Which Neighborhoods Have the Oldest Roofs (And How to Find Them House by House)

Every roofing owner I've ever talked to has the same map in their head: a few subdivisions they know are due, a couple they suspect, and a giant gray area of "somewhere out there" where the jobs probably are but nobody can point to the street. The question underneath all of it is the one you typed to get here. Which neighborhoods have the oldest roofs? Because if you knew that with any precision, you'd point your crew, your mailer, and your ad budget straight at them and stop spreading yourself across a whole city of roofs that mostly don't need you yet.

Here's the honest version, from years of watching this play out: "oldest neighborhoods" is the right instinct and the wrong unit. Roof age clusters by neighborhood because subdivisions get built in waves, but it does not stay clean. One street in a 1998 development got hit by a 2019 hailstorm and is now full of five-year-old roofs. The cul-de-sac behind it never got touched and is sitting on original 26-year-old shingles. The neighborhood-level answer gets you to the right side of town. It does not get you to the right doors, and the doors are where the money is.

So we're going to do both. First, how to actually find the older neighborhoods in your market using public data anyone can pull — build-year maps, permit records, county assessor data, and a handful of aerial tells that read roof age from the sky. Then how to drop from the neighborhood down to the house, which is where most roofers leave thousands of dollars on the table by knocking a whole subdivision instead of the third of it that's worn out. I'll give you the workflows, the numbers, the edge cases, and the mistakes I see good companies make. By the end you'll have a repeatable system, not a hunch.

Why "oldest neighborhoods" is the right instinct (and where it breaks)

Roofs don't age randomly across a city. They age in cohorts, and the cohort is the subdivision. A developer buys raw land, builds 180 homes over 18 to 30 months, and roofs them all with whatever shingle was spec'd for that phase. Those 180 roofs are effectively the same age, installed by the same handful of crews, sometimes from the same pallet. That's why a neighborhood-level signal works at all: find the build wave, and you've found a few hundred roofs that crossed the same line in the same year.

The standard asphalt shingle most of those homes wear is a 3-tab or an entry architectural shingle. Manufacturers rate those products for a couple of decades, and field life depends heavily on climate, slope, ventilation, and storm exposure. The National Roofing Contractors Association is blunt about this in its homeowner guidance: a manufacturer's warranty term is not a prediction of service life, and roofs routinely need replacement before or after the printed number depending on conditions. So when you hear "20-year shingle," read it as "a roof that enters the replacement conversation somewhere in a window around two decades, give or take a lot."

That window is your opening. A subdivision built in 1999 to 2001 with original shingles is, as of the mid-2020s, squarely in the replacement zone — old enough that sun, thermal cycling, and granule loss have done their work, and any decent storm in the last few years pushed a chunk of those roofs over the edge. A subdivision built in 2016 is not your market, no matter how nice the houses are.

Where the instinct breaks is uniformity. Three things scramble the neat "old neighborhood = old roofs" map:

1. Storm re-roofs. One hailstorm runs through a 25-year-old subdivision, insurance pays for a swath of replacements, and now half the neighborhood has new roofs and half doesn't — and the split follows the storm's path, not the property lines. The houses that didn't qualify or didn't file are still original. The neighborhood looks "done" from a windshield survey and absolutely is not.
2. Phased builds. Big developments go up in phases over five to eight years. Phase 1 and Phase 5 can be most of a decade apart. "This neighborhood" might be three different roof-age cohorts wearing the same entrance sign.
3. Pride re-roofs and flips. A scatter of homeowners replace early — a leak, a sale, an HOA push, a buyer's inspection. These are random across the map and they pepper holes in any clean cohort.

Keep that picture in your head: the neighborhood gets you to a cohort, the cohort has holes shot through it, and your job is to find the cohort and then find the holes so you skip them. Now let's build the find.

The data that actually tells you a roof's age

Before the workflow, you have to know what each data source actually says, because the single most expensive mistake in roof targeting is treating year-built as roof age. Let's separate the signal from the noise.

Year built is the floor, not the answer

Zillow, Redfin, the county site, and most public listings show year built. Year built is the year the house was constructed. On a brand-new house, the roof age equals the house age — those line up exactly once. After that first re-roof, they diverge forever, and year built tells you nothing about the current roof except that the roof cannot be older than the house. It's a floor.

