How to Find Which Streets Are Due for Re-Roofs (A Field Guide for Crews That Knock)

Every roofing company eventually hits the same wall. The phone slows down, the easy referrals dry up, and somebody on the team says it out loud: we need to go knock. Then comes the harder question nobody likes to answer honestly. Knock where?

Most crews answer it by gut. They pick a subdivision that "looks old," they remember a hailstorm that came through "a couple years back," and they spend a Saturday walking streets where two-thirds of the houses got new roofs eighteen months ago. By noon the rep is sunburned, demoralized, and sitting on four no-answers and one homeowner who already has a contractor. The territory wasn't bad. The targeting was bad.

Finding which streets are due for a re-roof is a solvable problem, and the contractors who solve it are quietly out-producing the ones who don't. It comes down to two questions you can actually answer with data instead of vibes. First: how old is the roof on each house, roughly? Second: what has the weather done to that specific roof since it went on? Stack those two answers across a neighborhood and the "due" streets light up. The rest of the county you leave alone.

What follows is the workflow real storm-restoration and retail crews use to do that. It covers reading roof age from aerial imagery, modeling storm exposure roof by roof, scoring streets so your best rep walks the best block first, and the operational stuff that actually decides whether the data turns into signed contracts: route density, knock timing, list hygiene, and the legal lines you do not cross. There are worked examples with real numbers, checklists you can hand to a canvasser tomorrow, and an honest accounting of where each method breaks down.

What "due for a re-roof" actually means

Before you can find due streets, you have to define "due" in a way that isn't wishful thinking. A roof is a candidate for replacement when one or more of these is true, and the more that stack up, the stronger the candidate:

• It is aging out of its service life. A typical 3-tab asphalt shingle roof in a sun-and-storm climate gives you somewhere around 15 to 20 years of real protection, not the 25 or 30 printed on the wrapper. Architectural (dimensional) shingles run longer, often 20 to 30 years in practice. Once a roof crosses roughly the two-thirds mark of its expected life, granule loss, brittleness, and sealant failure accelerate, and the roof starts failing in storms it would have shrugged off when new.
• It took a storm hit that did real damage. Hail that fractures the shingle mat, or wind that lifts and creases tabs, can age a roof a decade in fifteen minutes. A four-year-old roof can be a legitimate replacement candidate after a severe hail event, while the eleven-year-old roof two blocks over that the storm missed is not.
• It is showing visible distress from the ground or the air. Curling, cupping, missing tabs, exposed felt, sagging planes, moss lines, and patchwork repairs are all tells. From aerial imagery you can spot a surprising number of these.

The trap is treating any one signal as the whole answer. Age alone over-targets: plenty of 16-year-old roofs are perfectly serviceable and the homeowner will tell you so at the door. Storm alone over-targets too: a brand-new roof in the hail swath usually doesn't have a damage case worth pursuing. The streets you actually want are where age and exposure overlap — older roofs that also caught the storm — plus the standalone severe-storm hits regardless of age.

Hold onto that overlap idea. It is the entire game.

The two data layers that decide everything

Think of every neighborhood as two transparent maps stacked on top of each other.

Layer one is roof age. For each address, roughly how many years has the current roof been up there? You will never get an exact install date for a whole neighborhood — nobody has that — so you work in ranges. "This roof reads 18 to 22 years old" is an honest, useful answer. "This roof was installed on March 4, 2007" is a claim you cannot back up and should never make to a homeowner.

Layer two is storm exposure. For each address, what is the worst weather that roof has plausibly seen, and how hard? Hail size, wind speed, number of events, time since the last significant one. This is modeled — it is odds and physics, not proof of damage on any single roof. A storm map tells you a roof was probably exposed to 1.5-inch hail; it does not tell you that roof is damaged. You confirm damage with an inspection, every time.

Where the two layers are both "hot" — older roof, meaningful storm exposure — you have a due street. Let's build each layer.

Layer one: reading roof age across a neighborhood

You have four practical ways to estimate roof age, from cheapest-and-roughest to most precise. Use them in combination.

Method 1: Aerial and satellite imagery, read by eye or by software

This is the workhorse. High-resolution aerial imagery (the kind in mapping tools and dedicated roofing-measurement platforms) lets you see the roof surface clearly enough to judge its condition and approximate age without leaving your desk.

