the core and motion. Free, slight learning curve.
• CoCoRaHS — volunteer ground reports including hail in some areas. Free.
• County assessor / GIS parcel portal — year-built per parcel for housing-age targeting. Free in most counties.
• Recent aerial imagery — to spot already-replaced roofs and rule out new subdivisions.
• A canvassing app — to assign territory, mark door status, and prevent double-knocking. Paid, optional but high-leverage at scale.
• Per-roof storm intelligence — to automate the age + storm-energy join across every parcel and hand reps a ranked route. Paid; this is the layer RoofPredict sits in.

The free stack alone beats knocking the blob by a wide margin. The paid layers mostly buy you speed, which during a worked storm is the whole ballgame.

On the roof: confirm before you promise anything

Narrowing gets the right roofs in front of you. The inspection is where you earn the job, and where contractors get themselves in legal and reputational trouble if they are sloppy. Document conditions and estimates honestly; let the adjuster and insurer make the coverage call.

What real hail bruising looks like

On asphalt shingles, functional hail damage shows as:

• Bruises: a soft spot where the impact fractured the mat under the granules. Press it — a true bruise feels spongy, like a bruise on fruit, because the fiberglass mat cracked. This is the finding adjusters care about, because a fractured mat is the path to future leaks.
• Granule loss in a random, scattered pattern with sharp-edged, fresh craters — the asphalt underneath looks dark and clean, not weathered. Random and circular is hail; linear or directional may be mechanical or foot traffic.
• Mat fractures radiating from impact points.

Density matters. Adjusters typically assess hits within a 10-foot by 10-foot test square on each slope and each elevation, counting bruises to judge whether the slope is functionally damaged. Mark your squares with chalk, photograph each with the chalk circles and a size reference, and log hits per slope per elevation. Do this the same way every time so your documentation is consistent and credible.

A documentation packet that holds up

The inspections that get approved are the ones where the photo set tells a complete, coherent story without you in the room. Build the same packet every roof:

1. Address establishing shot — the house with a visible house number, so every later photo is anchored to a location.
2. Date proof — a phone shot with the date visible, or rely on the embedded photo metadata. Tie the damage to the storm date.
3. Collateral first — A/C fins, window screens, fence caps, downspouts, the mailbox. This places the hail and its direction before you ever touch the roof.
4. Each slope wide — one establishing photo per slope and elevation so the adjuster can orient.
5. The test square — chalked 10x10 on each slope, hits circled, with a coin or tape measure for scale.
6. Close-ups of representative bruises — a few clear macro shots showing the spongy fracture and the fresh, dark asphalt in the crater.
7. Soft-metal roof accessories — dented vents, turbines, drip edge, valley metal, with the dent direction consistent with your collateral reads.
8. A short written summary — storm date, stone size from the LSR, dominant impact direction, hits per slope, and your honest assessment of which slopes are functionally damaged.

Keep elevations labeled the same way every time (front/back/left/right or N/S/E/W) so a reviewer never has to guess which slope a photo shows. Consistency reads as competence, and competence gets your findings taken at face value.

Match the damage to the storm

The strongest inspections close the loop between the desk research and the roof. If your LSR said 1.75-inch hail came from the west-southwest, the bruises should be roughly that size, concentrated on the west and south slopes, with the soft-metal dents pointing the same way and the ground collateral on the west exposures. When every layer agrees, the story is airtight. When the roof shows tiny scattered marks on the east slope and your data said big hail from the west, something is off — either it is old damage, a different event, or weathering, and you should clear it rather than force it.

What is NOT hail (and will get you embarrassed)

New reps call these in as hail and burn their credibility with adjusters:

• Blistering: raised bumps from trapped gases/moisture in the asphalt. The pop leaves a crater but there is no mat fracture under it and the pattern relates to manufacturing, not a storm direction.
• Granule loss from normal weathering / foot traffic: smooth, no fracture, often in walk paths.
• Mechanical damage from prior work — straight lines, tool marks.
• Manufacturing defects and algae/lichen staining.

Know the difference cold. An adjuster who catches you marking blisters as hail discounts everything else you show them that day.

Use the accessories as corroboration

The shingles are the claim, but soft-metal accessories are your supporting evidence and they are harder to argue with: dents in the metal valley, the drip edge, vents, gutters, the furnace/water-heater roof caps, and especially the soft-aluminum turbine and box vents. If those are freshly dented and the dents face the same direction as your spatter reads, you have a coherent story: hail of size X came from direction Y and impacted the whole north-and-west exposure. Photograph the collateral on the ground (A/C fins, window screens, fence caps) too — it places the storm and the date.

Field safety and the legal lane

Two guardrails that keep you in business.

Safety. Steep, wet, or hail-bruised roofs are slip hazards, and ladder falls are a leading cause of serious injury in this trade. Follow basic fall-protection practice, set ladders at the right angle, and do not send anyone up a roof they are not equipped or trained to be on. A drone or pole camera does a lot of the swath-confirmation and even some inspection work without putting a body at height.