That sounds like a knock on year-built data. It isn't — it's the most powerful free signal you have for finding old neighborhoods, precisely because it reveals build cohorts. Just don't confuse the cohort signal (great) with the per-house roof age (which year-built can't give you once a re-roof has happened). Use year built to find the wave. Use other tools to find which houses in the wave still wear the original roof.

County assessor and parcel data: the bulk-cohort tool

Your county assessor or property appraiser publishes year built, square footage, ownership, and last sale for essentially every parcel. The U.S. Census Bureau's data on the age of the national housing stock confirms what you already feel in the field — a huge share of American homes were built in identifiable decade-long waves, which is exactly why cohort targeting works. Many counties let you download or query parcels in bulk and filter by year-built range. That single capability — "show me every parcel built 1998 to 2003 in this set of zip codes" — is the backbone of finding old neighborhoods at scale, and it's free or nearly free.

A practical note: assessor "year built" can be the year of the original structure even after major additions, and data entry quality varies by county. Treat it as accurate to within a year or two, which is plenty for cohort work.

Building permits: the closest thing to a roof's birth certificate

Here's the source most roofers underuse. When a roof gets replaced legally, somebody pulls a re-roofing permit. Most jurisdictions require one, and many publish permit records online — searchable by address, date, and permit type. A re-roof permit is the cleanest single record that a roof was replaced and roughly when.

Flip the logic and you have gold:

• A house in a 1999 cohort with no re-roof permit on record is a strong candidate for an original, ~25-year-old roof.
• A house in that same cohort with a 2020 re-roof permit is a five-year-old roof — skip it, it's not your job for years.

No permit system is complete. Some re-roofs happen without a permit (smaller jobs, lax enforcement, owner-pulled work that never closes out), and historical digitization varies. So "no permit" is a strong lean, not a guarantee — pair it with the build cohort and an aerial look. But permit data is the only public source that directly timestamps roof work rather than house construction, and learning your county's permit portal is one of the highest-leverage afternoons a sales manager can spend.

Aerial and satellite imagery: reading age from the sky

You can read a surprising amount of roof age off good aerial imagery without leaving your desk. The visual tells of an aging asphalt roof:

• Color fade and unevenness. A fresh architectural shingle reads dark and uniform. As granules wear, the surface lightens, mottles, and goes patchy — especially on south- and west-facing slopes that eat the most sun.
• Streaking and algae. Dark vertical streaks (Gloeocapsa magma) signal a roof that's been up long enough to grow biology — common on older roofs in humid regions.
• Surface texture. Curling, cupping, and granule loss can show as a rougher, broken-up texture versus the crisp lines of a newer roof.
• Patches and mismatches. A section that's a different shade is a repair or a partial — and a story worth knocking on.
• The contrast test. This is the most useful move: compare a roof to its neighbors in the same cohort. In a 1999 subdivision, the three roofs that read noticeably newer than the block are almost certainly recent re-roofs. The rest of the block reading uniformly weathered? That's your cohort, intact.

Imagery has limits. A clean, well-ventilated 18-year-old roof in a dry climate can look better from the sky than a poorly-vented 10-year-old one. Recency of the imagery matters — if the photo is three years old, it predates recent storm re-roofs. And you cannot reliably read exact age off a picture; you read a range and a relative ranking against the block. Which is exactly how you should think about roof age anyway: not a birthday, a window.

What none of these sources can do alone

Notice the gap. Year-built finds cohorts but goes blind after the first re-roof. Permits timestamp replacements but miss the un-permitted ones and the gaps in digitization. Aerial reads condition but only to a range, and only as fresh as the photo. Each source is strong exactly where the others are weak. The whole game — and the rest of this piece — is stacking them so the strengths cover the gaps. A house that's in an old build cohort, has no re-roof permit, and reads weathered from the air is about as confident an "original old roof" call as you can make from a desk. That's a door you knock first.

Step-by-step: find the oldest neighborhoods in your market

Here's the workflow I'd hand a new sales manager on day one. It takes a focused afternoon to set up and a couple hours to refresh per market. Work it in order.