What you are looking for, plane by plane:

• Color uniformity and granule loss. A newer asphalt roof reads dark and even. As it ages, granules wash into the gutters and the mat shows through, so the roof looks lighter, blotchy, and "thirsty." Streaky black discoloration is usually algae (Gloeocapsa magma), which is cosmetic but correlates with older, damp-prone roofs.
• Surface texture. Curling and cupping shingles cast tiny shadows that give an old roof a rough, shadowed texture compared to the flat sheen of a new one. On a sunny midday capture you can often see it.
• Repairs and mismatched sections. A patch of different-colored shingles, a tarped section, or a single replaced plane tells you the roof is already failing somewhere.
• Generation of imagery. This is the underrated trick. Most imagery providers keep a history of captures for the same address going back years. If you can find the capture where the roof visibly changed from old-and-light to new-and-dark, you have bracketed the install: it happened between those two dates. That turns a guess into a tight range.

The historical-imagery move deserves emphasis because it is how you separate a re-roofed house from its identical neighbors. Two houses built the same year in the same subdivision are not the same target if one was redone after a 2019 storm and one wasn't. Imagery history catches that; tract data does not.

One more practical note on imagery vintage. Aerial captures are not refreshed everywhere at the same cadence. Dense suburbs often get reflown every year or two; rural areas can go three or four years between captures. That matters because a roof replaced last spring may still show as old in last year's capture, and you'll knock a door the homeowner already closed on a competitor. So treat any single capture as provisional and confirm the freshest-looking roofs with a windshield pass before you commit reps. The freshest publicly available federal imagery (the kind in USGS EarthExplorer and the National Agriculture Imagery Program) is useful for history but usually lags the commercial providers on resolution and recency, so use it to extend the timeline backward rather than to judge today's condition.

Another field tell that imagery surfaces well: solar panels. A roof with panels was almost certainly sound when the panels went on (no installer puts panels on a failing roof), and removing and reinstalling panels for a re-roof is expensive enough that those homeowners delay. Flag solar homes as lower priority unless the panels are clearly old themselves, and never assume you can re-roof under panels without factoring the detach-and-reset cost into the conversation.

Method 2: County records and permit data

The assessor and the permitting office are public, free, and underused. Two fields matter.

• Year built. In a subdivision where most roofs are original, year-built is a strong proxy for roof age. If the tract went up in 2003 and you find no re-roof permits, a 2003 roof is around 22 years old and well into replacement territory.
• Re-roof / reroof permits. Many jurisdictions require a permit to tear off and replace a roof. Where they do, the permit record tells you exactly which addresses already got a new roof and when. That is a negative filter as valuable as any positive one: it crosses streets off your list so you stop knocking houses that just paid someone else.

The big caveat: permit compliance is wildly inconsistent. Plenty of re-roofs happen without a pulled permit, especially insurance jobs done fast after a storm. So absence of a permit does not prove the roof is old. Use permit data to confirm recent work and to break ties, not as your only age signal.

Method 3: Subdivision and builder patterns

Walk or drive a tract neighborhood and you'll notice the builder used the same shingle, same color, same pitch on a whole phase. That uniformity is a gift. When you confirm the age and condition of a handful of roofs on a block, you can reasonably extrapolate to the untouched neighbors. The exceptions — the houses that broke the pattern — are exactly the re-roofed ones, and they stand out precisely because they don't match.

Method 4: Ground-truth windshield survey

Nothing beats eyes on it, and you should always validate your desk work before committing a full day of canvassing. A windshield survey is a slow drive-through where a rep (not driving) photographs and rates roofs on a quick scale. You are checking that the streets your data flagged actually look the part, and you are catching things imagery missed: a north-facing plane that hides damage from overhead capture, a recent re-roof too new to show in last year's imagery, a homeowner with a contractor's sign in the yard.

Putting the age layer together: a worked example

Say you are looking at a 240-home subdivision built in 2004.

1. Pull year-built: nearly all 2004. Baseline roof age, if original, is about 21 years.
2. Pull re-roof permits for the tract: 38 addresses show permits between 2017 and 2023. Cross those off — they're not due.
3. Scan current aerial imagery on the remaining 202: 26 of them read clearly new (dark, even, no algae) despite no permit — likely unpermitted re-roofs. Cross those off too.
4. You're left with roughly 176 homes whose roofs read 18-plus years old and original. Now the age layer is hot across most of the subdivision, and you've already saved your reps from knocking 64 doors that would've been dead ends.

