How to Find Neighborhoods With Old Roofs: A Field Playbook for Roofing Contractors

Every roofing company has a story about the street that paid for the truck. One sales rep knocks a cul-de-sac on a whim, finds nine roofs all installed the same summer in 2006, and three weeks later the crew has booked out a month. The problem is that nobody can repeat it on purpose. They went back to the well, knocked the next subdivision over, and got nothing but barking dogs and 'we just did ours.'

Finding neighborhoods with old roofs is not luck, and it is not a vibe you get from driving around. It is a research problem with a small number of reliable inputs: when the houses were built, when the roofs were last replaced, what storms have rolled through since, and what the roofs actually look like from above today. Every one of those inputs is available to you, most of them for free, and the contractors who consistently fill their boards have simply built a habit of checking them before they spend a dollar on labor.

What follows is the workflow a sharp canvassing manager would hand a new rep. It covers the data sources, how to read each one, the math on roof lifespan that tells you what 'old' even means, how to layer storm history on top, and the mistakes that send crews to the wrong streets. There are worked examples with real numbers, a few checklists you can copy, and an honest look at where the easy methods break down. The goal is simple: by the end you should be able to open a county GIS site and a mapping tool and walk out with a ranked list of streets worth your crew's morning.

What 'old roof' actually means before you go looking

Before you can find old roofs, you need a working definition, because 'old' is the whole game and most reps carry a fuzzy one in their head. The honest answer is that it depends on the roofing material, the climate, the installation quality, and the weather the roof has eaten. But you can anchor on the dominant material in the United States: asphalt shingles cover the large majority of steep-slope residential roofs.

A standard 3-tab asphalt shingle roof is generally engineered for somewhere in the 15-to-20-year range under normal conditions. Architectural (dimensional) shingles, which became the default on new construction over the last two decades, are typically rated longer, often in the 25-to-30-year range, with premium products carrying longer warranties. Those are warranty and manufacturer figures, not guarantees of field performance. Real-world life is usually shorter than the printed number because of heat, ventilation problems, hail bruising, wind uplift, and the simple fact that the south- and west-facing slopes age faster than the north side.

Here is the practical translation for a canvasser. A neighborhood built in the late 1990s or early 2000s with original 3-tab shingles is almost certainly past or near the end of its service life right now. A subdivision built in 2008 with architectural shingles may still have years left unless a storm wore it out early. So 'old' is really two questions stacked together:

1. Is the roof old by the calendar? Years since the last install, compared against the expected life of whatever material is up there.
2. Has the roof been aged early by weather? Hail and wind can take a 12-year-old roof and make it functionally a 20-year-old roof in an afternoon.

The best targets sit where both answers point the same direction: an aging roof in a corridor that has also taken storm hits. A roof that is merely old will eventually need replacement and is a legitimate retail conversation. A roof that is old and storm-worn is the one where the conversation moves fastest, because the homeowner has a reason to act now and an avenue to address it through their own insurer if the damage warrants it.

A note on lifespan tables you will see online: treat them as planning ranges, not appraisals. The table below is a reasonable planning reference for steep-slope residential materials, but the spread is wide for a reason.

Why this matters for targeting: a tile or slate roof neighborhood is a poor canvassing target even if the houses are 40 years old, because the field is long-lived and the failure point is hidden underlayment, not the visible surface. An asphalt-shingle subdivision built 18 years ago is a prime target. Knowing the dominant material in a tract before you knock saves you from wasting a morning on streets that will never convert.

One more wrinkle that separates careful operators from the rest: the same shingle ages at wildly different speeds depending on where it sits. A roof in a hot southern climate with poor attic ventilation can cook through its service life years ahead of the warranty number, because trapped attic heat bakes the asphalt from below while the sun bakes it from above. The same product in a milder climate with good ridge-and-soffit ventilation can run close to or past its rated life. So when you anchor on a planning number, mentally shade it: in hot, sunny, high-UV regions, subtract a few years; in mild northern climates, the warranty number is closer to reality. This is why a 2006 subdivision in one metro is screaming for replacement while the identical-vintage subdivision two states north still has life left. Local climate is a multiplier on the calendar, and reps who carry one national lifespan number in their head misjudge both markets.

The five data layers that locate old roofs

Finding neighborhoods with old roofs comes down to assembling and overlaying five layers of information. None of them is sufficient alone. Together they turn guessing into a ranked list. Walk through them in order, because each one filters the next.