The legal lane. You document conditions and provide an estimate. You do not adjust, approve, or guarantee a claim outcome; you do not promise a roof will be "free" or that a deductible disappears; you do not present a hail forecast or model as proof a specific roof is damaged. Many states regulate how contractors may interact with insurance claims, and a number have rules specifically about roofing and insurance proceeds — check your state's Department of Insurance and contractor-licensing rules. The clean posture, the one that also builds long-term trust with homeowners and adjusters: the contractor documents, the insurer decides coverage, the homeowner owns the claim. Stay in that lane and your reputation compounds instead of catching fire.

Edge cases the desk data will not catch

The funnel handles the common storm cleanly. A few situations need extra judgment:

• Layered storms in one season. A neighborhood that took marginal hail in spring and a bigger event in summer can show mixed-age damage. Inspect carefully and document only what ties to the event you are working; do not blend two storms into one claim.
• Older hail, new urgency. Sometimes real damage from a prior storm went unworked because nobody knocked. The damage is legitimate but old. Be honest about timing; some policies and adjusters treat date-of-loss strictly, so document what you find and let the homeowner and insurer sort the timeline.
• Tile, metal, and wood roofs. The 10x10 asphalt method does not transfer. Tile cracks and chips, metal dents and may shed coating, wood splits along the grain. Each has its own damage signature; do not apply shingle logic to them.
• Impact-resistant (Class 4) roofs. These are built to resist exactly this. A Class 4 roof in a 1.25-inch swath may legitimately have no functional damage. Clearing it is the correct, honest call, and it protects your credibility for the next house.
• Steep or fragile roofs. Some roofs you should not walk at all. Use a drone or pole camera, or inspect from a ladder at the eave. A no-walk inspection with good optics beats a fall.
• Solar panels and complex roofs. Panels, lots of penetrations, and intricate valleys change both the inspection and the eventual repair scope. Flag them so your estimator is not surprised.
• HOA and architectural-control neighborhoods. Some subdivisions require approved colors and materials. Knowing that before you write an estimate saves a painful conversation later.

The neighborhood compounding effect

There is a reason pros work a swath densely instead of cherry-picking one house per street. Roofing is a neighbor business. When you do honest, visible work on a block — clean sites, a yard sign, a crew that shows up when it said it would — the houses around it watch. A homeowner who sees three neighbors get new roofs starts wondering about their own, and a referral from the guy next door costs you nothing and converts far better than a cold knock.

So when you narrow to a swath, saturate it rather than scattering. Concentrate your reps in the confirmed high-damage envelope, do excellent work, and let the block sell the block. The same desk-and-windshield narrowing that finds the swath also tells you where to plant that density for maximum referral spread. One well-worked subdivision is worth a dozen one-off jobs spread across a county, both in margin and in the pipeline it seeds for next season.

Common mistakes that waste the storm

A quick checklist of the errors that turn a real hail event into a losing week:

• Knocking the blob. Treating the whole MESH polygon as a target instead of narrowing to the swath.
• Ignoring wind drift. Knocking the radar centerline instead of the offset, downwind ground swath.
• Trusting report gaps as "no hail." Skipping under-reported rural or overnight-storm areas that actually got hammered.
• Skipping the windshield survey. Spending a day on roofs in a neighborhood the collateral damage would have cleared in twenty minutes.
• One-slope inspections. Clearing a roof on the visible slope when the storm-facing slope on the back is destroyed.
• Knocking new subdivisions. Working streets full of 4-year-old roofs that shrug off the hail.
• Marking blisters as hail. Torching adjuster credibility with bad findings.
• Slow sequencing. Letting the best streets sit until day 9 when competitors have already saturated them.
• Over-promising. Stepping out of the documentation lane and into claims-handling language that gets you in regulatory trouble.

Put it together: a repeatable storm-day playbook

Here is the whole method compressed into a sequence you can run every time a storm clears.

The night of / morning after

1. Pull SPC/NWS storm reports and the IEM archive for the date. Sort hail reports by size; read the path direction from the time sequence.
2. Open the radar archive. Trace the core centerline and storm motion. Nudge a 1-2 mile search band downwind of the core.
3. Search local social/CoCoRaHS for ground photos with addresses. Pin them.

Day 1, daylight

1. Drive transects across your band. Log collateral (dented metal, shredded foliage, spatter). Find and pin the swath edges. Read the dominant impact direction.
2. Pull assessor parcel age for the confirmed swath. Or open your per-roof tool and sort the swath by which roofs are due (old enough + in the high-energy zone).
3. Drop new-roof subdivisions. Build the ranked knock list, oldest-and-hardest-hit first, biased to the storm-facing side.

Day 1-3, knocking

1. Inspect with chalk and a 10x10 test square per slope per elevation. Document bruises (spongy = mat fracture), corroborate with soft-metal collateral, photograph everything with a size reference.
2. Clear the roofs that show only blistering/weathering. Honestly. Your no-damage calls protect your damage calls.
3. Stay in the lane: document and estimate; insurer decides; homeowner owns the claim.