Step 1: Define the market and pull the build-year map

List the zip codes you actually service — the ones a crew can reach without an hour of windshield time. For each, pull parcel data from the county assessor and filter to a target year-built range. A good default starting window in the mid-2020s is roughly 1990 to 2008: old enough that original roofs are deep in the replacement zone, recent enough that the homes are still standard re-roof candidates rather than full restorations. Adjust for your region's storm history and your typical job.

What you want out of Step 1 is a count of homes per build-year band per zip. You'll immediately see the pattern: a few zips light up with hundreds of homes in the 1998 to 2003 band, others are mostly post-2012. The lit-up zips are your hunting grounds.

Step 2: Cluster into actual neighborhoods, not zip codes

Zip codes are postal routes, not neighborhoods. Inside one zip you'll have a 1999 subdivision next to a 2015 one. Drop your filtered parcels onto a map and look for spatial clusters of same-era homes — those are real subdivisions. Most mapping tools color points by a field, so color by year-built band and the cohorts pop out as blobs of one color.

Name each cluster something your team will recognize — the subdivision name, a major cross street, an entrance landmark. "The Wexford Hills cohort, ~1999, 210 homes" is a usable target. "Parts of 75234" is not.

Step 3: Rank cohorts by opportunity, not by age alone

Oldest is not automatically best. Rank each cohort by a blend of factors:

• Roof age band — how deep into the replacement window the original roofs are.
• Density — homes per square mile. A tight 220-home subdivision beats a sprawl of large lots for knocking and route efficiency; your reps spend time at doors, not driving.
• Re-roof saturation — roughly what share already looks replaced (you'll estimate this in Step 4). A cohort that's already 70% re-roofed has thin pickings even if it's old.
• Home value and roof complexity — fit to your average ticket and crew capability. Steep, complex, high-value roofs are a different sale than a tract ranch.
• Drive time — proximity to your other active work so a crew chains jobs instead of crossing the city.

Score each cohort, even roughly, and you've turned a vague "older part of town" into a ranked target list. Top of the list is where the first knock and the first mailer go.

Step 4: Estimate re-roof saturation per cohort

This is the step that separates pros from spray-and-pray. For each top cohort, get a read on how much of it has already been re-roofed, because that's wasted effort you want to subtract before you spend a dime. Two ways, used together:

• Permit pull. Query your county permit portal for re-roof permits inside the cohort's date range and footprint. The count of permitted re-roofs against the cohort's home count gives you a saturation floor.
• Aerial scan. Open recent imagery over the cohort and eyeball the ratio of weathered roofs to obviously-newer ones using the contrast test from earlier. This catches the un-permitted re-roofs the portal missed.

A cohort that's 25 years old but only ~15% re-roofed is a rich target. One that's the same age but ~60% re-roofed has been worked — by storms, by competitors, or both — and you'll burn a lot of doors to find the remaining originals. Knowing this before you deploy a crew is the entire point.

Step 5: Layer in storm history

Old roofs and storm-worn roofs are two overlapping reasons a house is due, and the overlap is the sweet spot. NOAA's National Centers for Environmental Information and the NWS Storm Prediction Center publish historical severe-weather and hail data you can use to see which of your cohorts sat under significant hail or wind in recent years. The Insurance Institute for Business and Home Safety has extensive research on how hail and wind damage roofing materials — worth understanding so you know what a worn-then-beaten roof actually looks like.

The move: overlay storm tracks on your ranked cohorts. A 22-year-old cohort that also took a real hail event three years ago is a different animal than a 22-year-old cohort in a calm pocket — the storm accelerated aging across roofs that were already marginal. Just keep the framing honest with yourself and the homeowner: storm history tells you a roof was exposed and is worth inspecting; it is not proof of damage on any specific house. Damage is determined on the roof, and coverage is determined by the homeowner and their carrier. More on that boundary later, because it matters legally.

Step 6: Build the route and the skip list

Now you've got ranked cohorts with a saturation read and a storm overlay. Turn it into field instructions: an ordered list of cohorts, and within each, the streets to work first. Hand your crew the cohort, the talking point ("this neighborhood went up around '99, most of these roofs are original and at the end of their life"), and — critically — the houses to skip because they're already re-roofed. A canvasser who isn't wasting knocks on five-year-old roofs covers the productive doors twice as fast and quits a lot less often.