Layer 1: Year built (the subdivision age skeleton)

The single most powerful free signal is the year a house was built, because the original roof went on within a year of construction and a large share of homes still carry roofs that are one replacement cycle old at most. Year-built data is public and sits in two places:

• County assessor / property appraiser records. Nearly every county publishes parcel-level data that includes 'year built' (sometimes labeled 'effective year built,' which is different and discussed below). Many counties expose this through an online property search or a GIS portal where you can click any parcel and read its attributes.
• Plat and subdivision recording dates. When a developer records a subdivision plat, that date is public. Tract homes go up fast after platting, so the plat year is a good proxy for the build year of the whole street, especially in the postwar and modern suburban tracts that make the best canvassing targets because they were built in waves.

The reason tract subdivisions are gold: they were roofed in a single season by one or two builders using the same product. That means an entire street ages in unison. Find the build year, add the expected roof life, and you know roughly when the whole block comes due. A street built in 2004 with 3-tab shingles was statistically due for replacement around 2019-2024. Many of those homes replaced once on a storm in that window; many did not, and those are the doors.

Watch out for effective year built. Some assessors record an 'effective' age that gets reset by major renovations or, occasionally, by a permitted reroof. It is not consistent county to county. Use the actual construction year for the build skeleton, and use permits (Layer 2) to find reroofs, rather than trusting effective age to tell you a roof was replaced.

Layer 2: Building permits (the reroof history)

Year built tells you when the original roof went on. Permits tell you whether it has been replaced since, and when. This is the layer that separates pros from amateurs, because a 20-year-old subdivision where half the homes pulled reroof permits after a 2018 hailstorm is a very different target than a 20-year-old subdivision with no reroof permits on file at all.

Most jurisdictions require a permit for a full roof replacement, and many publish permit records. You are looking for the absence or presence of a reroof permit per address:

• No reroof permit since construction + old build year = original roof, prime target. The roof is one cycle old and never touched.
• Reroof permit dated recently = skip. That roof is new; do not waste the door.
• Reroof permit from 8-12 years ago on a 3-tab product = becoming a target again. It is entering a second decision window.

A second, underused read on permit data: it tells you which storms actually converted in a given tract. If you pull reroof permits for a subdivision and see a sharp spike of permits in a single year, cross-reference that year against storm history. When the spike lines up with a known hail date, you have confirmation that the storm produced real, permitted replacements, which tells you the remaining un-replaced roofs on that street are the homes that slipped through. That is a much sharper signal than a generic 'old neighborhood.' You are no longer guessing whether a storm mattered there; the permit record proves it did, and the homes without a matching permit are your list of the ones that got missed.

How to get permit data:

1. Open data portals. Larger cities and many counties run open-data sites where you can download building-permit datasets, filter by permit type (look for 'reroof,' 'roofing,' 'roof replacement'), and join them to addresses. Some let you pull a whole ZIP code at once.
2. Permit search portals. Where there is no bulk download, there is usually a web lookup where you enter an address and see its permit history. Slower, but fine for vetting a specific street before you commit a crew.
3. Public records request. For a systematic territory, a records request for reroof permits in a date range across a jurisdiction can hand you a spreadsheet. It is legal, routine, and many offices fulfill it for a small fee.

A caution that pros learn the hard way: permit compliance is uneven. In some areas a meaningful share of reroofs happen without a pulled permit, especially smaller contractors or fast post-storm work. So a missing permit is strong but not absolute evidence the roof is original. That is exactly why you do not stop at permits. You confirm with imagery and your own eyes.

Layer 3: Aerial and street-level imagery (visual age and condition)

Imagery is where you confirm what the records imply and catch what the records miss. There are two views, and you want both.

Top-down aerial / satellite imagery lets you read a roof's surface and estimate age and condition without leaving your desk. What to look for:

• Granule loss and color fade. Asphalt shingles lose their mineral granules as they age, and the surface shifts from a saturated, even tone to a washed-out, mottled gray. A street where every roof reads patchy and pale is an old-roof street.
• Streaking and dark staining. Black streaks are usually algae (Gloeocapsa magma), which is cosmetic but correlates with older, weathered surfaces that have sat long enough to colonize.
• Surface texture. Curling and cupping shingles cast tiny shadows that, on high-resolution imagery, give an old roof a rougher, less uniform texture than a crisp new one.
• Patch jobs and tarps. Mismatched sections, blue tarps, and partial repairs flag roofs that are failing and homeowners already worried about them.
• Mixed ages on one street. If you see a tract where most roofs look uniformly weathered but a handful look new, the new ones already replaced (probably on the last storm) and the weathered majority did not. That majority is your knock list.

