How to Use Storm Data to Plan Roofing Canvassing That Actually Closes

Most roofing companies treat storm data like a starting gun. The radar lights up, a hail report shows 1.5-inch stones over the north side of town, and within hours there are trucks rolling into the affected ZIP codes with no plan beyond "work the area." That approach made money fifteen years ago when fewer crews paid attention to weather. Today it mostly means six companies fighting over the same three streets while entire pockets of legitimately damaged, aging roofs sit untouched on the other side of the swath.

Storm data is not the edge anymore. Everybody has a hail map. The edge is what you do with the data: how you turn a messy swath outline and a pile of wind reports into a ranked list of doors your crew can work in order, with a route that respects drive time, a script that matches what actually hit each roof, and a way to skip the houses that will waste your reps' afternoons. That is the difference between a storm season that funds your year and one that breaks even after you pay for the gas and the door hangers.

What follows is the operational version. Not "download a hail map and go." The actual workflow restoration teams use to read the data, qualify the territory, build the door list, and put canvassers in front of the homeowners most likely to have a roof an adjuster will write. There are worked examples, the numbers that matter, the edge cases that cost crews days, and an honest accounting of where the data stops being useful and judgment takes over.

What "storm data" actually means (and which parts you can trust)

When a roofer says "storm data," they usually mean a single hail map from one vendor. That is the thinnest possible version. Real storm data is a stack of independent sources, each measuring something different, each with its own error bars. You get better targeting when you understand what each layer is actually telling you and where it lies to you.

The five layers, ranked by how much they should drive your decisions

1. Ground-truth hail and wind reports. The National Weather Service collects storm reports from trained spotters, law enforcement, and the public, and the Storm Prediction Center compiles them into a daily storm report log with point locations, estimated hail size, and measured or estimated wind gusts. These are the closest thing to ground truth you can get for free. The catch: they are sparse and biased toward population. A golf-ball hailstone that falls in a field next to a subdivision may never get reported, while the same stone over a busy intersection gets three reports. Use these as anchors, not as coverage maps.

2. Radar-derived hail estimates (MESH). Maximum Estimated Size of Hail is a radar product that infers hail size from the vertical structure of a storm. Commercial vendors repackage MESH into the smooth colored swaths you see on hail maps. It covers everywhere radar covers, which is its strength, but it is a model, not a measurement. MESH overestimates in some storm types and underestimates in others, and it cannot tell you whether a given stone actually struck a given roof at a damaging angle. Treat the swath as "where damage is plausible," not "where damage happened."

3. Wind field data. Straight-line wind and gust estimates come from surface stations, radar velocity, and modeled wind fields. Wind damage is more about sustained gusts and direction than a single peak number. A roof on the windward slope of a hill facing a 70 mph gust sees a very different load than the identical roof in the lee. Wind data is coarser than hail data, so you lean harder on roof-by-roof factors (orientation, exposure, age) when wind is the driver.

4. Roof age and condition. This is the layer most crews skip, and it is the one that decides whether a storm-touched roof actually converts. A 4-year-old architectural shingle and a 19-year-old three-tab can take the identical 1.25-inch hail on the same block and end up in completely different places: one sheds it with bruising you have to hunt for, the other loses granule mat and cracks at the mat into the kind of functional damage an adjuster writes. Storm data tells you where the energy went. Roof age tells you which roofs the energy actually wore out.

5. Property and ownership data. Owner-occupancy, year built, lot characteristics, and parcel boundaries. This is the layer that keeps you from knocking 40 rental doors where the owner lives three states away, or wasting a Saturday on a new build that took the same hail your aging targets did but will not produce a claim.

The mistake almost everyone makes is stopping at layers one and two. They get the swath, they see the reports, they go. The teams that consistently out-earn them fuse all five, and they weight roof age and condition far higher than the raw storm intensity, because a moderate storm over old roofs beats a violent storm over new ones every single time.

A quick reliability table

Why "work the swath" is the wrong unit of work

A hail swath can cover 40 square miles. Inside it might be 30,000 rooftops. If your plan is "work the swath," you have not made a plan, you have drawn a box around a problem. The swath is the territory. The unit of work is the door, and the doors inside any swath are wildly unequal.