Run that loop and the blob stops mattering. You are no longer reacting to a purple cloud on an app — you are reading the storm the way it actually behaved and putting your crew on the specific roofs the hail wore out, before anyone else gets there. That is the whole game: narrow fast, ground-truth honestly, rank by which roofs are due, and confirm on the ladder.

FAQ

How do I find out exactly where hail fell after a storm?

Start with the free public record: NWS Local Storm Reports and the SPC daily storm reports give you confirmed hail sizes, times, and locations. Then trace the storm core on the NOAA radar archive and nudge your search band downwind for wind drift. Finally, ground-truth it with local social-media hail photos (note the addresses) and a slow drive across the area looking for dented gutters, shredded foliage, and spatter marks. The reports tell you where hail was confirmed; they never prove hail was absent, because rural and overnight storms are badly under-reported.

Are consumer hail maps accurate enough to plan canvassing?

They are a useful starting hypothesis, not a target list. Most hail maps are built from MESH, a radar estimate of the largest hail aloft. It does not account for melt before the stone lands, wind drift that offsets the ground swath from the radar core, or how long and how densely the hail fell. Use the map to find the corridor, then narrow it with ground-truthing. Knocking the whole colored polygon is how crews waste a week.

What size hail actually damages a roof?

As a rough rule for asphalt shingles, functional damage starts becoming likely around 1-inch (quarter-sized) hail and gets serious past about 1.25 to 1.5 inches, with golf-ball (1.75-inch) and larger damaging almost anything. Pea- and dime-sized hail rarely produces the mat-fracturing bruises an adjuster pays for on a sound roof. But size is only one factor; an old, brittle 3-tab roof can be damaged by hail that a new architectural roof shrugs off.

Why do houses on the same street have totally different damage?

Two reasons. First, roof age and quality: an older, weathered shingle fractures at far lower impact energy than a recent architectural or impact-resistant roof next door. Second, slope orientation: hail comes in at an angle, so it hammers the storm-facing slopes and barely touches the others. Always read the impact direction from spatter and soft-metal dents, then inspect the storm-facing slopes hardest, even if the street-facing slope looks clean.

How do I tell real hail damage from blistering or normal wear?

A true hail bruise feels spongy when you press it because the fiberglass mat fractured under the granules, and fresh impacts leave sharp-edged craters with clean, dark asphalt and a random, scattered pattern. Blistering leaves craters too, but there is no mat fracture beneath and the pattern relates to manufacturing rather than storm direction. Weathering and foot-traffic granule loss is smooth with no fracture. Marking blisters as hail destroys your credibility with adjusters, so learn the difference before you knock.

How does RoofPredict help identify which roofs to knock?

It narrows the work to the address level for roofing contractors by combining a roof-age range estimated from aerial imagery with storm physics modeled per roof, then lets you sort a neighborhood by which roofs are most likely due. That turns a county-sized blob into a ranked knock list. Be clear on the limits: the age figure is a range, not an install date; a modeled storm is odds, not proof of damage; and it does not handle claims. You still confirm every roof on the ladder, the homeowner owns the claim, and the insurer decides coverage.

What is wind drift and why does it matter for canvassing?

Hail rarely falls straight down through a sheared storm. The updraft and crosswind carry stones downwind, so the heaviest ground accumulation lands offset from the radar reflectivity core, sometimes a quarter mile or more. If you knock the streets directly under the radar centerline, you will find weaker damage and not understand why. Read the storm motion and the largest ground reports, then shift your whole search band to the downwind, high-energy side of the core.

How fast do I need to move after a hailstorm?

Fast, because the best streets get saturated by competitors within days. The narrowing funnel exists precisely so you can spend hours at the desk and on a windshield survey to build a ranked knock list, then put your crew on the highest-probability roofs first while the storm is fresh. The difference between knocking a prime street on day 2 versus day 9 is often the difference between a full pipeline and picked-over leftovers.

Can I confirm a hail swath without getting on every roof?

Largely, yes. Soft targets damage at lower thresholds than roofs and are visible from the truck: dented gutters, downspouts, garage doors, A/C condenser fins, and mailboxes; torn window screens; shredded leaves and stripped bark on the storm-facing side; and circular spatter marks on decks and metal. A slow drive across the corridor logging this collateral maps the swath edges and the impact direction in about twenty minutes. A drone or pole camera then handles much of the per-roof confirmation without sending a body up a wet, bruised roof.

What should I avoid saying to homeowners to stay compliant?

Stay in the documentation lane. Do not adjust, approve, or guarantee a claim outcome; do not promise a roof will be 'free' or that the deductible disappears; and do not present a hail map or model as proof that a specific roof is damaged. Document conditions and provide an honest estimate, let the insurer decide coverage, and let the homeowner own the claim. Many states regulate contractor involvement in insurance claims, so check your state Department of Insurance and licensing rules.