The limitation, and you must respect it: you cannot read a precise install date off a photo. Imagery gives you a confidence range and a relative ranking, not a birthday. A roof that looks 18 from above might be 14 with bad ventilation or 22 with a good north slope. Treat aerial age as a band, not a number.

Street-level imagery (the drive-by photo layer in major mapping tools) adds the angle you cannot get from straight down. From the curb view you can sometimes see drip-edge rust, fascia and gutter condition, and the general upkeep of the home, which correlates with how a homeowner will receive you. It also lets you pre-plan: spot the corner lots, the dogs, the long driveways, and the cul-de-sacs before you ever park.

A worked read: open aerial imagery over a 1999-built subdivision. You count 40 homes. Twelve roofs read crisp and dark (replaced). Twenty-eight read pale, streaked, and patchy (original or near it). Cross-check three of those 28 against permit data and find no reroof permits. You now have high confidence that roughly 28 doors on that loop carry roofs at or past 20 years. That is a morning's work for a crew, sourced entirely from a chair.

Layer 4: Storm history (what aged the roofs early)

A roof that is old by the calendar is a retail conversation. A roof that has also been hit by hail or high wind is a faster conversation and may involve the homeowner's own insurer. Layering verified storm history onto your old-roof map is how storm-restoration companies decide where to surge after an event, and how smart retail companies decide which old neighborhoods will be easiest to convert.

The authoritative public sources:

• NOAA's Storm Prediction Center and the Storm Events Database publish historical hail and wind reports, including estimated hail size and wind speed by location and date. This is the public record of what actually happened, county by county, going back decades.
• The National Weather Service issues and archives severe-thunderstorm and warning data you can map against neighborhoods.
• Hail size thresholds matter. Damage to asphalt shingles generally begins to appear as stones get into the one-inch-and-up range, with bigger stones and higher fall velocity doing more. Roughly quarter-size (1 inch) is a common threshold where functional shingle damage starts to show up, and larger stones do progressively worse. Use reported hail size to gauge whether a given storm plausibly damaged roofs, rather than assuming every storm did.

How to use it: pull the storm history for your service area, note the dates and reported hail sizes for the last 10-15 years, and overlay those storm footprints on your old-roof neighborhoods. The overlap is your priority territory. An old subdivision inside a recent significant-hail swath is the highest-value target you can find, because the roofs are both worn out and freshly damaged.

The honest caveat: a hail report is a point observation, not a guarantee that a specific roof took functional damage. Storm reports are sparse and crowd-sourced; the absence of a report does not mean a street was spared, and the presence of one does not mean every roof on it is compromised. Storm data tells you where to look harder. The roof itself, inspected properly, tells you what is true. Never represent a forecast or a storm report as proof of damage to any individual home. It is a probability that justifies an inspection, nothing more.

Layer 5: Ownership and demographic filters (who is likely to act)

The last layer is about conversion, not detection. An old roof on a house that has been owner-occupied for 15 years converts differently than an old roof on a rental or a home that just sold. A few practical filters, all from public or low-cost data:

• Owner-occupied vs. absentee. Assessor data usually flags whether the owner's mailing address matches the property address. Owner-occupants make roofing decisions about their own home; absentee landlords are slower and more price-driven. Both are valid, but they are different pitches.
• Length of ownership. A homeowner who has been in the house 12 years has likely never reroofed and has the equity and attachment to act. A very recent buyer may defer because they just spent their cash on the purchase, or may be motivated because an inspection flagged the roof.
• Housing-age context from the Census. The American Community Survey publishes median year-built and owner-occupancy by area, which helps you rank whole ZIP codes or tracts before you drill into parcels. It is a coarse filter for prioritizing which subdivisions to research first.

This layer keeps you from knocking a beautifully aged roof on a vacant flip. It is the difference between a list of old roofs and a list of old roofs attached to people who can say yes.

Putting the layers together: a repeatable territory workflow

Here is the actual sequence, start to finish, that turns the five layers into a knock list. Run it the same way every time and it becomes a 60-to-90-minute routine per new area.