That's the neighborhood-level system end to end. It's repeatable, it's mostly free data, and it'll already put you ahead of any competitor still driving around looking for "old-looking" houses. But the saturation problem in Step 4 hints at the real frontier, so let's go there.

From neighborhood down to the house: the part most roofers skip

Here's the uncomfortable math. Take a 25-year-old, 200-home cohort that's 40% re-roofed. If you knock the whole neighborhood, 80 of your 200 doors are dead — five-to-fifteen-year-old roofs that won't buy from you for years. You're paying a rep (or postage) to reach 200 houses to find 120 real opportunities. Your effective cost per real door is 67% higher than it needs to be, your reps are getting "not interested" from people whose roofs are genuinely fine, and the morale tax on a new canvasser eating those no's is brutal.

The neighborhood signal got you to the cohort. It cannot tell the 120 from the 80. That's a per-house question, and answering it is where targeting goes from good to surgical.

What "per house" actually requires

To separate the real doors from the dead ones inside a cohort, you need three things resolved at the address level, not the neighborhood level:

1. A roof-age estimate for that specific roof — not the house's build year, the roof's age, expressed honestly as a range. "This roof reads 18 to 22 years" is an actionable, defensible statement. "This house was built in 1999" is not, because it ignores the re-roof.
2. Whether it's already been replaced — the permit/aerial check, resolved per address so you can hard-skip the new roofs.
3. Storm exposure on that roof — whether this roof, at this location, sat under hail or damaging wind, modeled tight enough to matter at the house level rather than "somewhere in the county it hailed."

Stack those three per address and you can rank every house in a cohort from "knock this first" to "don't bother." That ranked list is the actual product of all this work. The neighborhood map was the means; the ranked door list is the end.

The manual version, and its ceiling

You can do this by hand, and for a single hot cohort it's worth it. Pull the cohort's parcels, cross-reference each against permit records, eyeball each roof on recent aerial imagery, mark the obvious re-roofs as skips, and note the weathered ones as priorities. For 200 homes that's a few hours of careful desk work, and it'll dramatically outperform blind knocking.

The ceiling is obvious the moment you try to scale it. Doing that across 15 cohorts and 4,000 homes, refreshed as new imagery and storms come in, is a full-time job nobody on your team has. And the hardest of the three signals — modeling whether a specific roof actually caught a storm, versus just sitting in a county that had weather — is genuinely hard to do by eye. A hail map shows you where it hailed. It does not show you which roofs the hail actually wore out, because the same storm hits a steep south-facing slope very differently than a shallow north one next door. That gap is where per-house data tools earn their keep.

Where a per-roof data tool fits (an honest look)

This is the section where I tell you what software does and doesn't do for this problem, because the per-house step above is exactly what tools like RoofPredict are built to handle, and you should understand both the value and the limits before you lean on any of them.

The core idea is to do the address-level stack from the last section automatically across a whole area. You hand it your market, and it scores every roof on two things that matter most:

• Roof age as a range, estimated from aerial imagery rather than build year — so it sees past the re-roof that breaks year-built data. The output is a window like "18 to 22 years," not a fake-precise birthday, because honestly a range is all anyone can defend from imagery.
• Storm impact modeled per roof, not per county. Instead of "it hailed in this zip," it models the hail and wind against each individual roof — orientation, slope, exposure — to estimate which roofs a storm actually beat up versus which it mostly missed. That's the difference between a weather lookup and per-roof physics, and it's the part that's nearly impossible to do well by hand.

Put those together and you get the ranked door list directly: every roof in the area sorted by how due it is, with the already-re-roofed houses pushed to the bottom so your crew and your mail skip them. The pitch to your reps becomes concrete — "this roof reads 18 to 22 years and took two hail events; here's the homeowner report" — which is a wildly different door than "hi, we're roofers, how's your roof." A green canvasser knocking pre-ranked doors with a real talking point sounds like a veteran and, more importantly, makes money and stays.