Consider a real-shaped example. A swath shows 1.5-inch MESH across a band roughly three miles wide and eight miles long. That band contains:

• A 1990s subdivision with original or first-generation roofs now 20 to 30 years old.
• A 2019 master-planned community with roofs four to six years old.
• A commercial corridor with flat membrane roofs that do not behave like shingles at all.
• A stretch of custom homes with metal and tile that respond to hail completely differently.
• A pocket of rentals owned by an out-of-state LLC.

The same storm passed over all of it. The conversion potential is not even close. The 1990s subdivision is where the money is, because that is where storm energy met roofs that were already at or past their service life. The 2019 community will eat a week of your reps' time and produce a handful of marginal inspections. If your canvassers walk the swath "top to bottom" the way they were handed it, they spend most of their hours on the lowest-yield doors.

The job, then, is to take the swath and rank the doors inside it before anyone gets out of the truck. That ranking is the entire game.

The end-to-end workflow

Here is the sequence experienced storm teams run, from the first radar ping to a canvasser standing on a porch. Each step narrows the universe and raises the quality of the door.

Step 1: Confirm the event is real and worth chasing

Not every flashy radar return is a job. Before you commit gas and payroll, verify:

1. Pull the SPC storm reports for the date and county. Look for actual hail-size reports of 1 inch or larger, or measured wind gusts of 58 mph or more (the severe thresholds). A swath with zero ground reports is a yellow flag, not a green light.
2. Cross-check at least two sources. If a vendor swath shows 2-inch hail but the NWS log shows nothing above pea-size and no wind reports, something is off. Believe the convergence of independent sources, not the most dramatic single one.
3. Check the storm type. A fast-moving line that dropped marginal hail for two minutes is a different animal than a slow supercell that sat over a neighborhood pulsing 1.75-inch stones for fifteen. Duration and stone size together drive whether shingles took functional damage versus cosmetic spotting.

Worked example: A March storm shows a tidy 1.25-inch swath on the vendor map across the east side. SPC log shows two hail reports, 1 inch and 1.25 inch, both near the swath centerline, plus a 61 mph wind gust. That is a confirmable, workable event. You proceed. A different swath the same week shows 1-inch hail but no ground reports and no wind, over a brand-new subdivision. You let that one go.

Step 2: Draw the candidate area, then immediately shrink it

Take the confirmed swath and overlay it on a map. Now start subtracting:

• Cut the new construction. Anything built in roughly the last 8 to 10 years is unlikely to have a roof an adjuster will write for hail short of severe stone sizes. Year-built data gets you most of the way here.
• Cut the wrong roof types for the storm. If you do not run commercial flat work or you do not do tile and metal, those parcels are not your candidates today even though they sit in the swath.
• Cut the obvious dead-end ownership. Large rental clusters under one out-of-area owner are low-yield for canvassing because the decision-maker is not behind the door.

This is usually where a swath of 30,000 roofs becomes a candidate set of 6,000 to 9,000. You have not ranked anything yet. You have just removed the doors that are nearly guaranteed to waste time.

Step 3: Rank the surviving doors by likelihood of a writable roof

Now you rank what is left. The single best predictor inside a confirmed swath is roof age and condition, because age determines how a roof responds to a given impact. The ranking factors, in rough order of weight:

1. Estimated roof age (as a range). Roofs in the 15-plus-year band on standard asphalt shingles are your top tier. Roofs in the 10-to-15 band are second tier. Under 8 to 10 years, deprioritize unless stone size was large.
2. Storm intensity at that specific address, rather than merely "inside the swath." Proximity to the swath centerline and to confirmed ground reports matters.
3. Roof orientation and exposure relative to the wind/hail vector. The slope facing the storm took more energy.
4. Owner-occupancy. A homeowner behind the door converts better than a tenant who has to call a property manager.
5. Prior roof generation. Three-tab and older architectural shingles fail more visibly than newer impact-rated products.

The output of this step is a list sorted so that door #1 is the address most likely to have an aging roof that the confirmed storm actually wore out, and door #4,000 is the marginal one you may never reach. That ranked list is what you hand to dispatch and canvassers, not the raw swath.

