How to Find Hail-Damaged Homes to Canvass: A Roofing Contractor's Field Guide

Storm season is short, and the math is unforgiving. A hail event drops on a metro one afternoon, and within seventy-two hours every restoration company within three states has trucks rolling toward the same zip codes. The crews that come out ahead aren't the ones who knock the most doors. They're the ones who knock the right doors first, before the neighborhood is saturated, before the adjusters get backed up, and before the homeowner has already signed with the company that showed up Tuesday.

This guide is about the part of the job that happens before anyone climbs a ladder: figuring out which homes are actually worth your time. Not "which neighborhood got hit" in the vague sense everyone repeats at the supply house, but which streets, which roof ages, which exposures, and which addresses give you the best odds of a sit, an inspection, and a signed contract. I'll walk through how to read a storm, how to build a target list that isn't just a heat map you screenshotted, how to route your canvassers so they aren't burning daylight on dead streets, and the legal and ethical guardrails that keep you out of trouble with the state and the homeowner both.

A quick framing note before we get into it: finding hail-damaged homes is two problems stacked on top of each other. The first is where did the storm do damage — a weather-and-physics question. The second is which of those homes is a good prospect — a roof-condition and homeowner question. Most contractors treat the whole thing as one fuzzy guess. The good ones separate the two, because the tools and the data are completely different, and conflating them is exactly why crews waste weeks canvassing roofs that were either too new to be marginal or too far outside the real damage swath to convert.

What "hail damage worth canvassing" actually means

Before you spend a dollar on data or an hour on a route, get specific about what you're hunting. Hail does several distinct things to an asphalt shingle roof, and they are not equally sellable.

Functional damage is what an insurance adjuster is supposed to be looking for: bruising where the hailstone fractured the shingle mat and dislodged the protective granules, exposing the asphalt underneath to UV. A bruise feels soft, like the spot on an apple, and over the following months and years that exposed asphalt cracks and the shingle fails early. This is the damage that supports a legitimate claim. It is also frequently invisible from the street and sometimes invisible from the roof to an untrained eye.

Cosmetic damage is granule loss and surface marks that don't compromise the shingle's function. Many policies now carry cosmetic damage exclusions or endorsements specifically because carriers got tired of paying for roofs that were marked but not failing. Knowing whether the carriers active in your area write cosmetic exclusions changes which roofs are worth your sales effort.

Collateral or soft-metal evidence is the hail's signature on everything that isn't the roof: dents in gutters, downspouts, fascia, gutter aprons, A/C condenser fins, mailboxes, garage doors, window screens, and especially soft metals like aluminum and copper. Soft metals dent at lower energy than it takes to bruise a shingle, so collateral evidence tells you hail was present and roughly how big it was. A roof with no soft-metal damage anywhere on the property is a roof that probably didn't take functional hail, no matter what the homeowner heard from a neighbor.

When you're canvassing, your real job is to find homes where (a) hail of damaging size actually fell, (b) the roof is old enough and the right material to have taken functional damage, and (c) the homeowner hasn't already been worked. Everything below is in service of stacking those three odds in your favor.

The size threshold matters and is worth internalizing. Damage to typical asphalt shingles generally starts becoming likely once stones reach roughly one inch — about quarter size. The National Weather Service uses one inch as the severe-hail threshold for warnings, and that's a useful floor, not a guarantee. Stones in the 1.0 to 1.5 inch range produce marginal-to-clear damage depending on shingle age, brand, and impact angle. Once you're into 1.75 inches (golf ball) and up, you're looking at widespread functional damage on most roofs in the swath. Pea-sized and dime-sized hail almost never produces a sellable roof on its own. If your storm data says the dominant stone size was three-quarters of an inch, temper your expectations no matter how loud the neighborhood was.

Reading the storm before you read the streets

The single biggest mistake newer storm chasers make is canvassing based on where the storm was loud instead of where it did damage. Those are different maps. A cell can dump terrifying small hail and high wind that wakes up a whole subdivision and produces zero functional roof damage, while a quieter cell two miles away dropped 1.75-inch stones for four minutes and wore out three hundred roofs. The homeowners in the loud neighborhood are primed and ready to talk. The homeowners in the damaged neighborhood may not even know they were hit. Guess which one pays.

Here are the data sources that actually tell you where damage happened, roughly in order of how much I trust them.