Now the honest limits, because you should not buy any tool — this one included — on hype:

• It's a range, not an X-ray. A roof-age estimate from imagery is a window. A clean roof can read younger than it is; storm exposure is modeled odds, not a guarantee that a given roof is damaged. You still verify on the roof.
• Storm modeling is probability, not proof. "This roof likely took impact" is a reason to inspect, never evidence to wave at a homeowner or an insurer. Damage gets determined by an inspection; coverage gets determined by the homeowner and their carrier. A tool that tells you which roofs to look at is doing its job; it is not adjudicating anything.
• Imagery recency still bounds it. No model sees a re-roof that happened after the last aerial capture. Fresh ground-truth always wins.
• It sharpens outbound; it isn't leads. This is targeting — which of your doors to knock and which to skip — not a list of strangers sold to five competitors at once. It makes the prospecting you already do far more precise. It doesn't replace the knock, the inspection, or the sale.

Used inside the workflow above, the value is plain: it collapses the multi-hour manual per-house stack into a ranked list, and it adds the one signal you genuinely can't eyeball — per-roof storm physics. Used as a magic "give me customers" button, it'll disappoint you, same as any tool would. Target with it, then go do the work.

Worked example: turning a zip code into a route

Let me run a realistic example end to end so the abstraction has bones. Numbers are illustrative — plug in your own.

Market: Three adjacent zips you service, call them A, B, and C.

Step 1 — build-year pull. Assessor data, filtered to homes built 1990 to 2008:

Zip A and C are your hunting grounds; B is mostly newer and goes to the back of the line.

Step 2 — cluster into cohorts. Mapping the points by year-built band, Zip A breaks into three real subdivisions:

Step 3 — rank. Wexford Hills is oldest (~25 yr roofs), tight and dense, mid-value tract homes your crew runs all day. Stone Creek is younger (~19 yr) but bigger and starting to enter the window. Lakeview is too new to prioritize yet. Wexford ranks first.

Step 4 — saturation. Permit pull on Wexford's footprint, 1998–2024, returns 48 re-roof permits. Aerial scan suggests roughly another 20 re-roofs with no permit on record. Call it ~68 of 220 already replaced — about 31% saturated. That leaves ~152 likely-original roofs, deep in the replacement window. Rich.

Step 5 — storm overlay. Historical hail data shows a significant event crossed Wexford ~3 years ago. So those 152 original roofs aren't just old — they're old and recently beaten. That's your A-priority cohort, confirmed.

Step 6 — route and skip list. Without per-house resolution, you'd hand the crew "knock Wexford Hills, 220 homes" and let them eat ~68 dead doors. With per-house resolution — manual stack or a tool — you hand them the 152 ranked addresses, skip the 68 re-roofs, and lead every knock with "this neighborhood went up around '99, most roofs are original and worn, and a storm came through three years back; mind if I take a look?"

Same neighborhood. Same crew. One version wastes a third of its knocks and grinds reps down; the other points them only at doors that can actually buy. That delta — multiplied across every cohort, every season — is the whole reason this work pays.

Common mistakes pros make (and how to avoid them)

Even good companies fumble this. The repeating ones:

Treating year built as roof age

The original sin. A 1999 house with a 2021 re-roof is a 1999 house and a four-year-old roof. If your targeting can't tell those apart, you'll knock new roofs all day. Always resolve re-roofs via permits and aerial before you call a house "old."

Knocking the whole subdivision

Covered above, but it's worth repeating because it's the most expensive habit in the trade. The neighborhood is the cohort; the cohort has 20 to 60% of its roofs already replaced. Subtract them before you deploy or you're paying full freight for partial opportunity and burning out your reps on doors that were never going to buy.

Chasing only the absolute oldest

The 1985 cohort sounds like the best target until you pull it and find it's 70% re-roofed, full of second-generation roofs, and worked to death by every competitor for two decades. A slightly newer cohort that's mostly original and unworked often out-produces the oldest one. Rank by opportunity, which blends age with saturation, density, and fit — not by age alone.

Using stale imagery

If your aerial is three years old, it's blind to every re-roof and every storm since. You'll mark replaced roofs as targets and miss fresh storm damage. Always check the capture date, and weight recent ground-truth — a rep's eyes on the street — over an old photo.

Ignoring storm history on old cohorts

Age and storm exposure compound. An old cohort that also took recent hail has roofs that crossed the line fast. Roofers who target on age alone miss that the storm did half their qualifying for them — and roofers who target on storms alone miss that a fresh roof under that same storm probably shrugged it off. You want the overlap.