Step 4: Cut the ranked list into walkable routes

A ranked list is not yet a route. A canvasser cannot teleport between the best doors scattered across eight miles. You have to balance door quality against drive and walk efficiency.

The practical method:

• Cluster the top-tier doors geographically. Find the neighborhoods where high-ranked doors are dense. A block where 18 of 24 homes are 1990s-vintage roofs inside the confirmed swath is gold, because a canvasser works it on foot with no windshield time.
• Sequence clusters by yield, not by distance from the office. Start reps in the densest high-tier cluster, even if it is across town.
• Set a door count per rep per day and size the route to it. A canvasser working detached suburban homes realistically gets through 60 to 120 doors a day depending on contact rate, conversation length, and walkability. Cul-de-sac suburbs are faster than long rural lots.
• Leave the low-density top-tier doors for windshield routes worked by your stronger closers who can justify the drive.

Step 5: Match the script to what hit that roof

This is where storm data stops being a map and becomes a conversation. A canvasser who knocks a wind-damaged door with a hail pitch sounds like every other truck in the neighborhood. A canvasser who says, in effect, "the line that came through Tuesday hit this side of the street with the strongest gusts, and on roofs the age of yours the wind tends to lift and crease shingles along the ridge," sounds like someone who knows the specific roof.

More on the script later. The point for the workflow is that the door list should carry, per address, what kind of storm hit it so the rep can lead with the right damage mechanism.

Step 6: Inspect, document, and let the process work

The canvasser's job is to earn an inspection and document conditions honestly. From there the roofer documents the roof's condition and provides an estimate; the insurer decides coverage; the homeowner owns and files their own claim. Storm data got you to the right porch. It does not decide the claim, and pretending otherwise on a doorstep is how companies get themselves in trouble. Keep the doorstep honest: you are there because the data suggests this roof may have storm-related wear worth a closer look, not because anyone is owed anything.

Reading hail and wind data without fooling yourself

The biggest waste in storm canvassing comes from believing the prettiest map. Smooth, confident color swaths feel authoritative. They are models. Here is how to read them like a practitioner instead of a tourist.

Hail size thresholds that actually matter for shingles

Stone size is not linear in its effect on asphalt shingles. A rough field-practitioner framing:

Notice that age moves the threshold. The same 1-inch hail that is cosmetic on a 5-year-old roof can be functional on a 22-year-old roof whose granule mat is already thin and whose shingles are brittle. This is exactly why a storm map alone is a weak targeting tool and a storm map fused with roof age is a strong one.

Why the swath edges lie

MESH swaths are smoothed. The real hail footprint is patchy, with streaks and gaps driven by storm motion and updraft pulses. Two consequences:

1. Doors just inside a swath edge may have seen nothing, while a streak a mile outside the colored boundary got hammered. Always weight confirmed ground reports over the swath outline near edges.
2. The centerline is your friend. Damage density is usually highest along the swath core and degrades toward the edges. When you rank doors, proximity to the core is a real signal.

Wind is about direction, not only speed

A single peak-gust number tells you little about which roofs got worked. What matters:

• Direction of the strongest gusts. Windward slopes take the load; leeward slopes are sheltered. A door list that ignores orientation treats every house the same when the storm did not.
• Exposure. Homes on ridgelines, at the edge of a development facing open ground, or with tall mature trees removed nearby take more wind than identical homes tucked into a sheltered interior block.
• Duration and gust cycling. Sustained buffeting fatigues sealant strips and lifts tabs; a single gust may do less than a long pulse of repeated 55 to 65 mph gusts.

A practitioner reading wind data asks "which slopes, on which streets, faced the strongest sustained gusts," not "what was the highest number on the map."

Building the ranked door list: a concrete worked example

Walk through a realistic build so the abstraction lands.

The event. A late-spring supercell. SPC log shows three hail reports (1.25, 1.5, 1.5 inch) clustered along a northeast-moving track, plus two wind gusts (59 and 64 mph). The vendor swath shows a 1.25-to-1.5-inch band about two miles wide and six miles long crossing the north side of a mid-size metro.