NOAA Storm Prediction Center storm reports

The SPC publishes daily storm reports compiling hail, wind, and tornado reports from across the country, including the estimated or measured stone size for each hail report. This is your free, authoritative starting point. Pull the day's report for your metro, note the reported stone sizes and the time stamps, and you've got the skeleton of the event. Two caveats: these reports are sparse (they depend on a human reporting in, so rural and overnight storms are underreported), and reported sizes skew toward the dramatic because people report the big stone they found, not the average. Treat SPC reports as confirmation that a damaging event occurred and a rough size ceiling, not as a complete damage map.

Radar-derived hail products (MESH and friends)

The higher-value signal comes from radar. Dual-polarization radar lets meteorologists distinguish hail cores from heavy rain, and derived products like Maximum Estimated Size of Hail (MESH) produce a gridded estimate of the largest hail the radar thinks fell at each point. This is what most commercial hail-mapping vendors are repackaging when they sell you a swath map. MESH is genuinely useful — it gives you a spatial picture of the core instead of a handful of scattered points — but understand its limits. MESH estimates the radar's guess at the maximum stone aloft, which is not the same as what reached the ground, and it's known to over- and under-estimate in specific situations (high freezing levels, storm motion, beam height far from the radar site). Use the swath to draw your outer boundary and to find the high-energy core. Don't treat a MESH grid cell's number as gospel for a specific roof.

Ground truth: soft metals and spotter reports

Nothing beats getting a person to the edge of the swath and checking soft metals. When you or a lead canvasser drives the suspected boundary and inspects gutters, downspouts, AC fins, and mailbox dents at a few addresses, you find the real edge of functional hail in a way no radar product can give you. This is the step almost everyone skips because it's slow, and it's exactly the step that separates crews who waste two days on the wrong streets from crews who don't. Spend the first morning of a new event ground-truthing the boundary, then build your routes inside the confirmed line.

Hail size, ground truth, and what to expect

Here's a working reference for translating reported or radar-estimated stone size into a canvassing expectation. Treat it as directional, because shingle age, brand, slope, and impact angle all move the line.

The lesson buried in that table: stone size and roof age interact. A 1.0-inch event on a brand-new subdivision is a waste of your week. The same event on a neighborhood of twenty-year-old roofs is a goldmine. Which brings us to the second half of the problem.

Where roof age changes the equation

A storm doesn't damage every roof in its path equally, and the biggest variable after stone size is roof age and condition. A two-year-old architectural shingle has a full, flexible mat and tight granule embedment; it shrugs off marginal hail. A seventeen-year-old roof of the same brand has lost granules, the asphalt has hardened and gotten brittle, and the same hailstone that bounced off the new roof fractures the old one. This is not a small effect. On marginal events, roof age is often the deciding factor between a sellable roof and a no-damage inspection.

This is why "which roofs are due" data is so valuable for canvassing, and why it's a different question from "where did it hail." If you can walk into a confirmed hail swath already knowing which addresses on each street carry older roofs, you've collapsed two filters into one route. You're not only inside the damage boundary — you're inside it and skipping the new construction that's going to give you a polite no.

The traditional way to estimate roof age is painful: county appraisal records sometimes list a roof permit date, but permits lag, get missed, and don't capture re-roofs done without a pull. MLS and prior listing photos sometimes show roof condition. Driving the street and eyeballing for curling, streaking, and obvious wear works but is slow and unreliable from a moving truck. None of these scale to a metro-wide event in the seventy-two-hour window when speed actually matters.

This is the gap where modern aerial-imagery analysis earns its place in a canvassing workflow, and I'll come back to how RoofPredict fits here specifically once we've laid out the full process. The short version: estimating roof age from aerial imagery across an entire storm footprint, as a range per address, turns the slow part of targeting into a list you can route on day one.

Building your target list, step by step

Let's put the pieces together into an actual workflow. This is the process I'd run for a fresh hail event, assuming you got radar alerts or a vendor ping that something significant just dropped.

Step 1: Confirm the event happened and is worth chasing

Within the first hours, pull the SPC storm report and any radar/MESH product you have access to. You're answering one question: did this storm produce stones of one inch or larger over a populated area large enough to be worth deploying crews? If the dominant signal is three-quarter-inch hail over farmland, stand down. If it's 1.5-inch-plus over a suburb of fifteen-year-old roofs, you're going.

Step 2: Draw the swath and the core

Using your radar product, outline the outer boundary of damaging hail (your one-inch line) and then mark the high-energy core inside it (the 1.75-inch-plus zone). You now have a tiered map: the core is your priority-one canvass area, the broader swath is priority two, and everything outside the line you ignore no matter how many calls come in from there.