That is the age layer doing its job: it didn't just tell you where to go, it told you where not to.

What roof material does to your age math

The "15 to 20 years" rule of thumb is for standard asphalt shingles, which cover the vast majority of the residential housing stock you'll canvass. But the moment you target a neighborhood with different materials, your age math has to shift, and reps who don't adjust waste knocks.

• 3-tab asphalt: the shortest practical life, roughly 15 to 20 years, less in brutal sun or hail country. These are your bread-and-butter re-roof targets and they age visibly.
• Architectural / dimensional asphalt: thicker, roughly 20 to 30 years in practice. A 15-year-old architectural roof is often not due yet on age alone, so lean harder on the storm layer there.
• Wood shake: ages on a different curve, splits and cups, and is increasingly restricted by fire codes; a due wood-shake neighborhood is a real opportunity but a different sales conversation.
• Tile (concrete or clay): the tile itself can outlast the house, but the underlayment beneath it typically fails around 20 to 30 years, and the homeowner often doesn't know the tile is fine while the waterproofing is shot. These are high-ticket, low-volume targets that age math from year-built handles well.
• Metal: 40-plus years; usually not a re-roof target unless storm-damaged or a failed earlier install.

The operational takeaway: before you score a neighborhood on age, know what's on the roofs. A subdivision of architectural shingles built in 2010 is not yet a strong age play in 2025, while a 3-tab tract from 2008 absolutely is. Imagery plus a quick windshield pass tells you the material; let it recalibrate your age thresholds per neighborhood instead of applying one number everywhere.

Layer two: modeling storm exposure roof by roof

Age tells you which roofs are tired. Storm exposure tells you which tired roofs got pushed over the edge, and which young roofs unexpectedly belong on your list. This layer is where most contractors are sloppy, because the public storm data is coarse and easy to misread.

The public sources and what they actually tell you

• NOAA's Storm Prediction Center and the National Weather Service publish hail and wind reports. These are point reports — a trained spotter or a measured gauge said "golf-ball hail here at this time." They are real but sparse. A storm can drop damaging hail on a street where no human ever reported it, so the absence of a report does not mean the absence of hail.
• Radar-derived hail estimates (MESH and similar products) fill the gaps. Weather radar estimates the maximum expected hail size across a continuous grid, so you get coverage everywhere the storm passed, rather than only where someone happened to report. This is far more useful for canvassing because it gives you a swath, not dots.
• Wind data comes from measured gusts at stations plus radar and damage surveys. Straight-line wind and the occasional tornado leave damage patterns that follow the storm track.

The core limitation to internalize: storm data describes what the sky did, not what your roof looks like now. A 1.75-inch hail estimate over a neighborhood means roofs there were probably exposed to damaging hail. Whether any specific roof has a fractured mat, bruising, or spatter that holds up to inspection is a separate question you answer on the roof.

Reading a hail swath like a pro

When you overlay a hail estimate on your territory, you're looking for:

• The core of the swath, where estimated hail size is largest. Damage probability is highest here.
• Hail size thresholds. Soft metals (gutters, vents, AC fins) start denting around 1 inch. Asphalt shingle functional damage becomes common as you climb past roughly 1.25 to 1.5 inches, and gets serious past 1.75 to 2 inches. Below 1 inch you are usually looking at cosmetic spatter, not a replacement case.
• Directionality. Hail driven by wind hits one slope harder. The west and south faces often take the worst of it in many storm setups, which is why a roof can look fine from the street and be wrecked on the back slope.
• Recency. A storm from last month is a live opportunity with intact evidence and an open insurance window. A storm from three years ago has mostly been claimed, repaired, or weathered past the point of a clean case. Know your state's claim-filing deadlines, because they cap how far back a storm is still actionable.

The overlap that defines a due street

Now stack the layers. For each address you have, roughly:

• Roof age range (Layer one)
• Max plausible hail / wind exposure and recency (Layer two)

Four buckets fall out:

The streets where the first bucket clusters are your re-roof streets. That is the answer to the question. Everything else in the workflow is about acting on it efficiently and legally.

