How to Find Roofs Due for Replacement in an Area (A Contractor's Playbook)

Every roofing company has a version of the same problem. You have a sales team, a finite number of door-knocking hours, and a service area with thousands of houses in it. Most of those houses have a roof that is fine. A small slice have a roof that is genuinely worn out, storm-beaten, or aging out of its service life. The whole game is figuring out which slice, and getting in front of those homeowners before three other companies do.

The lazy answer is "knock everything." That works, in the sense that if you knock 400 doors you will eventually find a roof that needs replacing. It is also the most expensive lead in roofing, because you are paying a salesperson's time and morale to ring 380 doorbells belonging to people whose roof has another decade in it. The good operators do something different. They treat the territory like data, narrow it down before anyone laces up their boots, and send crews to the addresses where the odds of a sale are stacked in their favor.

This is a field guide to doing exactly that: how to find roofs due for replacement in an area, using the signals that actually correlate with a replacement, the tools that surface those signals, and the workflows that turn a map into a route. It is written for the person who has to make the number, not for someone writing a textbook. Where there is a shortcut, I will tell you. Where a popular tactic is mostly noise, I will tell you that too.

What "due for replacement" actually means

Before you can find roofs that are due, you have to be honest about what "due" means, because the word gets used three different ways and they lead to three different sales motions.

Due by age. The roof has reached or passed the end of its expected service life for its material. A standard 3-tab asphalt shingle roof is commonly cited at roughly 15 to 20 years; architectural (dimensional) asphalt shingles in the 20 to 30 year range; and premium materials longer. The National Roofing Contractors Association and most material warranties frame these as ranges, not guarantees, because climate, ventilation, slope, and install quality move the number a lot. A south-facing, poorly ventilated roof in Phoenix ages faster than a shaded north slope in Portland. So "old" is a probability statement, not a verdict.

Due by damage. A storm, hail event, wind event, or chronic leak has compromised the roof regardless of its age. A four-year-old roof can be due for replacement the morning after a baseball-sized hail event. This is the storm-restoration motion, and it runs on a completely different clock than the age motion.

Due by transaction. The homeowner is about to sell, just bought, is refinancing, or got a non-renewal notice from their insurer that flagged the roof. These roofs may not be the oldest or the most damaged, but the homeowner has a forcing function, which makes them some of the easiest closes you will ever get.

A real targeting system blends all three. The roofs that are old AND in a storm swath AND owned by someone who just got an insurance letter are the ones you want at the top of the route. Everything in this guide is about scoring addresses against those three axes and stacking them.

The honest math of a roofing territory

Let me put numbers to the problem so the rest of the article has stakes. Take a suburban service area of 20,000 single-family homes. Assume an average roof service life of 22 years and a roughly even age distribution (it is never perfectly even, but bear with me). That implies somewhere around 900 to 1,000 roofs cross into "replacement-likely by age" territory every year, plus whatever a storm season adds on top. That is your real addressable market this year, maybe 5% of the doors.

If your team knocks blind, you are hunting a 5% population by checking doors one at a time. If you can pre-score the territory and concentrate knocking on the streets where that 5% is actually a 20% or 30%, you have not changed how hard your reps work. You have changed what they walk into. That delta, the difference between a 5% door and a 25% door, is the entire return on everything below.

The signals that actually predict a replacement

There is a long list of things that feel like they should predict a worn-out roof and a shorter list that actually do. Here is the honest ranking from years of watching what correlates with a signed contract versus what just looks good on a slide.

Tier 1 signals (strong, worth building a route around)

Roof age as a range. The single best predictor, full stop. Not because age guarantees failure, but because age is the base rate everything else multiplies. A 23-year-old asphalt roof in a hail-prone county is a fundamentally different prospect than a 6-year-old one, before you know anything else. The trick is that you almost never get a clean "this roof was installed in 2003" date. What you can get, and what you should aim for, is a defensible range: roughly 18 to 24 years old, derived from imagery, permit records, and material type. Treat age as a range, never a single date, and your conversations with homeowners stay honest.

Recent severe-weather exposure on that specific roof. Note the wording: on that specific roof, not "a storm passed through the zip code." Hail and wind damage is intensely local. A hail core can drop golfball-sized stones on one street and pea-sized on the next street over. Wind funnels and accelerates around terrain and buildings. The contractors who win storm work are the ones who can say "this address took the worst of it" instead of "there was a storm in your area," which every homeowner has already heard from five door-knockers.