Step 3: Ground-truth the boundary

Send a lead to the edge of your drawn swath and check soft metals at a handful of addresses along the boundary. Confirm or adjust your line based on what you actually find on gutters and AC units. This single morning of work prevents your whole team from canvassing a phantom edge for two days.

Step 4: Overlay roof age and remove the dead inventory

Now filter the confirmed swath by roof age and condition. Remove or de-prioritize new construction and recently re-roofed homes. On a marginal event this is the difference between a 30% inspection rate and a 10% one. On a strong event it still saves you from wasting time and lets you sequence the highest-probability streets first.

Step 5: Sequence routes for saturation, not only geography

Saturation is your enemy. The neighborhoods nearest major roads and the loudest neighborhoods get worked first by everyone, so they're saturated fastest. Counterintuitively, the quieter damaged neighborhoods — the ones in the core that didn't sound dramatic — are often the least-worked and highest-converting on day three and four. Sequence your routes to hit the high-age, high-energy, low-competition streets while they're fresh, and leave the obvious loud-but-marginal subdivisions for fill-in.

Step 6: Pre-load each canvasser with a clean route, not a map

Don't hand a rep a heat map and say "go knock that." Hand them a turn-by-turn route of specific addresses worth knocking, with the roofs you've already filtered out removed. A canvasser who knocks forty pre-qualified doors will out-produce one who wanders eighty random ones, because the forty convert at a far higher rate and the rep stays in rhythm instead of getting demoralized by no-damage homes.

A worked example

Say a cell drops on the north side of a metro at 4:50 p.m. SPC reports two ping-pong-sized stones and one golf-ball report near a major intersection. Your MESH product shows a six-mile-long core of 1.75-inch-plus running northeast, with a broader 1-inch swath maybe twelve miles long. You ground-truth the southwest edge the next morning and find clear AC-fin denting and gutter dimpling two blocks past where your drawn line was, so you push the boundary out. Inside the confirmed swath you've got a mix: a 2019-built subdivision of 600 homes (skip the bulk of it), an established neighborhood of roughly 1998-2006 builds (priority one), and a stretch of mixed-age homes along an arterial that's already crawling with competitor trucks (deprioritize — saturated). Your day-one routes go to the established neighborhood, core first, oldest roofs first. That's a targeted list, not a heat map. The difference shows up in your contract count by the weekend.

Routing and managing canvassers in the field

Finding the homes is half the battle; getting humans to knock them efficiently is the other half. A few hard-won operational notes.

Territory assignment. Break the swath into bite-sized territories of a few hundred doors each and assign one rep per territory for the duration of the event. Reps who own a territory build neighbor-to-neighbor momentum ("I'm already doing your neighbor's roof at 412") that random coverage destroys. Overlapping reps in the same blocks confuse homeowners and burn goodwill.

Track knocked doors. Whether you use a dedicated canvassing app or a shared map, log every door's outcome: not-home, not-interested, inspection-set, inspected, signed. This prevents double-knocking, lets you circle back to not-homes at better hours, and gives you real conversion data per neighborhood so you can pull reps off dead territory faster.

Hit the right hours. Door-knocking conversion is heavily time-of-day dependent. Late afternoon into early evening on weekdays and mid-morning to mid-afternoon on weekends catch the most people home. Use your dead midday hours for not-home callbacks, drone or street-level roof pre-screening, and supplement paperwork — not for first-touch knocking on streets where nobody's home.

Set the inspection, don't sell the roof at the door. The door conversation's only job is to earn a roof inspection, not to close a job. Reps who try to diagnose and sell on the porch talk themselves out of inspections. The pitch is some version of: this neighborhood took confirmed hail of a damaging size, a lot of roofs in here have wear that the storm could have pushed over the edge, and I'd like to get up there and document the actual condition for you at no cost. Honest, low-pressure, and it gets you on the roof where the real selling happens.

Measure leading indicators. Track doors-knocked, contacts-made, inspections-set, and inspections-completed per rep per day, not only signed contracts. Contracts lag by days because of adjuster scheduling. If you only watch contracts, you'll fire a good rep who's actually filling the pipeline. Watch the leading metrics and the contracts follow.