Where per-roof modeling tools fit

Doing all of the above by hand — pulling permits, scrubbing imagery house by house, overlaying NOAA reports and radar hail, then merging it into one ranked list — is absolutely possible. Plenty of strong companies do it manually, and you should understand the manual version because it's how you sanity-check anything automated. But it is slow, and at county scale it does not realistically get done every week by a sales manager who also has crews to run.

This is the gap platforms like RoofPredict are built for. The idea is to do both layers at the address level, automatically, and hand you a ranked map instead of a stack of tabs. For each roof it estimates an age range from aerial imagery (a range, never a fake install date), and it models storm exposure per roof — the specific hail and wind physics that passed over that individual address — rather than painting a whole zip code with one broad storm. Then it scores and ranks streets and doors so your canvasser walks the highest-probability block first instead of guessing.

A few honest things about what that does and doesn't do, because the category is full of overclaiming:

• It ranks, it does not prove. A high score means a roof is statistically likely to be old and storm-exposed. It is a prioritization signal, not a verdict on damage. You still get on the roof and inspect. The forecast is odds, not evidence.
• Age is a range, by design. Aerial imagery can bracket when a roof changed; it cannot read an install date. Anyone selling you exact dates from imagery is selling you confidence they don't have.
• It does not replace judgment, relationships, or a clean inspection. It replaces the part where you waste Saturdays driving streets that looked promising and weren't. The closing still depends on your reps, your workmanship reputation, and an honest assessment at the door.
• It is not a lead-buying service and it does not knock for you. It does not hand you a homeowner who raised their hand. It tells you which doors are worth your own crew's knock. The work is still yours.

Whether you build the layers yourself or use a tool to assemble them, the underlying logic is identical: age range plus per-roof storm exposure, ranked. Use whichever gets a correct, current ranked list in front of your reps each morning. The tool is a means to that list, nothing more.

The retail (non-storm) play that crews underwork

Most of the attention in this trade goes to storm chasing, because a fresh hailstorm is a loud, time-boxed opportunity. But the steadier, less competitive money is in aging-roof retail targeting, and the same age layer feeds it. These are the streets where roofs are simply at the end of their service life and the homeowner is going to replace eventually whether or not a storm ever shows up.

The retail play has real advantages. There is no claim clock, so you can work these blocks any week of the year, which smooths out the feast-or-famine cycle that wrecks storm-only companies. Competition is thinner because the chasers ignore non-storm neighborhoods. And the homeowner conversation is cleaner: you're not navigating an insurance claim, you're discussing a roof that's visibly worn out and a straightforward replacement. Margins on a retail re-roof are often healthier than a beaten-down insurance scope.

How to find them: run the age layer on subdivisions that are 18-plus years past their build date with few re-roof permits, then sort by visible condition from imagery (granule loss, curling, algae, patches). The densest clusters of original, end-of-life roofs become your no-clock backfill list. Work them in the slow weeks between storms with an honest, condition-led opener — "we're inspecting roofs in the neighborhood and yours is showing its age; want a free, no-pressure assessment?" — and you keep crews busy and cash flowing year-round.

A balanced book of business runs both: fresh storm swaths when they come, aging-roof retail streets to fill the calendar in between. Companies that live and die on storms alone have wild revenue swings and lay off good crews in quiet seasons. The age layer is what makes the retail half of that book possible.

Turning a ranked list into a canvassing plan

A ranked list of due addresses is not yet a plan. Reps don't knock spreadsheets; they walk streets. Translating scores into routes is where a lot of the real productivity gain hides, and it's almost entirely about density.

Why density beats raw score

Imagine two streets. Street A has six top-priority homes scattered among forty average ones. Street B has twenty top-priority homes in a row. Street B is a better route even if Street A has your single highest-scoring house, because your rep spends the day at doors, not in the truck. Knock-to-knock walking time is the silent killer of canvassing productivity. A tight route of good doors will out-produce a sparse route of great doors almost every time.

So when you build routes, cluster. Group adjacent high-priority addresses into walkable loops of roughly 30 to 60 doors — about what one rep can work in a focused two-to-three-hour block. Order the loop so the rep is always walking to the next nearby door, not crisscrossing.