1. Pick a candidate area from coarse data. Use Census/ACS median year-built and owner-occupancy to choose ZIP codes where the housing stock skews 15-30 years old and ownership is mostly owner-occupied. This is your shortlist of subdivisions worth real research.
2. Pull the build-year skeleton. Open the county GIS or assessor portal. Identify subdivisions inside your shortlist built in the target window (for 3-tab, roughly 18-26 years ago; for architectural, roughly 22-30 years ago). Note the plat year for whole tracts to confirm they were built in one wave.
3. Subtract the recently reroofed. Pull permit data for those subdivisions. Flag and remove addresses with recent reroof permits. Keep addresses with old build years and no reroof permit on file. This is your raw candidate list.
4. Confirm with aerial imagery. Scan the candidate streets top-down. Down-rank roofs that read new (already replaced without a permit on file) and up-rank roofs reading pale, streaked, patched, or tarped. Note any tile/slate streets and drop them.
5. Overlay storm history. Pull NOAA/SPC hail and wind history for the area. Mark which candidate streets fall inside significant-hail or high-wind footprints from the last 10-15 years. These move to the top.
6. Apply ownership filters. Tag owner-occupied vs. absentee and length of ownership so reps know which pitch to lead with per door.
7. Sequence the routes. Build a walking route that minimizes backtracking, front-loads the highest-confidence streets, and respects practical constraints (school-zone timing, no-soliciting registries, gated communities, dog-heavy blocks). Hand the rep a ranked list, not a heat map they have to interpret on the fly.

The output of one pass is a stack-ranked list of streets and addresses, each tagged with build year, reroof status, an aerial age band, a storm flag, and an ownership note. That is a canvassing plan you can defend, repeat, and improve.

Scoring instead of sorting

Reps work faster when the list is scored rather than merely sorted. A simple points model keeps everyone calibrated and removes argument about which street is 'better.' One workable scheme:

Total each address and knock highest-first. The exact weights matter less than the discipline of using the same ones every time, because consistency is what lets you compare last month's streets to this month's and learn which signals actually predict booked jobs in your market. After a season of tracking outcomes against scores, tune the weights toward whatever your closed-job data says matters most. A model you adjust with real results beats anyone's gut.

A worked example, end to end

Make it concrete. Suppose you run a crew in a mid-size metro and you want next week's routes.

Step 1. ACS data shows three ZIP codes where median year built is around 2002-2006 and owner-occupancy is over 80 percent. You shortlist those three.

Step 2. In the county GIS you find a 96-home subdivision platted in 2003. Sampling parcels confirms build years cluster in 2003-2004. The dominant product, from a quick street-view check, is architectural shingle, so expected life is long-ish, but 21-22 years in puts many of them into the early-failure window, especially if storms hit.

Step 3. Permit data shows 31 of the 96 homes pulled reroof permits, most of them dated to a single year that lines up with a hailstorm. That leaves 65 homes with no reroof permit on file.

Step 4. Aerial imagery of those 65: about 50 read weathered and original, eight read newly replaced (likely unpermitted reroofs you remove), and seven are ambiguous. You keep roughly 57 as live candidates.

Step 5. NOAA storm history shows two events over the subdivision in the last decade, one with reported hail around 1.75 inches. The whole tract sits inside that footprint, so every candidate gets a storm flag. This is now a high-priority subdivision: aging roofs, documented significant hail, and a clear set of homes that did not replace when their neighbors did.

Step 6. Ownership tags show 48 of the 57 are long-tenure owner-occupants. Those lead the list.

Result: a single afternoon of desk research produced 57 vetted doors, 48 of them prime, on one loop, with a storm narrative that justifies a free inspection at each. Compare that to sending the same crew to drive and guess. The research did not replace the inspection or the knock; it aimed them.

Run the economics on that, because it is the whole argument for doing the work. Say a rep knocks 40 doors in a morning. On a random subdivision with no research, maybe a third of those doors even have a roof worth talking about, and the rep burns most of the morning on homes that recently replaced, rentals, or roofs with years of life left. On the vetted loop above, nearly every door has an aging or storm-worn roof attached to a long-tenure owner. The contact-to-qualified ratio is not marginally better; it is several times better. If a rep books one inspection for every eight qualified conversations, the difference between a list where 12 of 40 doors are qualified and a list where 35 of 40 are qualified is the difference between roughly one inspection and four from the same morning of walking. The labor cost is identical. Only the aim changed.