Where RoofPredict fits in this workflow

Everything above has a slow, manual version and a faster, data-driven version. RoofPredict is built for the targeting steps — Step 4 and Step 5 above — where the bottleneck is figuring out which roofs inside a confirmed swath are actually worth knocking.

Two things RoofPredict gives a canvassing operation:

First, a roof-age range per address estimated from aerial imagery across a whole area. Instead of guessing roof age from a moving truck or chasing incomplete permit records, you get a per-house estimate that lets you remove new construction and prioritize aging roofs before you build a single route. Because marginal hail events are decided by roof age, this turns the most labor-intensive filter into a list. The honest limit worth stating plainly: it's a range, not a manufacture date or an install certificate. RoofPredict isn't reading a permit; it's estimating a band (say, an 11-to-16-year window) from imagery, and you should treat it as a prioritization signal, not a fact you repeat to a homeowner or write on a contract.

Second, storm modeled per roof, not only where it passed. The phrase we use is that we model the storm on each roof, not only the path it took across the county. Radar swaths and MESH grids tell you the regional core; per-roof storm modeling estimates the energy that each individual roof's slopes and exposures likely took, which is closer to how damage actually distributes house by house. That helps you sequence the genuinely higher-odds addresses to the top of each route instead of treating every house inside a swath polygon as identical.

Used together, the workflow becomes: confirm the event with NOAA/radar data, draw and ground-truth your swath the way you always would, then let per-address roof age and per-roof storm modeling rank the doors inside the confirmed line so your reps spend their hours on the roofs the storm most likely wore out and the roofs that were already aging out.

The honest framing on what this is and isn't: RoofPredict is a targeting and prioritization tool. It tells you which doors are worth knocking and ranks routes. It does not, and cannot, tell you a specific roof is damaged — only a physical inspection establishes that. A storm model gives you odds, not proof. The documentation of actual conditions happens when your inspector gets on the roof, the homeowner decides whether to file, and the insurer decides coverage. RoofPredict gets you to the right ladder faster; it doesn't replace what happens at the top of it.

Reading a roof's hail vulnerability beyond age

Roof age is the single biggest condition variable, but it isn't the only one, and the pros who consistently out-target everyone else stack several signals. When you're deciding which roofs inside a confirmed swath are worth knocking, age tells you a lot, but these factors move the line in ways worth knowing before you knock.

Shingle type and weight. Not all asphalt shingles take hail the same way. Older three-tab shingles are thinner, lighter, and lose granules faster, so a fifteen-year-old three-tab is more vulnerable than a fifteen-year-old laminate architectural shingle. Designer and impact-resistant (Class 4) shingles are specifically built to resist hail bruising and will survive stones that wreck a standard roof; if a neighborhood was built or re-roofed with IR shingles, your conversion will be lower even at age. You can't always tell shingle class from the street, but a generation of newer subdivisions installed with builder-grade three-tab will canvass very differently from a custom-home neighborhood with thick designer shingles.

Slope and orientation. Hail rarely falls perfectly straight down; it usually comes in at an angle driven by the storm's wind. That means the slopes facing into the storm's wind direction take the most energy and the most hits, while the leeward slopes may be nearly untouched. For canvassing this matters in two ways: first, a roof with steep slopes facing the storm direction took more punishment than a low-slope roof of the same age; second, when your inspector gets up there, the storm-facing slopes are where the documentation is. Knowing the storm's wind direction (available from the same radar and surface data that gave you the swath) tells you which roof faces to prioritize and helps you predict which homes on a street took the worst of it.

Prior wear and maintenance. A roof that was already streaking, curling, blistering, or shedding granules before the storm is far more likely to show clear functional damage after marginal hail, because the mat had already lost its flexibility. Neighborhoods with visible deferred maintenance — moss, algae streaking, obvious patching — convert better on marginal events than pristine, well-maintained roofs of the same nominal age.

Exposure and tree cover. Heavily tree-covered lots can have roofs partially shielded by canopy, which both reduces the hail that reached the roof and complicates documentation (debris, shading in photos). Open-exposure homes on the windward edge of a neighborhood often took the cleanest, most-documentable hits. This is a minor factor but worth a mental note when two otherwise-identical streets convert differently.

The practical takeaway: roof-age data gets you most of the way to a good target list, and layering storm direction on top of it (so you know which roof faces took the energy) sharpens it further. This is exactly the gap per-roof storm modeling is meant to close — instead of treating every roof in a swath polygon as equally hit, you account for how each roof's geometry and orientation interacted with the actual storm.