Crossing the legal line on storm claims

This one can cost you a license, so read it carefully. When you knock storm-worn old roofs, your job is to document conditions and provide an estimate. It is not to interpret, file, manage, negotiate, or guarantee an insurance outcome. Telling a homeowner their roof is damaged and you'll "get their claim approved" or talking about their deductible can put you over the line into unlicensed public adjusting in many states — the Stonewater Roofing matter in Texas (2024) is a pointed reminder that even calling yourself a claims "specialist" can violate those rules. Stay in your lane: you document what you see, you give an honest estimate, the homeowner owns the claim, and the insurer decides coverage. Use storm data to decide which roofs to inspect, never as proof of damage to wave at anyone. When in doubt, get counsel for your state before any claims-adjacent language goes on a flyer or out of a rep's mouth.

Refreshing never

A cohort map built once decays. Roofs get replaced, storms hit, new imagery drops. Treat your target list as living — re-pull permits and re-scan imagery on a cadence (quarterly is reasonable for active markets) so you're not knocking last year's map.

Matching the channel to the cohort: knock, mail, or your old list

Finding the old cohorts is half the win. The other half is deciding how to work each one, because a ranked door list feeds three very different motions, and the smartest companies route each cohort to the channel it fits.

Door knocking

Knocking is highest-conversion and highest-cost-per-touch, so spend it on your densest, highest-priority cohorts — the ones where age, low saturation, and recent storm exposure all line up. The economics only work when your reps aren't wasting steps. A pre-ranked door list with the re-roofs stripped out lets a rep work the productive houses back to back instead of grinding down a whole street of mixed roofs. Give each rep the cohort's age story and a per-house talking point so a green canvasser walks up sounding like someone who already knows the roof, because the difference between "how's your roof today" and "this neighborhood went up around '99 and most of these are original" is the difference between a door slammed and a ladder going up.

Direct mail

Mail is lower-cost-per-touch and scales to cohorts that are too spread out to knock efficiently — larger lots, longer drives, or simply more homes than a crew can cover before the season turns. The targeting discipline is identical: you are paying for every piece, so every piece that lands on a five-year-old roof is wasted postage. Mail the likely-original roofs in a cohort, suppress the re-roofs, and your cost per genuine opportunity drops hard. A useful frame for owners: one re-roof is worth thousands, so the only reason to mail a house is that its roof can actually buy one. Build-cohort plus permit-suppression plus an aerial read is what lets you mail tight instead of carpet-bombing a zip and hoping.

Your own old list

The cheapest "cohort" you own isn't on any map — it's in your CRM. Every estimate you ever wrote and didn't close, every customer from years back, every inspection that didn't convert. Run those addresses through the same age-and-storm lens. A homeowner you quoted six years ago on a then-15-year-old roof is now sitting on a 21-year-old roof, possibly storm-worn, and you already have their name, their address, and a relationship. That's found money with zero ad spend, and it's the one motion that never feels like renting customers from anyone. Re-scoring your old book on roof age is often the single fastest return in this entire process, because the prospecting cost was already paid years ago.

The point across all three: the data you build to answer "which neighborhoods have the oldest roofs" isn't a knock tool or a mail tool or a CRM tool. It's a targeting layer that makes all three sharper. Decide the channel by cohort density and drive time; decide the doors by the per-house stack.

Setting up the data without a data team

Most roofing companies don't have an analyst, so here's how to run all of the above with the people you already have.

The assessor pull. Nearly every county property appraiser has an online search, and a growing number offer bulk export or an open-data portal. If yours doesn't export cleanly, a sales manager can still search by subdivision and read year-built ranges directly. For multi-county markets, the data lives in several portals, so keep a simple per-county notes file: which site, how to filter year built, whether bulk export exists. That reference saves hours every refresh.

The permit portal. Search your jurisdiction's building department site for "permit search" or "permit portal." Learn two things: how to filter to roofing or re-roof permit types, and how to search by address or area. Some portals are clunky and some charge for bulk records, but even address-by-address lookups on your top cohort are worth the time, because no other public source timestamps a roof replacement.