A simple street-scoring formula you can run by hand

If you want a number per street without any software, score each street like this:

1. Age points: count of homes on the street with roofs 15+ years old and original.
2. Storm points: count of those homes inside a meaningful hail/wind swath (≥1.25 in hail or ≥60 mph wind), recent enough to be actionable.
3. Overlap bonus: homes that are both old and storm-exposed, counted again (so overlap is weighted double).
4. Density divisor: divide by the street's length so you reward tight clusters over spread-out ones.

Street score = (age points + storm points + overlap bonus) / blocks of length. Rank your streets by that number, walk the top of the list first. It's crude, it's transparent, and it beats gut feel every single time.

Sequencing the day

• Best block, best rep, best window. Don't burn your top-scoring street on a Tuesday at 2 p.m. when nobody's home. Save it for a high-contact window (more on timing below) and put your strongest closer on it.
• Warm the rep up. Some managers send reps to a medium block first to get reps loose, then onto the prime block once they're sharp. Others go prime-first to capture the freshest energy. Test both with your team.
• Leave-behind on no-answers. Half or more of doors won't answer. A branded door hanger with a clear, honest message turns a no-answer into a possible callback. Track which streets the callbacks come from — that's feedback on your targeting.

Tracking knocks so the data compounds

The targeting work is wasted if the result of each knock evaporates. Every door your reps touch is a data point, and capturing it turns one season's effort into next season's edge. At minimum, log per address: knocked / no-answer / not-interested / inspection-booked / inspected / signed, plus a quick reason on the dead doors (recent re-roof, has-a-contractor, new build, renter who can't decide). Cheap canvassing apps and most roofing CRMs handle this on a phone at the door.

Two payoffs. First, no-answers become a re-knock list you work in the low-contact windows instead of abandoning, and re-knocks convert at a meaningful rate because timing, not interest, was the real reason for the miss. Second, the dead-door reasons feed back into your targeting: if a "top-priority" street is logging fifteen recent-re-roof crosses, your age layer for that tract is stale and you fix it before sending reps back. Over a year this loop is the difference between a team that gets sharper each storm and one that repeats the same misses.

Keep the logging dead simple, though. A canvasser fumbling a complicated form at the door knocks fewer doors. One screen, big buttons, ten seconds per address. Sophistication in the office, simplicity at the door.

Timing: the layer everyone forgets

You can have a perfect list and still flop because you knocked at the wrong time. Timing splits into two scales.

Storm timing (the macro window)

After a significant hail or wind event, there is a window where roofs have fresh, documentable damage and homeowners are alert to it. Move in that window with an honest condition assessment and you're meeting demand, not manufacturing it. Wait too long and three things happen: the damage weathers and gets harder to distinguish from age, the insurance filing window narrows or closes, and competitors saturate the neighborhood. Know your state's claim deadlines and work the fresh swaths first.

For aging-roof (non-storm) targeting, there's no clock — those streets are due whenever you get to them — so they're your steady backfill between storms. The mix of "storm-fresh" and "steady aging" work is what keeps a crew busy year-round instead of feast-or-famine.

Knock timing (the micro window)

Contact rate is everything in canvassing, and it swings hard with the clock:

• Weekday late afternoon to early evening, roughly 4:30 to 7:30 p.m., is the reliable workhorse window. People are home, it's not too late to be rude, and you get face-to-face contact instead of door hangers.
• Saturday mid-morning to mid-afternoon is strong, especially for homeowners who work standard hours. Sunday is more sensitive — respect that it's off-limits for many households.
• Weekday mornings and mid-days are low-contact and best used for re-knocks of no-answers, drop-offs, and windshield surveys rather than fresh prospecting.
• Respect local solicitation rules and posted signs. "No soliciting" means skip it. Some municipalities require a canvassing permit; some HOAs forbid door-to-door outright. Knowing the rules before you walk keeps your reps out of trouble and your brand out of the local complaint feed.

What pros get wrong

A decade of watching crews canvass surfaces the same mistakes over and over. Avoiding these is most of the edge.

Treating a zip code as a target. Storm and age both vary block by block. "We're working the 750-whatever" is not a plan; it's an excuse to wander. Get down to streets and addresses or you're leaving most of the value on the table.

Chasing the loudest storm instead of the actionable one. The storm that made the news isn't always the one with the best combination of damage, recency, and low competition. Sometimes the quieter event two counties over, that nobody else worked, is the better territory.