Visible condition from imagery or the curb. Granule loss showing as dark patches, curling or cupping shingles, missing tabs, patched repairs, streaking, sagging deck lines, moss in wet climates. Modern aerial and street-level imagery lets you eyeball a surprising amount of this before you knock. It is not a substitute for getting on the roof, but it is an excellent filter.

Tier 2 signals (useful as multipliers, weak on their own)

Neighborhood install cohort. Tract subdivisions get roofed in waves. If a development was built in 2001 and 60% of the original roofs are still original, that whole neighborhood is aging in lockstep and is worth a concentrated push. Conversely, if you can see half the street already has bright new shingles, the cohort has already turned over and the urgency is lower.

Property transaction activity. Recent sale, listing, or a building-permit pull next door (people roof when their neighbor roofs). New owners and sellers have forcing functions.

Insurance signals you can legitimately observe. Some homeowners will tell you, unprompted, that their carrier sent a notice about the roof. You cannot and should not pretend to know someone's policy status. But when a homeowner volunteers it, that is a Tier 1 close sitting inside a Tier 2 signal.

Tier 3 signals (mostly noise, treat with suspicion)

Home value or income data alone. A wealthy neighborhood is not a roof-replacement neighborhood. Wealthy people often have newer or premium roofs. Income tells you about ability to pay, which matters for the close, but it tells you almost nothing about whether the roof is due.

"Storm hit the county" alerts with no per-roof resolution. A county-level or zip-level storm alert means a storm touched somewhere in a 30-mile box. By the time you knock on it, so has everyone else, and most of the doors did not take real damage. Useful as a starting flag, dangerous as a targeting strategy.

Generic "homeowner" lists from data brokers. Lists sorted by "homeowner, age 35+, owns home 10+ years" will technically skew older-roof, but the correlation is so loose you are back to nearly-blind knocking with extra steps and a data-broker invoice.

The operating principle: build routes on Tier 1, use Tier 2 to break ties and stack, and never let Tier 3 drive the bus.

Method 1: Read roofs from aerial and satellite imagery

This is the highest-leverage skill in the whole playbook because it is free or cheap, scales to thousands of roofs, and you can do it from a desk at 9 PM. The goal is not a perfect diagnosis. It is triage: sorting a neighborhood into "worth a knock," "maybe," and "skip."

What you can actually see from above

With good-resolution aerial imagery (the kind in most mapping tools and dedicated roofing measurement platforms), a trained eye can read:

• Overall color uniformity. A roof that has aged unevenly, with blotchy lighter and darker areas, is often shedding granules unevenly. Uniform deep color usually reads younger.
• Streaking and staining. Dark vertical streaks (often algae, Gloeocapsa magma) signal an older roof in humid climates. Not damage by itself, but a strong age tell.
• Patches and mismatched sections. A rectangle of different-colored shingles screams "prior repair," which means prior problems.
• Missing or displaced shingles. Visible gaps, especially in a consistent direction, point to wind events.
• Sagging or wavy ridge and deck lines. A roof plane that is no longer straight suggests deck or structural issues, which push toward full replacement.
• Roof complexity and material. You can identify asphalt vs. tile vs. metal vs. wood shake, count facets and penetrations, and estimate squares, all of which feed your bid before you ever knock.

What you cannot see, and why honesty matters

Imagery does not show you granule depth, mat brittleness, soft decking, or the bruising from hail that only reveals itself on a hands-on inspection. You can flag a roof as a strong candidate from above. You cannot declare it damaged or due. Pretending otherwise is how reps end up making claims they can't back up on the roof, which is both a sales liability and, in storm work, a compliance problem. Imagery qualifies the door. The inspection makes the call.

A practical imagery triage workflow

Here is a repeatable desktop process a sales manager can hand to a junior rep:

1. Pick a target neighborhood (we will cover how to choose it in Method 4). Pull up high-resolution aerial imagery for the area.
2. Cross-reference build year. County assessor and parcel data (usually free online) gives you the year the home was built. For original roofs, build year is your starting age. Pair it with imagery condition to refine the range.
3. Score each roof 1 to 3 on a simple rubric: 3 = old build year plus visible condition signs (streaking, patching, color blotch); 2 = one signal present; 1 = clearly newer or recently re-roofed. Drop the scores into a spreadsheet or a map layer keyed to the address.
4. Check historical imagery if available. Some platforms let you scrub back through years of captures. If a roof looks the same in a 2009 image as today, and the home was built in 2002, you now have a roof that is very likely 22-plus years old. Historical imagery is one of the most underused age tools available.
5. Build the knock list from the 3s first, then 2s on the same streets to keep routes tight.