The aerial layer. Free satellite and aerial imagery covers most markets, and the key habit is checking the capture date before you trust what you see. For a single hot cohort, a careful desk scan comparing each roof to its block is entirely doable by one person in an afternoon.

The storm layer. Pull historical hail and wind events for your area from the public NOAA and Storm Prediction Center tools, note the dates and rough footprints of significant events in the last several years, and keep them handy to overlay on cohorts.

Where it gets hard, and where to spend money. The manual stack works beautifully for one cohort and collapses under scale. The two genuinely hard parts are resolving roof age per address across thousands of homes, and modeling whether a specific roof caught a storm versus just sitting in a county that had weather. Those are the points where a per-roof data tool earns its cost — not because you couldn't do it by hand, but because doing it by hand across your whole market, refreshed every quarter, is a job nobody on your team has time for. Spend manual effort on the cohort-finding, which is cheap and high-leverage; spend money on the per-house resolution, which is the expensive, repetitive part. That split keeps your costs honest and your targeting sharp.

Putting it on a cadence: making this a system, not a sprint

The difference between a company that did this once and one that runs on it is cadence. A workable rhythm:

• Quarterly: Re-pull assessor and permit data for your top zips. Re-rank cohorts. Update saturation estimates. This catches the new re-roofs and keeps your skip lists honest.
• After every significant storm: Overlay the new storm track on your cohorts within a few days. The overlap of "old cohort" plus "fresh storm" is your highest-priority knocking the moment crews are free — and it's time-sensitive, because every other roofer is looking at the same weather radar.
• Continuously: Feed rep observations back in. A canvasser who finds a street is mostly re-roofed should be able to flag it so the office updates the map. Your reps are ground-truth sensors; use them.
• Annually: Reassess your year-built window. As time passes, the replacement zone slides forward — the 2008 cohort that was too new three years ago is entering the window now. Slide your filter with it.

The payoff of cadence is compounding. Year one you're working from a static map and it already beats the windshield. By year two, with permits and storms layered and rep feedback flowing, your target list is sharper than anything a competitor running on instinct can touch — and your reps are knocking pre-qualified doors, closing more, and sticking around. That last part is underrated: targeting precision isn't only a marketing win, it's a retention win, because nothing burns out a canvasser faster than a day of knocking roofs that were never going to buy.

A quick-start checklist

If you do nothing else from all of the above, do this in order:

1. List your serviceable zips. Only the ones a crew can reach without burning an hour driving.
2. Pull assessor data, filter to ~1990–2008 build years. Find which zips light up with old-cohort homes.
3. Map the points, color by build-year band, find the subdivision clusters. Name them.
4. Rank cohorts by opportunity — age, density, value fit, drive time — not age alone.
5. For your top 3 cohorts, pull re-roof permits and scan recent aerial. Estimate saturation; subtract the replaced roofs.
6. Overlay storm history. Flag old cohorts that also took recent hail or wind.
7. Resolve to the house where it counts — manually for one hot cohort, or with a per-roof tool across many — so you knock the originals and skip the re-roofs.
8. Hand crews the ranked doors and the skip list, plus a real talking point.
9. Stay legal on storm claims — document and estimate, never adjudicate.
10. Put it on a cadence — quarterly data refresh, post-storm overlays, rep feedback.

That's the whole system. The neighborhood-level question you started with — which neighborhoods have the oldest roofs — is the right place to begin, and the steps above answer it concretely with data anyone can pull. But the money is one level deeper, at the house. The roofers who win the next decade aren't the ones who know which neighborhoods are old. They're the ones who know which doors are due, who skip the re-roofs, who walk in with a real reason, and who turn their own streets into work they own instead of leads they rent or storms they wait on. Find the old cohorts first. Then go find the houses.

FAQ

How do I find out which neighborhoods have the oldest roofs in my city?

Start with county assessor data, which lists year built for nearly every parcel. Filter to an older build-year window (a common starting range in the mid-2020s is roughly 1990 to 2008), then map the results and look for spatial clusters of same-era homes. Those clusters are real subdivisions built in a single wave, which means their roofs are effectively the same age. That gives you the old neighborhoods. To know which specific houses still wear the original roof, you then layer in re-roof permit records and recent aerial imagery.

Can I just use Zillow or the county site to find old roofs?