Pre-canvass roof screening: drones, street imagery, and the truck

There's a step between "this address has an old roof in the swath" and "knock this door" that the sharper operations have started inserting: a quick pre-screen of the roof itself before a rep ever walks up. The goal is to weed out the roofs that are obviously not going to inspect well and to give the rep something concrete to open the conversation with.

Aerial and satellite imagery pre-screen. Before routing, a quick look at recent aerial imagery of a target street tells you which roofs are obviously new (bright, uniform, full granule color), which look worn (streaked, patchy, faded), and which have already been recently replaced (a fresh roof on an otherwise older street is a skip — they just got a new roof, they're not filing). You can do a surprising amount of triage from imagery alone before committing a rep's time.

Drone pre-inspection. A licensed drone operator can pre-screen roofs at a rate no climber can match, capturing high-resolution imagery of multiple slopes without anyone on a ladder. For canvassing, drones serve two purposes: they let you triage which roofs actually show hail signature before you invest in a full climb-up inspection, and they keep reps off marginal roofs for the screening pass, which is a real safety win. The limit is that drone imagery shows surface bruising and collateral well but can miss the subtle mat fractures that only a hands-on test square reveals, so a drone pre-screen sets up the climb-up; it doesn't replace it.

The truck pass. The oldest method still works: a slow drive-through with an experienced eye picks up curling, missing shingles, prior patches, obvious storm debris in yards, and tarps from neighbors already in the claim process. A street with three tarps and a competitor's yard signs is a street that's both confirmed-damaged and saturating fast — knock it now or skip it. A truck pass is also how you spot the tell-tale signs that a neighborhood is already deep in the restoration cycle: dumpsters, magnetic-sign trucks, and stacks of bundles on driveways.

Sequence these so the cheap screens come first: imagery triage at the desk, then a truck pass or drone screen on the priority streets, then reps on the doors that survived both filters. Every roof you eliminate before a rep walks up is time that rep spends on a door more likely to convert.

Building a target list when there's no fresh storm

Storm chasing is feast-or-famine, and the best storm-restoration companies don't sit idle between events. There's a second mode of canvassing that doesn't depend on a fresh hail event at all: working aged-out roofs and old hail in your established market.

Here's the logic. Roofs don't only fail from this week's storm. A neighborhood that took a real hail event two or three years ago is full of roofs that were marginally damaged then, never inspected, and have been quietly degrading ever since — the bruised spots have weathered, the granule loss has accelerated, and what was a borderline claim is now a clearer one. Layer on the roofs that are simply aging out of their service life regardless of storms, and you have a year-round canvassing target that nobody is competing for because there's no fresh-storm gold rush drawing crews to it.

To work this, you flip the process. Instead of starting with a storm and filtering by roof age, you start with roof age and historical storm exposure and look for the overlap: neighborhoods where roofs are old enough to be near or past end-of-life and that took a documented hail event in the past few years. Historical storm data (NOAA's Storm Events Database catalogs past events) plus per-address roof age gives you exactly that overlap. These are doors you can knock on a quiet Tuesday in a month with no storms, with a completely honest pitch: this neighborhood took hail a couple years back, a lot of roofs in here are reaching the end of their life, and you'd like to document the current condition.

This off-season targeting is where roof-age data earns its keep even more than during a fresh event, because there's no swath map to lean on — the entire target list is the roof-age-plus-historical-storm overlay. It keeps reps productive in the dead months and builds the kind of steady, non-storm-dependent pipeline that separates companies that survive a slow hail year from companies that fold during one.

Cost, ROI, and how targeting changes the math

It's worth being concrete about why targeting beats volume in dollars, because the instinct under storm pressure is always to flood more doors at the problem. Walk through the unit economics qualitatively.

A canvasser's productive time is the constraint. In a day a rep can physically knock some bounded number of doors — call it sixty to eighty in good conditions. The question isn't how many doors exist; it's what fraction of the doors that rep knocks are worth knocking. If half the doors on an untargeted route are new construction, recently re-roofed, outside the real damage edge, or in a saturated block, then the rep spends half their finite day on doors that cannot convert. Targeting doesn't make the rep knock more doors; it changes the mix so a much larger share of the same number of knocks land on convertible homes.

The effect compounds through the funnel. Every door that can't convert costs a knock, but a marginal door that looks plausible costs much more: it costs the inspection (a climb-up, photos, an hour), the adjuster meeting, and the supplement effort, all of which dead-end when the roof turns out clean or the carrier denies on a cosmetic exclusion. Filtering those out before the rep knocks saves not only the knock but the entire downstream cost of chasing a job that was never going to close. The most expensive lead is the one you fully work and lose.