Raw universe. The swath contains roughly 11,000 rooftops.

Subtract step (Step 2):

• Remove ~3,200 roofs built since 2016 (two newer master-planned communities).
• Remove ~900 commercial and multifamily flat roofs (the company does steep-slope residential only).
• Remove ~600 rental-cluster doors under two out-of-area LLCs.
• Candidate set: ~6,300 doors.

Rank step (Step 3):

• Tier 1 (top priority): ~1,400 doors that are owner-occupied, on asphalt shingles estimated at 15-plus years, within half a mile of the swath core or a confirmed ground report.
• Tier 2: ~2,100 doors, roofs 10 to 15 years, owner-occupied, anywhere in the confirmed band.
• Tier 3: ~2,800 doors, younger roofs, edge-of-swath, or mixed occupancy. Work only if you run out of Tier 1 and 2.

Route step (Step 4):

• Tier 1 doors cluster heavily in two 1990s subdivisions on the north and northeast. Those become foot-canvass routes for the first three days, sized at ~90 doors per rep per day.
• A scattered set of ~300 Tier 1 doors in custom-home pockets goes to two senior closers as windshield routes.
• Tier 2 becomes the following week's plan; Tier 3 is contingency.

Script step (Step 5):

• The two subdivisions sit on the windward (southwest-facing) side of the track, so reps lead with combined hail-and-wind language for southwest-facing slopes, and carry the confirmed stone sizes and gust numbers so they can speak to the specific event.

The outcome of all that: instead of dropping ten reps into an 11,000-roof swath to wander, you put them on roughly 1,400 of the best doors first, clustered for foot efficiency, with a script tuned to the actual storm vector. That is what "using storm data to plan canvassing" means in practice. The map was step one of six.

Where the roof-age layer comes from, and how RoofPredict fits

Everything above hinges on one layer most crews cannot easily get: reliable, address-level roof age and condition fused with the storm that actually hit each roof. Hail maps are commodities. Year-built data from public records tells you when the house was built, not when the roof was last replaced, which means every re-roof since construction is invisible to it. A 1994 house with a 2021 roof and a 1994 house with its original roof look identical in county data, and they are opposite ends of your priority list.

This is the gap RoofPredict is built to close. It reads aerial imagery to estimate a roof's age as a range per address (not an exact date, because nobody can read a precise install date off a photo), and it models the storm physics on each individual roof rather than just telling you where the swath passed. Instead of a hail map that shows where it hailed, you get a per-roof picture of which roofs the storm likely wore out, paired with how old each roof already was. That is exactly the fusion the workflow above depends on: the storm layer and the roof-age layer joined at the address, so your door list is ranked by which roofs are genuinely due rather than by which ZIP the radar lit up.

Used inside the workflow, that looks like this. After you confirm the event (Step 1) and draw the candidate area (Step 2), instead of hand-cutting new construction with crude year-built data, you get a per-address roof-age range and a per-roof storm-impact read for the whole swath. Step 3, the ranking, stops being a guess and becomes a sort: the aging roofs the storm actually hit float to the top, the young or re-roofed homes drop, and your reps walk the highest-yield doors first. It also makes a green canvasser sound like a veteran, because the per-home talking point is sitting in their hand before they climb anything.

The honest limits matter, and a tight trade compares notes, so here they are plainly. A roof-age estimate from imagery is a range, not a guaranteed install date; you will occasionally find a recent re-roof the imagery had not caught up to, or an old roof that was patched rather than replaced. The storm model gives you odds that a given roof took damage, not proof; the canvasser still has to get on the roof and document what is actually there, and the adjuster still decides coverage. RoofPredict is not a lead service and does not hand you customers; it sharpens the outbound you already do by telling you which doors are worth the knock. If anyone sells you storm targeting as a guarantee of damage or a guarantee of a claim, walk away. The value is in concentrating your reps' hours on the right roofs, which is a large and real edge, not in eliminating the inspection.

The canvassing script: matching the door to the storm

A ranked, routed door list puts your rep on the right porch. The conversation closes the inspection. The teams that convert best treat the script as a function of the data, not a memorized speech.