Ignoring the negative filters. Knowing where not to knock — recent re-roofs, brand-new construction, houses already flying a competitor's sign — is as valuable as the positive list. Reps who skip dead doors fast knock more live doors per hour.

Confusing exposure with damage at the door. Telling a homeowner "the radar shows you have hail damage" is wrong on the facts and a compliance problem. The radar shows exposure. You document the actual condition on the roof. Keep those words straight or you'll get yourself and the homeowner in trouble.

Over-relying on age and annoying the wrong people. A 17-year-old roof that's been meticulously maintained and never stormed may have years left, and the homeowner knows it. Lead with condition and an honest read, not a presumption that old equals due.

No feedback loop. The crews that compound are the ones who feed results back into targeting: which streets converted, which were already saturated, which scored high but flopped. Without that loop you make the same targeting error every season.

Burning out reps on bad routes. Sparse, low-contact routes destroy morale faster than rejection does. Density and timing protect your people, and your people are the asset that actually closes.

Edge cases that break the simple model

The age-plus-storm overlap is the right backbone, but a few situations don't fit it cleanly, and knowing them keeps your reps from chasing ghosts or skipping real opportunities.

Mixed-age subdivisions. Some neighborhoods built out in phases over five or six years, so "year built" smears across a range and the original-roof assumption gets fuzzy. Read condition from imagery rather than leaning on build year here, and expect more re-roof variation block to block.

Rental and absentee-owned streets. Renters can't authorize a roof, and absentee owners are slow and often out of market, so a street that scores well on age can convert poorly. Cross-reference owner-occupancy where you can; a high-age street that's 60 percent rentals is a worse route than the score suggests.

HOA and architectural-review neighborhoods. Some HOAs dictate shingle color and brand and require approval before any re-roof. That's not a reason to skip them, but the conversation and timeline differ, and a few HOAs ban door-to-door outright, so check before you send reps.

Newly developed exurban edges. Brand-new construction shows up as zero-age and should be skipped, but the model can mislabel a half-built street. A windshield pass catches the difference between a finished 2024 tract (skip) and a 2006 tract that just happens to border it (work).

Storm damage that imagery can't see. Overhead aerial capture is weak on the exact things that matter most for a hail case: granule loss in the field of the shingle, mat bruising, and soft-metal denting are often invisible from straight above. So storm exposure is a reason to inspect, never a substitute for the inspection. A roof that looks untouched from the air can be a clear replacement on the ladder, and vice versa.

Distinguishing age wear from storm damage on the roof. This is where reps and even some adjusters get sloppy. Age wear is uniform and gradual: granule loss spread evenly, curling at the edges across whole planes, brittle tabs everywhere. Hail damage is random and impact-shaped: scattered bruises and fractures with no pattern, often with matching dents on the soft metals and spatter marks on painted surfaces. Wind damage is directional: creased and lifted tabs concentrated on the windward slopes and edges. Reading those signatures correctly is what separates an honest, defensible inspection from a stretch, and it's what protects your reputation when an adjuster meets you on the roof.

The compliance lines you do not cross

Storm-restoration canvassing operates in a regulated space, and the rules exist for good reasons. None of this is legal advice — check your own state's department of insurance and your local ordinances — but these principles keep ethical contractors out of trouble:

• The contractor documents conditions and provides estimates. The insurer decides coverage. The homeowner owns the claim. Stay in your lane. You inspect, you document, you bid, you do excellent work. You don't adjust the claim, you don't promise what will or won't be covered, and you don't handle the homeowner's claim for them.
• Don't promise to cover, waive, rebate, or absorb a deductible. In most states that's illegal, and it's a fast way to lose your license and your reputation. The deductible is the homeowner's responsibility, full stop.
• Don't promise a "free roof." Whether insurance pays is the insurer's call based on the homeowner's policy and the documented damage. Promising a free roof is both a compliance violation and a setup for an angry customer.
• Don't present a storm forecast or exposure map as proof of damage. It's a prioritization signal and a reason to inspect, not evidence. Damage is established on the roof and in the documentation.
• Honor solicitation rules, permits, and posted signs. "No soliciting" is a hard stop. Pull canvassing permits where required.
• Be honest about roof age. A range from imagery is a range. Don't manufacture a precise install date you can't support, and don't tell a homeowner their roof is "shot" when your own read is a probability.