There's also a saturation-timing cost that pure volume can't address. In the first seventy-two hours of a significant event, the homes are fresh and conversion is high; by day seven the neighborhood is worked and conversion has collapsed. Volume strategy spreads your reps thin across the whole swath, so they reach the high-value quiet-core streets late, after competitors. Targeting lets you concentrate reps on the highest-odds streets while they're still fresh, capturing the conversion premium of being early where it matters instead of being early everywhere and on-time nowhere.

And targeting protects your scarcest resource: good reps. Storm canvassing has brutal turnover, and the fastest way to lose a promising rep is to park them on a dead route where they get nothing but no-damage inspections and slammed doors. A rep on a pre-qualified route wins more, stays in rhythm, and sticks around for the next event. The retention math alone justifies the targeting investment, before you count a single extra contract.

None of this requires inventing numbers to believe. The structure of the funnel makes it true: same knocks, better mix, fewer dead-end inspections, earlier arrival on the good streets, and reps who don't quit. That's where targeting pays.

The legal and ethical guardrails you cannot skip

Storm restoration is one of the most heavily scrutinized corners of contracting, and for good reason — it has attracted enough bad actors that many states passed specific laws aimed at roofers after disasters. Getting the canvass right legally is part of getting it right, period. None of this is legal advice; consult counsel and your state regulators, but here's what you need to have on your radar.

Don't run afoul of insurance-practices laws

Many states have rules, often originating from public-adjuster statutes and unfair-claims-practices acts, that restrict what a roofer can say and do regarding a homeowner's insurance claim. The clean and safe division of roles is this: the roofer documents the roof's condition and provides a repair or replacement estimate; the homeowner owns and files their own claim; the insurer decides coverage. Stay on your side of that line. A contractor inspects and documents and estimates. A contractor does not adjust the claim, does not negotiate coverage on the homeowner's behalf as if they were a licensed public adjuster (unless they actually are one and disclose it), and does not promise a specific outcome.

Things to never say at the door or on the roof

• Don't promise to "get the homeowner a free roof." The roof isn't free; the insurer may pay for a covered loss and the homeowner still owes their deductible. Promising a free roof is both misleading and, in many states, a flag for the kind of conduct anti-fraud statutes target.
• Don't offer to waive, absorb, pay, or "work around" the homeowner's deductible. Rebating or absorbing deductibles is illegal in many states and is insurance fraud. Don't even hint at it.
• Don't present a storm forecast, a radar map, or a hail report as proof that a specific roof is damaged. The data shows hail was probable in the area; only an inspection establishes condition on a given roof. Conflating the two oversells and can cross into misrepresentation.
• Don't claim to act for the insurance company, or imply the carrier sent you.
• Don't pressure a homeowner to sign a contract or an assignment of benefits on the spot under a manufactured deadline. Many states give consumers a cancellation window on storm-related contracts specifically because of high-pressure door tactics; know your state's rescission rules and honor them.

Solicitation, permits, and the local rules

Many municipalities require door-to-door solicitors to register and carry a permit, and some have no-solicitation ordinances and registries you're legally obligated to respect. Get your canvassers permitted where required, give them visible ID and branded gear, and keep them off no-knock lists. It's a small cost that prevents citations and the reputational damage of being the company whose reps got run off by code enforcement.

Honest documentation

When your inspector is on the roof, document what's actually there: date-stamped photos, chalk circles on representative test squares, soft-metal collateral, and an honest count of hits per slope. Honest documentation protects you in the supplement process, protects you if a claim is ever scrutinized, and is frankly the thing that builds the referral reputation that outlasts any single storm. The contractors who are still standing five storm seasons later are the ones whose documentation an adjuster trusts.

Safety: your crew on roofs and in the field

Canvassing season means people on ladders and roofs all day, often steep-slope and often in a hurry. Falls are the leading cause of death in construction, and roofing is among the most dangerous trades by injury rate. OSHA's fall-protection requirements apply, and "we were just doing a quick inspection" is not an exemption. Set a hard rule: nobody goes on a roof they can't safely access and move on, drones and edge-of-eave assessment exist for a reason, and a marginal too-steep or too-wet roof is documented from the ground or not that day. A canvasser is worth more healthy next week than injured this afternoon for a maybe-roof. Train reps on ladder setup, weather (don't put people on roofs in heat-index extremes or when storms are redeveloping), and dog/hostile-homeowner protocols. The cost of one serious fall dwarfs the margin on a dozen roofs.