There is a second-order benefit that does not show up in a single day's math: rep morale. Canvassers quit because the work feels like rejection roulette. A rep working a vetted, scored list hears 'we still have the original roof, come take a look' often enough that the job stays survivable. Retention of good canvassers is one of the hardest problems in this business, and feeding them better streets is one of the cheapest fixes.

Where this gets faster: per-roof age ranges and storm modeling

The workflow above is real and it works, but anyone who has run it across a whole metro knows the friction: assessor portals, permit systems, imagery, and storm databases all live in different places, in different formats, with different coverage and quality. Stitching them together by hand for one subdivision is a morning. Doing it for an entire territory, and keeping it current, is a full-time job that most companies never staff.

This is the gap RoofPredict is built to close. It is not a lead-buying service and it does not sell you a list of homeowners. It does the layering you just read about, at scale, and ranks roofs house-by-house so your crews knock the doors most likely to be due. Practically, that means two things working together:

• A roof-age range per address, estimated from aerial imagery rather than a single guessed date. The output is an honest band, not a birthday, because you cannot read an exact install date off a photo and any tool that claims otherwise is overselling. A range is what the imagery can actually support, and a range is enough to rank a street.
• Storm physics modeled per roof, layering historical hail and wind against each individual roof so you can see which roofs a given storm most plausibly wore out, expressed as odds rather than certainty. A modeled hit is a reason to inspect, never a substitute for the inspection or a claim of damage on any specific home.

Put those together and the seven-step workflow collapses into a ranked map: the roofs that are aging out by the calendar and the roofs a storm most likely accelerated, surfaced as routes instead of raw data. Your reps spend their hours knocking instead of researching, and the homes they knock are the ones the data says are due.

The honest limits are worth stating plainly, because over-promising here gets companies in trouble. An age range is a range; it will be wrong on individual homes that reroofed without a permit or installed an unusually durable product. A storm model gives odds; it does not prove damage and cannot tell you a roof has a valid insurance claim. Only a proper inspection establishes the actual condition of an actual roof, and only the homeowner's own insurer decides coverage. Used the right way, the data points your crews at the right streets and the inspection does the rest. That division of labor, data to target and inspection to verify, is exactly how the best storm-restoration and retail teams already operate; the tooling just makes it repeatable.

What experienced canvassers get wrong

Even crews that do research make the same handful of mistakes. Avoiding these is most of the edge.

Chasing the storm and ignoring the age. A fresh hailstorm pulls every company in the region into the same swath. The savvier move is to cross the storm footprint with roof age. Inside a hail swath, the old roofs are the easiest conversions and the ones competitors overlook while they blanket everything. Age is the filter that thins the competition.

Trusting aerial age as an exact date. Reps who treat a satellite read as gospel knock confidently on roofs that were quietly replaced last year and skip roofs that look okay from above but are failing underneath. Imagery is a ranking tool with a confidence band. Confirm with permits where you can, and let the inspection be the final word.

Knocking tile and slate streets. A 45-year-old tile neighborhood looks like a jackpot of 'old roofs' and is almost always a waste, because the tile outlives the homeowner and the real failure point is hidden underlayment. Identify the dominant material before you build the route and drop the long-life materials.

Ignoring effective year built confusion. Reps pull 'year built,' see an effective age that was reset by a remodel, and either skip a genuinely old roof or knock a new one. Use actual construction year for the skeleton and permits for the reroof history; never let effective age stand in for roof age.

Forgetting permit blind spots. Treating 'no permit' as absolute proof of an original roof leads to dead doors where unpermitted reroofs happened. Treating 'has permit' as absolute proof of a new roof is safer but still misses repairs versus full replacements. Permits are strong evidence, not certainty.

Skipping the legal and courtesy layer. No-soliciting ordinances, registries, gated communities, and HOA rules are real, and ignoring them burns goodwill and occasionally gets a rep cited. Build the no-knock list into the route before the crew leaves. Door-to-door sales are regulated; know your local rules and any applicable consumer-protection requirements like the buyer's right to cancel on in-home sales.

Confusing 'old roof' with 'easy sell.' An old roof is a qualified prospect, not a closed deal. The homeowner with a 22-year-old roof has lived under it for 22 years and may be perfectly comfortable waiting until it actively leaks. Research finds you the roofs that should be replaced; it does not find you homeowners who have decided to act. That gap is closed at the door with a real inspection, photo documentation, and an honest condition report, not by the data. Reps who expect the list to sell for them get discouraged fast. The list earns the conversation; the rep earns the job.