Lead with the specific storm, not a generic pitch

Homeowners in a hit neighborhood have already had three trucks knock. The generic "we noticed storm damage in your area" pitch is now background noise. A rep armed with the data can be specific:

• Reference the actual date of the storm.
• Reference the actual mechanism that hit their slope (the southwest-facing wind load, the 1.5-inch hail along this street).
• Reference, honestly, why this roof in particular is worth a look: its age range plus the confirmed storm. "Roofs on this block are in the range where this kind of hail tends to do more than it looks like from the curb" is a true, specific, non-hype opener.

Keep it honest and keep it legal

This is where companies get themselves in trouble, so be deliberate. The canvasser's job is to earn a free inspection and document conditions accurately. What a rep should never do on the doorstep:

• Promise the homeowner a "free roof" or that insurance "will" pay. The insurer decides coverage; you do not.
• Make any promise about the deductible.
• Present a storm forecast or a swath map as proof the roof is damaged. It is odds and a reason to inspect, not evidence of a claim.
• File, handle, or negotiate the claim on the homeowner's behalf in ways that cross into public-adjusting. The roofer documents the roof's condition and provides an estimate; the homeowner files and owns their claim; the insurer decides.

Many states regulate this tightly, and several have rules about contractors and insurance claims, deductible rebating, and what a contractor may do on a storm door-knock. Know your state's department of insurance rules and your state's contractor and consumer-protection statutes before you write a script. Honesty is not only the ethical position, it is the durable sales position: the rep who says "the storm is a reason to look, not a guarantee" out-converts the rep who over-promises and gets shut down by a skeptical homeowner who has heard the pitch.

A doorstep flow that respects the data

1. Open with the specific storm and date. Establish you are not a generic canvasser.
2. Tie it to their roof's age range. Specific and honest.
3. Offer the free inspection, framed as documentation, not a promise of payout.
4. If they decline, leave a branded report or QR that shows what the data suggests about their roof, so you stay in their head for the day they are ready.
5. Log the outcome (no-knock, not home, declined, inspection set) so the route data improves.

A quick do/don't for storm canvassing language

Routing and crew logistics: turning the list into worked doors

A brilliant ranked list dies if the logistics are sloppy. The operational layer is where storm seasons are won or lost on margin.

Door density beats door quality past a point

A Tier 1 door 25 minutes from the next Tier 1 door is worth less, in dollars per hour, than a Tier 2 door three houses away. Drive time is the silent killer of canvassing economics. When you cut routes, optimize for clustered density first, then quality within the cluster. The math is simple: a rep who knocks 90 doors a day at a 30 percent contact rate has 27 conversations; a rep who spends half the day driving to better doors and knocks 45 has 13 or 14. The denser route usually produces more inspections even if the individual doors are slightly lower tier.

Size routes to realistic door counts

Plan with honest throughput numbers and adjust to your market:

These are doors attempted, not contacts. Contact rate (someone answers) typically runs 25 to 40 percent depending on time of day and neighborhood, which is why afternoon and early-evening windows matter.

Sequence the day around contact rate

Morning hours have low contact rates in many residential areas. A common pattern: use mid-morning for the lower-tier or windshield doors where you expect fewer answers anyway, and concentrate the highest-tier clusters in the late-afternoon and early-evening window when people are home. You are spending your best doors during your best contact hours.

Track outcomes per door and feed it back

Every knock should produce a logged outcome. Over a season, that data tells you which neighborhoods, roof-age bands, and storm types actually convert for your crews, and it sharpens the next ranking. Teams that log nothing repeat the same misallocations storm after storm.

Common mistakes that quietly cost crews money

A list of the failure modes that show up again and again, and what to do instead.

Chasing the biggest swath instead of the best swath

The most violent storm of the month is also the one every competitor in three states is driving toward. A moderate storm over a city full of aging roofs, with fewer trucks chasing it, often produces a better season than a baseball-hail event swarmed by out-of-town crews. Rank events by age-of-roofs-times-intensity-divided-by-competition, not by stone size alone.