Common mistakes that quietly kill storm seasons

A grab-bag of the failure modes I see most often, so you can check yourself against them.

Chasing the loud storm instead of the damaging storm. Covered above, but it's the number one error, so it's worth repeating. The neighborhood that's blowing up your phone is often the marginal one. Trust the swath and the soft metals over the volume of inbound noise.

Canvassing new construction. Subdivisions built in the last few years will give you friendly conversations and almost no sellable roofs. If you're not filtering for roof age, you're paying reps to knock roofs that physically can't have taken functional damage from a marginal event.

Saturating the obvious blocks and ignoring the quiet core. Everyone hits the main-road-adjacent neighborhoods first. The damaged-but-quiet interior streets are where the un-worked inventory is on day three. Route accordingly.

No tracking, so reps double-knock and you can't tell good territory from bad. If you don't log door outcomes, you're flying blind. You'll re-knock not-interesteds, miss not-home callbacks, and keep reps parked on dead streets because you can't see the conversion data.

Selling at the door. Reps who diagnose damage and pitch a full job on the porch talk homeowners out of inspections. The door's job is the inspection. Full stop.

Treating the storm map as a damage certificate. A radar swath is a probability surface, not a list of damaged roofs. Walking up to a homeowner and saying "our data shows your roof is damaged" oversells, sets up an awkward inspection when you find a clean roof, and edges toward misrepresentation. Say hail was confirmed in the area and you'd like to document the actual condition.

Ignoring the cosmetic-exclusion reality. If the dominant carriers in your market write cosmetic damage exclusions, a marked-but-functional roof isn't a claim. Know the carrier landscape so you're not selling inspections that can't convert to approved jobs.

Burning reps with bad routes. A canvasser handed eighty random doors converts worse and quits faster than one handed forty pre-qualified ones. Targeting isn't just efficiency; it's retention. Reps stay when they're winning.

A simple pre-deployment checklist

Before you roll trucks on a new event, run this:

1. Confirmed via SPC report and/or radar that stones reached one inch or larger over a populated area? If no, stand down.
2. Drawn the outer swath (1-inch line) and the core (1.75-inch-plus)?
3. Ground-truthed the swath boundary with soft-metal checks at a few edge addresses?
4. Filtered the confirmed swath by roof age, removing or de-prioritizing new construction and recent re-roofs?
5. Sequenced routes by energy and age, with quiet-core priority and saturated/loud-marginal blocks deprioritized?
6. Assigned bite-sized territories, one rep each, with a door-tracking method live?
7. Verified solicitation permits and no-knock compliance for every municipality in the swath?
8. Briefed reps on the script (set the inspection, don't sell or diagnose at the door) and the do-not-say list (no free roof, no deductible games, no forecast-as-proof)?
9. Confirmed fall-protection and weather/safety protocols for inspectors?
10. Set leading-indicator tracking (doors, contacts, inspections-set, inspections-done) so you can manage the week, not only count contracts at the end of it?

If you can check all ten, you're deploying a targeted operation instead of a swarm, and the difference will show up in cost-per-signed-job by the end of the event.

Putting it together

Finding hail-damaged homes to canvass comes down to refusing to treat "the storm hit this area" as a single fact. It's two questions — where did damaging hail actually fall, and which roofs in that footprint were vulnerable enough to take functional damage — and the contractors who separate them and answer both win the season. Use NOAA and radar data to draw and confirm the swath. Use soft-metal ground truth to find the real edge. Use roof-age and per-roof storm modeling to rank the doors inside the line so your reps spend their finite hours on the roofs the storm most likely wore out and the roofs that were already aging out anyway. Stay scrupulously clean on the legal and ethical side, keep your people safe, and track leading indicators so you can manage in real time.

That's the whole game: fewer, better doors, knocked in the right order, before the neighborhood saturates. If you want to compress the targeting steps — roof age per address and storm modeled per roof so you can route to the highest-odds doors first — that's exactly what RoofPredict is built to do. It won't tell you a roof is damaged (only your inspector and the homeowner's claim can establish that), and roof age comes as a range, not a date. What it does is get your crew to the right ladders faster, so the short storm season actually pays. Visit https://roofpredict.com/ to see which roofs in a storm footprint are worth knocking first.