Over-weighting the newest storm. There is a reflex to chase only the most recent event because it is fresh in everyone's mind. But roofs damaged in a storm two or three years ago, on homes that never inspected, are often better prospects than a brand-new swath crawling with competitors. Old damage does not heal. A roof compromised by hail two seasons back is still compromised, and that homeowner has had no one knock with a credible reason to look. Your storm-history layer covers years rather than only last week, for exactly this reason.

Pitching condition you have not verified. The fastest way to a complaint, and to trouble with state regulators, is telling a homeowner their roof is damaged or their claim will be approved before anyone has inspected the roof. The roofer documents conditions and provides an estimate. The insurer decides coverage. The homeowner owns the claim. Stay on the right side of that line and your storm narrative stays an invitation to inspect, not a promise.

A canvassing-prep checklist you can copy

Before a crew rolls into a new neighborhood, the desk work should be done. Run this list:

• Subdivision build year confirmed (plat year and parcel sampling agree)
• Dominant roofing material identified (asphalt = go; tile/slate = usually drop)
• Expected roof life mapped against build year (is the tract in or past its first cycle?)
• Reroof permits pulled; recently reroofed addresses removed
• Aerial imagery scanned; roofs ranked by visual age band
• New-looking roofs without permits flagged as 'probably replaced'
• NOAA/SPC storm history pulled; storm-flagged streets moved up
• Ownership tagged (owner-occupied vs. absentee, length of tenure)
• No-solicit registries, HOA, and gated-community constraints loaded into the route
• Route sequenced to front-load highest-confidence streets and minimize backtracking
• Reps briefed on the per-door narrative (old roof, storm flag, ownership note)

If every box is checked, the crew walks into a neighborhood already knowing which roofs are most likely due and why. That is the whole point: spend the cheap hour on research so you do not spend the expensive day guessing.

Reading roofs from above: a deeper field guide

Because aerial imagery is the layer reps most often misread, here is more detail on translating what you see into a defensible age band.

Color and saturation. New asphalt shingles read deep and even, whether the color is charcoal, weathered wood, or driftwood. As granules shed, the underlying asphalt and fiberglass mat show through unevenly, and the roof lightens and goes blotchy. A roof that has gone from rich to chalky is well into its second decade in most climates.

Slope-by-slope differences. The south and west slopes take more sun and weather faster. On older roofs you will often see one slope noticeably more faded than another on the same house. That asymmetry is itself an age signal: a brand-new roof is uniform; an old one is not.

Granule loss at valleys and edges. Water concentrates in valleys and at eaves, accelerating wear there. On imagery, darker, shinier streaks down valleys can indicate exposed asphalt where granules washed away. Combined with overall fade, that pushes the band older.

Hail-pattern clues after a storm. Severe hail can leave a roof looking subtly speckled or scoured on imagery captured soon after an event, and it often triggers a wave of tarps and replacements visible across a tract. If you see a neighborhood where replacements suddenly cluster in one recent period, a storm drove them, and the homes that did not replace are your list.

Adjoining clues. Look at gutters, skylights, and chimney flashing where resolution allows. Rust and staining around penetrations track with overall roof age and neglected maintenance.

The discipline is to convert all of this into a band: 'this roof reads 18-24 years and original,' not 'this roof is 20.' Reps who internalize the band instead of a false-precision number make better calls and get embarrassed less often at the door.

A note on imagery freshness. Aerial and street-level imagery is not live; it was captured on some date that can be anywhere from months to a few years old depending on the provider and the area. Always check the capture date if the tool exposes it, and assume a roof could have changed since. A street-view image from three years ago will not show last year's replacements or last month's storm tarps. This is a real failure mode: a rep down-ranks a roof that 'looks original' on stale imagery, not realizing it was replaced eighteen months ago, and wastes the door. When imagery and permit data disagree, trust the more recent source, and let the physical drive-by or inspection settle ties. Stale imagery is still useful for the build-skeleton read, because granule fade does not reverse, but it is unreliable for catching very recent changes.

Resolution sets your ceiling. The amount of detail you can read depends entirely on the imagery resolution available for that area, which varies a lot between dense metros and rural fringe. In a high-resolution metro you can pick out individual curling shingles and valley wear. On low-resolution rural imagery you may only get overall color and obvious tarps. Calibrate your confidence to what the pixels actually support, and lean harder on permit and build data where the imagery is coarse.