Treating year-built as roof age

County records tell you when the house was built. Re-roofs are invisible to them. Crews that target purely on year-built waste time on recently re-roofed homes and miss old houses with original roofs that look young in the records. Roof age has to come from something that sees the actual roof.

Working the swath "top to bottom"

Handing a canvasser the raw swath and saying "start at the north end" guarantees they spend most of their day on average doors. Always work the ranked, clustered list, not the geographic outline.

Ignoring orientation on wind events

Reps who knock both sides of a street identically on a wind-driven storm waste half their pitches. The windward slopes took the load. The data should route reps to the slopes that actually got worked.

Over-promising on the doorstep

The fastest way to torch a neighborhood's trust (and invite regulatory trouble) is reps promising free roofs and guaranteed payouts. It converts worse, not better, with homeowners who have already been pitched, and it puts the company at legal risk. Honest, specific, data-grounded pitches win the long game.

Letting the data go stale

A door list built off a three-week-old storm, worked after every competitor has already knocked it twice, is a different (and worse) asset than the same list worked in the first days. Storm targeting has a freshness curve. The roof-age targeting, by contrast, holds up between storms, which is the point of working aging roofs even in quiet weeks.

Skipping the post-storm "quiet" targeting entirely

Many crews go dormant between storms and starve. The same roof-age data that ranks doors inside a swath also ranks the aging roofs across your whole service area on a calm week. Storm data plus roof age is a storm-season tool; roof age alone is a year-round one. The companies that stay busy use both.

Reading storm type, not only storm size

Two storms can drop the same peak hail size and leave completely different damage on the ground. The storm's structure and behavior decide how much energy a roof absorbed, and a practitioner reads that into the targeting.

Supercells versus squall lines versus pulse storms

Supercells are the discrete, rotating storms that produce the largest, most damaging hail and often sit or track slowly. A supercell that crawls over a neighborhood pulses repeated large stones in the same place, so the damage is concentrated and severe along a relatively narrow path. These are high-yield events where they hit aging roofs, because the energy delivered to a given roof is high. Your targeting tightens around the core track.

Squall lines and bow echoes are long lines of storms driven by wind. They produce widespread straight-line wind damage and often smaller, fast-moving hail. The damage signature is wind-led: lifted and creased shingles, missing tabs, debris impact, with hail as a secondary factor. Targeting on these events leans heavily on wind direction and exposure rather than the hail core.

Pulse or air-mass storms are the common summer afternoon storms that flare up and collapse. They can drop brief small-to-moderate hail over a limited footprint. These are marginal events for most roofs and only worth working where they overlap genuinely old roofs near a confirmed report.

Why duration is the underrated variable

A roof that takes 1.25-inch hail for two minutes and a roof that takes 1.25-inch hail for twelve minutes are not in the same place afterward, even though a swath map labels both "1.25 inch." Repeated impacts on the same shingle compound granule loss and mat fracture. Storm motion (how fast the cell moved) is the best available proxy for duration: slow-moving cells dwell, fast lines pass. When you read the data, a slow-mover over old roofs is a stronger signal than a faster cell of the same peak size. This is one more reason the per-roof storm read matters more than the headline number on the map.

Cross-referencing and quality-checking your data

The discipline that separates pros from tourists is verification. Before a single rep walks, the data should survive a few sanity checks.

Triangulate every event

Never act on a single source. The reliable pattern is to require agreement across at least two independent layers before committing crews:

1. Vendor swath says where and how big.
2. NWS/SPC ground reports confirm something real happened at points inside it.
3. Local signal (social media photos of hail, news reports, your own driving) confirms it on the ground.

When all three agree, you have a confirmable event. When the swath is dramatic but the ground reports and local signal are silent, you have a model artifact, and you let it go. Crews that skip this step burn entire weeks on storms that barely happened.

Watch for the population-density bias

Ground reports cluster where people are. A hailstorm over a sparsely populated stretch may show almost no reports while having dropped serious stones, and a marginal storm over a dense suburb may show many reports for small hail. Adjust mentally: in low-density areas, lean on the radar swath and your own ground verification; in high-density areas, do not let a pile of small-hail reports talk you into a non-event.