The contractors who treat compliance as a feature, not a friction, win the long game. Homeowners can smell a hustle, and referrals follow the crews that played it straight.

A full worked example, start to finish

Let's run one territory end to end so the workflow is concrete.

The setup. A 1.75-inch hail core clipped the northeast quadrant of a suburb eleven days ago. You have three reps and want two solid days of canvassing out of it.

Step 1 — Define the storm footprint. Pull the radar hail estimate and the NWS/SPC reports. The damaging core (≥1.5 in) covers about 1,900 homes across nine subdivisions. The fringe (1 to 1.5 in) covers another 3,000 you'll treat as secondary.

Step 2 — Layer in age. Across the 1,900 core homes, you bracket roof age from imagery history and year-built. Roughly 1,100 read 12-plus years and original. About 500 read newer or recently re-roofed (cross off the new-and-stormed-light ones unless imagery shows fresh hail-era change). About 300 are uncertain — flag for windshield check.

Step 3 — Find the overlap. Inside the damaging core, the ~1,100 older-and-exposed homes are your top bucket. The ~500 newer-but-exposed homes are a secondary storm play worth inspecting. That's your priority universe: about 1,600 doors, with 1,100 of them top-tier.

Step 4 — Build dense routes. Cluster the 1,100 top homes into walkable loops. Three subdivisions hold the densest clusters — call them roughly 700 of the 1,100 in tight blocks. Those three become Day One. Each rep gets two loops of ~40 doors for the prime evening window.

Step 5 — Sequence and time it. Day One, prime window 4:30 to 7:30 p.m., strongest reps on the densest blocks. Day Two morning: windshield-survey the 300 uncertain homes and re-knock Day One no-answers; Day Two evening: work the next-densest clusters.

Step 6 — At the door. Honest opener: a storm came through, you're inspecting roofs in the neighborhood, you'd be glad to take a look and document the condition at no cost. On the roof you assess actual damage. If it's there, you document it thoroughly with photos and measurements. The homeowner decides whether to file; the insurer decides coverage; you provide the estimate and, if hired, the work.

Step 7 — Feed results back. End of Day Two, log it: which loops converted to inspections, which converted to signed work, which were already saturated by a competitor, which scored high but came up empty. That log sharpens the next storm's targeting. Over a season, the feedback loop is worth more than any single storm.

The difference between this and "go knock the northeast side" is the difference between 1,600 deliberate doors and 5,000 random ones. Same storm, a fraction of the windshield time, far more inspections per rep-hour.

A canvasser's pre-walk checklist

Hand this to every rep before they leave the truck:

• I have the ranked door list / map for my specific loop, not merely "the neighborhood."
• I know which homes are top-priority (old + storm-exposed) versus secondary.
• I've noted the cross-off homes (recent re-roofs, new builds, competitor signs) so I don't waste knocks.
• I know the storm date and the actionable window, and I will speak about exposure, not "proven damage."
• I will not promise deductible help, a free roof, or guaranteed coverage.
• I've checked for no-soliciting signs and any local permit requirements.
• I'm walking in the high-contact window and have door hangers for no-answers.
• My route is dense and ordered so I'm always walking to the nearest next door.
• I have my inspection kit, camera, and a clean way to document condition honestly.
• I know how I'm logging results so tonight's data sharpens tomorrow's targeting.

A manager's weekly targeting rhythm

For the person building lists, a simple cadence keeps it sustainable:

• Monday: Review last week's storm activity and any new hail/wind swaths. Update the actionable-storm list and drop anything aged past your state's filing window.
• Tuesday: Refresh the age layer on new target areas — imagery scan, year-built, re-roof permits. Cross off the negatives.
• Wednesday: Build and densify routes from the overlap. Assign loops to reps by skill and block quality.
• Thursday–Saturday: Canvass the prime windows. Re-knock no-answers in low-contact windows.
• Sunday/Monday: Log conversion by street and feed it back into next week's scoring.

Keep that rhythm and "where do we knock?" stops being a Saturday-morning guess and becomes a standing answer your team trusts.

Bringing it together

Finding the streets due for a re-roof isn't mystical and it isn't luck. It's two data layers — roof age as a range, and storm exposure modeled per roof — stacked to find the overlap, then turned into dense, well-timed, compliant routes that respect both the homeowner and your reps' time. You can assemble those layers by hand with public records, imagery, and NOAA data, or you can lean on a per-roof platform like RoofPredict to rank the doors for you. Either way the logic is the same, and the payoff is the same: your crew spends its hours at the doors that are actually due, not the ones that merely looked it from the street.