Storm timing and the replacement decision window

Understanding when homeowners actually replace helps you time your knock. A roof rarely gets replaced the day it turns old. It gets replaced when something forces the decision: a leak, a real-estate transaction, a neighbor's new roof, or a storm.

Storms are the biggest forcing function in hail and high-wind regions. After a significant event, a neighborhood splits into three groups within roughly 6 to 18 months:

• Fast movers who inspect, file with their insurer if warranted, and replace quickly. These are often gone before slower companies even canvass.
• Fence-sitters who know something happened but have not acted. This is the richest group for a well-prepared rep, and the window can stay open for a year or more, sometimes bumping against insurer reporting timelines that vary by policy and state.
• Non-movers who never inspected, often because no one knocked or they assumed they were fine. Old roofs in this group, especially original roofs that skipped the post-storm replacement wave their neighbors took, are exactly what your layered research surfaces.

The practical lesson: the best old-roof targets in a storm region are frequently the homes that did not replace when their street did. Your imagery shows them as the weathered roofs surrounded by new ones. Your permit data confirms no reroof. Your storm history explains why the inspection conversation is timely. That trio is the most repeatable setup in storm-restoration canvassing.

Building a territory you can work for years

Finding one good neighborhood is a day. Building a territory is a system. A few principles keep the pipeline full past the first hot street:

Track what you find. Every time you research a subdivision, save the build year, material, permit status, storm flags, and outcome. Over a season you build a map of your whole service area graded by roof age and storm exposure. Next year's research is faster because last year's is on file, and you can revisit subdivisions as they age into the window.

Re-clock aging tracts. A subdivision that was 14 years old and too new last year is 15 this year and entering the window. Keep a calendar of tracts crossing into prime age so you are first on the street when it ripens, not last.

Watch new storms against your map. When a storm hits, you should already know which neighborhoods inside the footprint carry old roofs, because you mapped them before the storm. That lets you surge intelligently in the first 48 hours instead of blanketing and burning labor.

Respect the homeowner and the law. A territory is a long-term relationship with a community. Reps who knock honestly, document conditions truthfully, leave a clean estimate, and never overstate damage or coverage build a referral base that outlasts any single storm. Reps who oversell torch a neighborhood and sometimes the company's license with it. The data tells you where to go; your integrity at the door determines whether you can go back.

Work the same streets in waves. A subdivision rarely converts on one pass. The first visit you catch the homeowners already worried about their roof. The fence-sitters need a second touch weeks later, often after they have watched a neighbor's crew work. A door-hanger left on the no-answers, a follow-up pass, and a callback list turn a single knock into a campaign. Companies that knock a street once and write it off are leaving most of the qualified doors for whoever comes back. Your research told you those roofs are old; that fact does not expire because nobody was home on Tuesday.

Mine your own job history. Your closed jobs are a free training set. Pull every replacement you have sold in the last two years, look up the build year, material, permit history, and storm exposure of each, and find the common profile. Most companies discover their bread-and-butter job is a specific vintage and material in a specific band of neighborhoods. Once you know that profile, you can go find every other subdivision in your territory that matches it. You are no longer guessing what 'a good neighborhood' looks like; you are cloning the ones that already paid you.

A quick territory-grading rubric

When you scout a new area, grade it fast on five questions before investing real research time:

1. Is the dominant material asphalt shingle? (If tile/slate, stop.)
2. Is the housing stock in or just past one roof cycle by build year?
3. Is it a tract built in waves, or scattered custom homes? (Waves are better.)
4. Has it taken at least one significant hail or wind event in the last decade?
5. Is it mostly owner-occupied with reasonable tenure?

Five yeses is an A-grade territory worth a full workup. Three or four is worth a look. Two or fewer, move on. This rubric is the thirty-second filter that keeps you from pouring an afternoon of parcel research into a neighborhood that was never going to convert.

Bringing it together

Finding neighborhoods with old roofs stopped being guesswork the moment property data, imagery, and storm history all came online. The method is straightforward even if the execution takes discipline: define what 'old' means for the material in front of you, build the build-year skeleton from assessor and plat data, subtract the recently reroofed using permits, confirm and rank with aerial imagery, overlay verified storm history to find the roofs that aged early, filter by ownership to find who can act, and sequence it all into a route your crew can walk.