Re-verify roof age against reality on the ground

Whatever your roof-age source, treat early knocks as calibration. If your reps keep finding recently re-roofed homes flagged as old, or genuinely old roofs flagged as young, the data needs a correction or the imagery is stale for that area. The first day or two of any campaign is partly a field test of the data, and good teams adjust the ranking when the ground tells them something different.

The economics: what good targeting is actually worth

It helps to put rough numbers on why this work pays for itself, using conservative, illustrative figures (plug in your own).

A simple cost-per-inspection comparison

Say a canvasser is fully loaded at a certain daily cost and attempts 90 doors a day. Compare two approaches over a five-day week with ten reps:

Unranked "work the swath": Reps wander the geographic outline. A large share of their doors are new construction, wrong roof types, or dead-end ownership. Suppose 35 percent of attempted doors are low-yield. Effective productive doors per rep drop, contact-to-inspection conversion is lower because pitches do not match the roof, and the crew sets fewer inspections per week.

Ranked, routed, storm-matched: The same reps work clustered top-tier doors first, with pitches matched to the actual storm. Fewer wasted doors, higher conversion per conversation, less windshield time. Even a modest lift, say moving from a low conversion rate to a moderate one, plus cutting wasted doors, can change the inspections-per-week number substantially without adding a single rep.

The point is not a magic multiplier. It is that your most expensive resource is rep-hours, and targeting redirects those hours from the worst doors to the best ones. The cost of better data is small against the cost of paying ten people to knock the wrong houses for a week.

The hidden cost: rep churn

There is a second economic effect that owners feel but rarely quantify. A green canvasser who gets handed the swath and walks new-construction doors all day gets no traction, makes no money, and quits inside a month. A green canvasser handed the right doors with a specific talking point sets inspections, earns commission, and stays. Reducing rep churn by giving people winnable doors is one of the highest-leverage uses of good targeting, because the cost of recruiting and training a replacement dwarfs the cost of the data.

Stack the tooling sensibly

A workable storm-canvassing stack usually has four parts, and you do not need to overbuy any of them:

The storm data and the canvassing app are common. The roof-age-and-storm-fusion layer is the one most crews are missing, and it is the one that does the heavy lifting in the ranking step.

A storm-canvassing readiness checklist

Before you launch reps into a fresh event, run the list:

• Event confirmed against at least two independent sources (vendor swath plus NWS/SPC reports).
• Storm type, stone size, and wind direction noted.
• Candidate area drawn and trimmed of new construction, wrong roof types, and dead-end ownership.
• Doors ranked by roof-age range first, storm intensity at the address second, orientation and occupancy after.
• Top-tier doors clustered into walkable foot routes; scattered top doors assigned to closers as windshield routes.
• Routes sized to realistic per-rep door counts for the neighborhood type.
• Day sequenced so the best clusters land in the best contact-rate hours.
• Scripts matched to the storm mechanism and the roof-age range per cluster.
• Script reviewed for legal/compliance language (no free-roof, no deductible, no payout, no claims handling).
• Outcome-logging set up per door for feedback into the next ranking.
• Knowing your state department-of-insurance and contractor rules before the first knock.

How the pieces fit together

The through-line of all of this is one idea: a hail map is the least important input, even though it is the one everyone fixates on. The map tells you where a storm passed. It does not tell you which roofs it wore out, which homeowners are behind the doors, which slopes took the load, or where your reps should physically stand. The contractors who out-earn the swarm are the ones who treat the storm map as a single layer, fuse it with roof age and condition and ownership, rank the doors, cluster them into efficient routes, and put a green-or-veteran canvasser on the best porch first with a specific, honest, storm-matched pitch.

Storm data did not make targeting harder; the volume of crews who now have the same maps did. The differentiator moved one layer down, to the address-level fusion of how old this roof is and what this storm actually did to it. Get that layer right and the rest of the workflow, the routing, the scripting, the logistics, compounds on top of it. Get it wrong, and you are a faster truck arriving at the same wrong doors as everybody else.