The companies that win the next storm season won't be the ones that knock the most doors. They'll be the ones that knock the right ones first.

FAQ

How can I tell how old a roof is without climbing on it?

Work in ranges, not exact dates. Read condition from high-resolution aerial imagery (granule loss, curling, algae streaks, patch repairs), then bracket the install by finding the historical capture where the roof changed from old-and-light to new-and-dark. Cross-check with county year-built data and re-roof permits. Together these give you an honest age range like "18 to 22 years," which is all you need for targeting.

Does a storm map prove a roof is damaged?

No. Storm data — radar hail estimates, NOAA reports, measured wind — describes exposure: what the sky did over an area. It tells you a roof was probably hit by damaging hail or wind, which is a strong reason to inspect. It does not establish that any specific roof has a fractured mat or bruising. You confirm actual damage on the roof, and you should never tell a homeowner the map proves they have damage.

What hail size actually damages an asphalt shingle roof?

Roughly speaking, soft metals like gutters and vents start denting around 1 inch. Functional shingle damage becomes common as hail climbs past about 1.25 to 1.5 inches and gets serious past 1.75 to 2 inches. Below 1 inch you're usually looking at cosmetic spatter rather than a replacement case. Wind-driven hail hits one slope harder, so a roof can look fine from the street and be damaged on the back face.

Is roof age or storm exposure the better targeting signal?

Neither alone — the overlap is what matters. Age alone over-targets because plenty of older roofs are still serviceable. Storm alone over-targets because new roofs in a swath rarely have a strong case. The best streets are where older roofs and meaningful, recent storm exposure coincide, plus standalone severe-storm hits regardless of age. Stack both layers and target the overlap.

Why does route density matter more than knocking the single best house?

Walking time between doors is the silent productivity killer in canvassing. A tight loop of 30 to 60 good doors lets a rep stay at doors instead of in the truck, so it almost always out-produces a sparse route with one great house and a lot of driving. Cluster your highest-priority addresses into walkable loops and order them so the rep always walks to the nearest next door.

When is the best time to knock doors for roofing?

Contact rate drives everything. Weekday late afternoon to early evening, about 4:30 to 7:30 p.m., is the reliable window. Saturday mid-morning to mid-afternoon is strong; Sunday is sensitive and off-limits for many households. Use low-contact weekday mornings and middays for re-knocking no-answers and windshield surveys, not fresh prospecting. Always honor no-soliciting signs and local permit rules.

What does RoofPredict actually do, and what are its limits?

It builds both targeting layers at the address level and ranks streets and doors so your crew works the highest-probability block first. For each roof it estimates an age range from aerial imagery and models storm exposure per roof rather than painting a whole zip code with one storm. Its honest limits: it ranks, it doesn't prove damage; age is a range, never an exact install date; it doesn't knock for you or sell you raised-hand leads; and you still inspect every roof. It's a prioritization tool, not a verdict.

How do I avoid knocking houses that already got a new roof?

Use negative filters as hard as positive ones. Pull re-roof permits to confirm recent work, scan current aerial imagery for roofs that read clearly new (dark, even, no algae), and watch for competitor yard signs and tarps during a windshield survey. Permit data is imperfect because many re-roofs go unpermitted, so combine it with imagery. Crossing off dead doors fast is how reps knock more live ones per hour.

What can I legally say to a homeowner after a storm?

Stay in the contractor's lane: you document conditions and provide estimates, the insurer decides coverage, and the homeowner owns the claim. Speak about storm exposure and the actual condition you document on the roof, not "proven" damage from a map. Never promise to cover or rebate a deductible, never promise a free roof, and never guarantee coverage. Check your state's department of insurance for specifics; these are common rules but vary by state.

How far back is a storm still worth working?

It depends on your state's claim-filing deadlines and how fast the neighborhood gets saturated and repaired. Fresh events have intact, documentable damage and an open insurance window, so they're the priority. As a storm ages, damage weathers and gets harder to distinguish from normal wear, competitors move in, and filing windows narrow or close. Work the freshest actionable swaths first and use aging-roof streets as your steady, no-clock backfill.