Do that by hand and one subdivision is a productive morning. Do it across a metro and keep it current, and you are running a research department most companies cannot staff. That is the work RoofPredict automates, returning a per-address roof-age range and per-roof storm modeling as ranked routes, with the honest understanding that ranges are ranges and storm odds are odds, and that the inspection, not the data, establishes the truth about any individual roof.

Either way, the principle holds. Spend the cheap hour pointing your crews at the roofs the calendar and the weather say are due, and let the knock and the inspection do what only they can. The street that pays for the truck is out there every week. The difference between the companies that find it on purpose and the ones that stumble onto it once a year is nothing more than the research they were willing to do before they parked.

FAQ

What is the single best free way to find old roofs in a neighborhood?

Start with county assessor or GIS year-built data layered against expected roof life. Tract subdivisions are roofed in one wave, so the build year plus the expected life of the dominant material (roughly 15-20 years for 3-tab asphalt, 25-30 for architectural) tells you when a whole street comes due. It costs nothing and narrows your search before you ever drive.

Can I tell exactly how old a roof is from satellite or aerial imagery?

No. Imagery gives you a confidence band, not an install date. Granule loss, color fade, streaking, curling, and patches let you rank roofs as relatively old or new and estimate a range, but ventilation, slope orientation, and product quality all shift real age. Use imagery to rank and prioritize, then confirm with permit data and a physical inspection.

How do building permits help me find replaceable roofs?

A reroof usually requires a permit, and many jurisdictions publish permit records. An old subdivision where homes have no reroof permit on file likely still has original roofs, which are prime targets. Homes with recent reroof permits should be skipped. Permit data is strong evidence but not absolute, since some reroofs happen without a permit, so confirm with imagery and inspection.

How do I combine storm history with roof age to pick the best streets?

Pull NOAA Storm Prediction Center and Storm Events data for hail and wind in your area over the last 10-15 years, then overlay those footprints on your old-roof neighborhoods. The overlap, old roofs inside a significant-hail or high-wind swath, is the highest-value target because the roofs are both worn out and freshly stressed. Treat a storm report as a reason to inspect, never as proof of damage to any specific home.

What hail size actually damages asphalt shingle roofs?

Functional shingle damage generally starts to appear once stones reach roughly one inch (about quarter size), with larger stones and higher fall velocity doing progressively more harm. Smaller hail is more likely cosmetic. Reported hail size from NOAA helps you judge whether a given storm plausibly damaged roofs, but only an inspection confirms damage on a particular roof.

Why are tract subdivisions better canvassing targets than mixed neighborhoods?

Tract subdivisions were built and roofed in a single season by one or two builders using the same product, so the entire street ages in unison and comes due together. That gives you many qualified doors on one loop. Mixed-age neighborhoods require vetting house by house and produce scattered, less efficient routes.

Should I avoid neighborhoods with tile or slate roofs?

Usually, yes, for replacement canvassing. Tile and slate can last 50 years or more, and their real failure point is hidden underlayment rather than the visible surface, so an old-looking tile street rarely converts to a full replacement. Identify the dominant material before building a route and focus on asphalt-shingle tracts.

What does RoofPredict do that I cannot do myself with free data?

You can assemble year-built, permit, imagery, and storm data by hand for one subdivision in a morning. RoofPredict does that layering at scale and ranks roofs house-by-house, returning a roof-age range per address from aerial imagery plus storm physics modeled per roof, delivered as routes. It does not sell leads or claim exact dates or prove damage; ranges are ranges and storm odds are odds, and the inspection still establishes the truth on any roof.

How long after a hailstorm do homeowners stay good targets?

A neighborhood typically splits within about 6 to 18 months into fast movers who replace quickly, fence-sitters who know something happened but have not acted, and non-movers who never inspected. Fence-sitters and old-roof non-movers stay good targets for a year or more, though insurer reporting timelines vary by policy and state, so timing matters.

What are the legal limits I should know before knocking doors?

Door-to-door sales are regulated. Respect local no-soliciting ordinances and registries, HOA and gated-community rules, and consumer-protection requirements such as the buyer's right to cancel certain in-home sales. Critically, never tell a homeowner their roof is damaged or their insurance claim will be approved before an inspection. The roofer documents conditions and estimates, the insurer decides coverage, and the homeowner owns the claim.