Work the doors the storm actually wore out, in the order they deserve, with a pitch that respects both the homeowner and the law. That is what it means to use storm data to plan canvassing, and it is a skill, not a download.

FAQ

What is the most important piece of storm data for planning canvassing?

Counterintuitively, it is not the hail map. The single best predictor of whether a storm-touched roof converts is the roof's age and condition, because age determines how a roof responds to a given impact. A moderate storm over 20-year-old roofs beats a violent storm over 5-year-old roofs almost every time. Use the hail and wind data to define where damage is plausible, then rank the doors inside that area by roof age first.

How reliable are the hail swath maps roofers buy?

They are directional, not precise. Most commercial swaths are built on radar-derived MESH estimates, which model hail size from a storm's vertical structure rather than measuring what actually hit the ground. Treat the swath as 'where damage is plausible,' weight confirmed NWS/SPC ground reports more heavily near the edges, and remember the real hail footprint is patchy and streaky even when the colored swath looks smooth.

Why can't I just use the year a house was built to find old roofs?

Because county year-built records tell you when the house was built, not when the roof was last replaced. Every re-roof since construction is invisible in that data. A 1994 house with a 2021 roof and a 1994 house with its original roof look identical in public records but sit at opposite ends of your priority list. You need roof age that comes from something that actually sees the roof, such as an estimate from current aerial imagery.

How many doors should a canvasser realistically knock per day?

It depends on the neighborhood. Dense suburban cul-de-sacs allow 90 to 130 doors attempted per rep per day; standard detached suburban runs 60 to 100; large-lot or semi-rural drops to 25 to 50 and should go to your closers as windshield routes. Those are doors attempted, not contacts; contact rates typically run 25 to 40 percent, which is why late-afternoon and early-evening hours matter.

Should I chase the biggest storm or a smaller one?

Not always the biggest. The most violent storm is also the one every out-of-town crew is racing toward, so you fight a swarm for the same streets. A moderate storm over a city full of aging roofs, with fewer competitors chasing it, often produces a better season. Rank events by roof age times intensity divided by competition, not by stone size alone.

How do I match my canvassing pitch to the storm data?

Carry, per address, what kind of storm hit that roof so the rep leads with the right mechanism. On a wind event, route reps to the windward slopes and talk about lifting and creasing; on a hail event near the swath core, talk about granule loss and bruising on aging roofs. Reference the actual storm date and stone size, and tie it to the roof's age range. Specificity separates you from every generic 'we noticed damage in your area' truck.

What can RoofPredict do that a hail map can't?

A hail map shows where a storm passed. RoofPredict estimates each roof's age as a range from aerial imagery and models the storm physics on each individual roof, so you get a per-address read of which roofs the storm likely wore out, paired with how old each roof already was. That fusion is what lets you rank doors by which roofs are genuinely due rather than by which ZIP the radar lit up. It is not a lead service and it does not replace the inspection; the estimate is a range and the storm model gives odds, not proof.

Is it legal to tell a homeowner the storm damaged their roof?

Be careful with that framing. A storm forecast or swath map is a reason to inspect, not proof of damage, and you should present it that way. The roofer documents the roof's actual condition and provides an estimate; the homeowner files and owns their claim; the insurer decides coverage. Never promise a free roof, a payout, or anything about the deductible, and do not handle or negotiate the claim in ways that cross into public adjusting. Many states regulate this tightly, so check your state department of insurance and contractor rules before writing a doorstep script.

How fresh does storm data need to be to be useful for canvassing?

Storm targeting has a freshness curve. A door list built off a three-week-old storm, worked after competitors have already knocked it twice, is a much weaker asset than the same list worked in the first days. Storm-driven canvassing rewards speed and organization. Roof-age targeting, by contrast, holds up between storms, which is why working aging roofs is a year-round play even in quiet weather.

What should I do for canvassing between storms?

Switch from storm-fused targeting to roof-age targeting across your whole service area. The same address-level roof-age data that ranks doors inside a swath also ranks the aging roofs in your market on a calm week. Storm data plus roof age is a storm-season tool; roof age alone keeps your crews busy and your pipeline alive between events, so you are not stuck in the feast-or-famine cycle of waiting on weather.