FAQ

What hail size actually damages a roof enough to canvass for?

Functional damage to asphalt shingles generally becomes likely around one inch (quarter size), which is also the National Weather Service severe-hail threshold. Stones from 1.0 to 1.5 inches produce marginal-to-clear damage depending on roof age and brand, while 1.75 inches (golf ball) and up cause widespread functional damage on most roofs. Pea- and dime-sized hail rarely produces a sellable roof on asphalt. Below one inch, temper expectations regardless of how loud the storm was.

Where do I get free hail data to find which areas got hit?

Start with the NOAA Storm Prediction Center daily storm reports, which list hail size and time for reported events. Pair that with radar-derived products like MESH (Maximum Estimated Size of Hail) if you have access through a weather service or vendor. SPC reports confirm an event happened and give a rough size ceiling; radar gives you the spatial swath. Always ground-truth the boundary with your own soft-metal checks before deploying.

Why do crews waste time canvassing the wrong neighborhoods?

The most common error is chasing where the storm was loud instead of where it did damage. A cell can drop terrifying small hail and high wind over one neighborhood with zero functional roof damage, while a quieter cell nearby dropped damaging stones. The loud neighborhood is primed to talk but won't convert; the quiet damaged one is the real opportunity. Trust the radar swath and soft-metal ground truth over the volume of inbound calls.

How does roof age affect whether a roof took hail damage?

Roof age is often the deciding factor on marginal events. A new architectural shingle has a flexible mat and tight granule embedment and shrugs off marginal hail, while a fifteen-plus-year-old roof of the same brand has lost granules and gone brittle, so the same stone that bounced off the new roof fractures the old one. Filtering a confirmed swath by roof age lets you skip new construction and prioritize aging roofs, which dramatically raises your inspection-to-contract rate.

How do I find roof age across a whole storm area quickly?

Traditional methods (county permit records, MLS photos, eyeballing from the truck) are slow and incomplete, and they don't scale to a metro-wide event in the seventy-two-hour window when speed matters. Aerial-imagery analysis estimates a roof-age range per address across an entire footprint, letting you remove new construction and prioritize older roofs before you build routes. Treat the result as a prioritization range, not a manufacture date or install certificate.

What's the difference between a hail swath map and per-roof storm modeling?

A swath or MESH map shows the regional hail core — roughly where the largest stones fell across a county. Per-roof storm modeling estimates the energy each individual roof's slopes and exposures likely took, which is closer to how damage actually distributes house by house. The swath tells you the neighborhood; per-roof modeling helps you rank the specific addresses inside it so your reps knock the highest-odds doors first instead of treating every house in the polygon as identical.

Can I tell a homeowner their roof is damaged based on storm data?

No. Storm data shows hail of a certain size was probable in the area; only a physical inspection establishes whether a specific roof is actually damaged. Presenting a forecast, radar map, or hail report as proof of damage oversells and can cross into misrepresentation. The honest door pitch is that confirmed damaging hail fell in the neighborhood and you'd like to document the roof's actual condition at no cost.

What can't I say to a homeowner when canvassing after a storm?

Don't promise a free roof (the homeowner still owes their deductible), don't offer to waive, absorb, or work around the deductible (illegal and fraudulent in many states), don't present storm data as proof of damage, don't imply the insurance company sent you, and don't pressure a same-day signature under a manufactured deadline. The clean division of roles: the roofer documents condition and estimates, the homeowner owns and files the claim, and the insurer decides coverage. Consult your state regulators and counsel for specifics.

Do I need a permit to door-knock for storm restoration?

Often yes. Many municipalities require door-to-door solicitors to register and carry a permit, and some maintain no-solicitation ordinances and registries you're legally required to respect. Permit your canvassers where required, give them visible branded ID, and keep them off no-knock lists. Check the rules for every municipality in your swath before deploying; citations and reputational damage from code-enforcement run-ins cost far more than the permits.

How should I manage canvassers to avoid wasting their time?

Hand each rep a pre-qualified, turn-by-turn route of specific addresses worth knocking, not a heat map. Assign bite-sized territories of a few hundred doors, one rep each, and log every door's outcome to prevent double-knocking and enable not-home callbacks. Knock during high-contact hours (weekday late afternoon, weekend midday), make the door's only goal setting an inspection rather than selling, and track leading indicators (doors, contacts, inspections set and completed) since contract counts lag by days.