Those sites show year built, not roof age. Year built equals roof age only on a house that has never been re-roofed. After the first replacement they diverge permanently, and a re-roof is invisible on a listing or a county year-built field. Year built is excellent for finding old build cohorts but cannot tell you which houses in that cohort already got new roofs. Use it to find the wave, then use permits and imagery to find which homes still have the original roof.

What's the difference between an old neighborhood and a neighborhood with old roofs?

An old neighborhood is a build cohort, and it works as a signal because subdivisions get roofed in waves. But that cohort almost always has holes shot through it: storm re-roofs follow a storm's path, big developments are built in phases years apart, and individual homeowners replace early for leaks or sales. So an old neighborhood typically has 20 to 60 percent of its roofs already replaced. Finding the neighborhood is step one. Finding which houses inside it still have the original, worn roof is the step that actually saves you money.

How can building permits help me find old roofs?

When a roof is replaced legally, someone usually pulls a re-roof permit, and many jurisdictions publish those records online, searchable by address and date. A re-roof permit is the cleanest public timestamp that a roof was replaced and roughly when. The powerful move is the inverse: a house in an old build cohort with no re-roof permit on record is a strong candidate for an original, aging roof, while a house in that cohort with a recent re-roof permit is a roof to skip. Permits are not complete, so treat no permit as a strong lean rather than a guarantee, and pair it with an aerial look.

Can you really tell a roof's age from aerial or satellite imagery?

You can read a range and a relative ranking, not an exact date. Aging asphalt roofs show color fade and mottling, dark algae streaking, rougher texture from granule loss, and visible patches. The most useful technique is the contrast test: compare a roof to its neighbors in the same build cohort. In a late-1990s subdivision, the roofs that read noticeably newer than the block are almost certainly recent re-roofs, and the uniformly weathered ones are likely original. Limits apply: a clean, well-ventilated roof can look younger than it is, and imagery is only as current as its capture date.

What build years should I target to find roofs that are due for replacement?

It depends on your climate and storm history, but a common starting window in the mid-2020s is homes built roughly 1998 to 2008, whose original asphalt shingles are deep in the typical replacement window. Remember that a shingle's warranty term is not a service-life prediction, so think in ranges, not exact ages. Reassess your window annually, because the replacement zone slides forward over time: cohorts that were too new a few years ago keep entering the window as their original roofs age.

Should I just knock the whole old subdivision?

No, and this is the most expensive habit in roof targeting. A 25-year-old cohort is commonly 20 to 60 percent already re-roofed, so knocking the whole neighborhood means a large share of your doors are five-to-fifteen-year-old roofs that will not buy for years. You pay full cost to reach those dead doors, and your reps burn out eating no's from people whose roofs are genuinely fine. Estimate saturation with permits and aerial first, build a skip list of the replaced roofs, and point your crew only at the likely originals.

How does storm history fit into finding old roofs?

Age and storm exposure compound. An old roof that also sat under recent hail or damaging wind crossed into replacement territory faster than an old roof in a calm pocket. Public severe-weather and hail data from sources like NOAA and the NWS Storm Prediction Center lets you overlay storm tracks on your old cohorts and prioritize the overlap. Keep the framing honest, though: storm exposure tells you a roof is worth inspecting, not that any specific roof is damaged. Damage is determined on the roof, and coverage is decided by the homeowner and their insurer.

How is per-roof targeting different from a hail map?

A hail map shows where it hailed across an area. It does not show which individual roofs the hail actually wore out, because the same storm hits a steep, south-facing slope very differently than a shallow, north-facing one next door. Per-roof modeling estimates storm impact on each specific roof using factors like orientation, slope, and exposure, and pairs it with a roof-age range read from imagery. The result is a ranked list of which doors are most likely due, rather than a blanket area that hailed. It points you to roofs to inspect; it does not prove damage.

Is a roof-targeting tool the same as buying leads?

No. Buying leads means purchasing contact info for strangers, often the same homeowner resold to several competitors at once. Roof targeting sharpens the outbound prospecting you already do by telling you which of your own doors to knock or mail and which to skip, based on roof age and storm exposure. It does not hand you customers or replace the knock, the inspection, or the sale. It makes your existing prospecting far more precise so your crew and your mail reach the roofs that are actually worn out instead of the whole street.