One caution on imagery dates: the "satellite" view in consumer mapping tools can be one to four years stale, and capture dates vary by area. For age estimation that staleness is fine. For post-storm damage assessment it is useless, because the imagery predates the storm. Know which question you are asking.

Method 2: Use age and permit records to anchor the range

Imagery tells you what a roof looks like now. Records tell you how old it probably is. Together they turn a guess into a range you can defend on a doorstep.

County assessor and parcel data

Nearly every U.S. county publishes parcel data with the year a structure was built, and much of it is searchable online for free or available as a bulk download. Year built is your anchor for any roof that is still original. The catch: you do not know if the roof is original. Plenty of 1995 homes have a 2015 roof on them. That is exactly why you pair build year with imagery and permits, not use it alone.

Building permit records

This is the most precise public signal of roof age that exists, and most reps never touch it. When a homeowner pulls a permit to re-roof (required in most jurisdictions, though enforcement varies), it creates a dated public record. Many cities and counties publish permit data online; some offer bulk exports or APIs.

Use permits two ways:

• To exclude. A re-roof permit from 2019 means that address is off your list for a decade. Removing recently-roofed homes from a knock list is as valuable as adding old ones, because it stops you wasting time and annoying people who just spent money on a roof.
• To find the cohort. If you see a cluster of re-roof permits in a neighborhood after a hail event five years ago, the homes without a permit in that cluster are the holdouts, still on storm-damaged or aging roofs, and they are prime targets.

Permit data is messy. Coverage, formats, and field quality vary wildly between jurisdictions, and not every re-roof gets a permit pulled. Treat it as a strong refining signal, not a complete census.

Material type sets the clock

Once you know or can estimate the material, you know which service-life clock to run. Rough, widely-cited service-life ranges (always ranges, always climate-dependent):

The tile note matters in tile-heavy markets: the tiles can be fine while the underlayment beneath them has failed, which is a real replacement opportunity that pure age-of-tile thinking misses. Knowing your local material mix tells you which clock to even bother running.

Method 3: Layer in storm and severe-weather history

For a huge share of roofing companies, the biggest replacement events are not slow aging, they are storms. Hail and high wind can take a roof from "another decade left" to "due now" in fifteen minutes. So any serious targeting system has to fold in weather, but the resolution at which you do it is everything.

The resolution trap

Most free storm data is coarse. NOAA's Storm Prediction Center and the National Weather Service publish hail and wind reports, and there are public storm-report databases, but the native resolution is often a point report or a county-level event. "1.75-inch hail reported near [town]" tells you a storm happened in the vicinity. It does not tell you that the roof at 412 Oak Street took a direct, damaging hit while 418 Oak Street, a hundred feet away, barely got grazed.

This matters because hail and wind damage is street-by-street and even roof-by-roof. Hail cores are narrow and erratic. Wind interacts with terrain, tree lines, and neighboring structures. The whole problem with zip-level or county-level storm alerts is that they send your entire team to a 30-mile box where maybe a few streets actually got hammered, and you find out which streets only by knocking all of them, alongside every other contractor who bought the same alert.

What good storm targeting looks like

The better approach is to model the storm against each individual roof: how much energy hit that specific roof plane, given the storm's path, the hail size and density along it, the wind field, and the roof's exposure and slope. Instead of "a storm passed through your area," you get "the roofs in this two-block pocket took the highest hail energy in the event, and these specific addresses were in the core." That is a route, not a region.

You can approximate some of this manually by overlaying detailed radar-derived hail-swath data on a parcel map, but the radar-to-roof translation is where it gets hard and where dedicated tools earn their keep. The key mental shift: stop thinking in storm areas and start thinking in storm-energy per roof. A forecast or a storm model gives you odds that a given roof was damaged. It is never proof of damage. Only an inspection establishes damage. Keep that line bright, in your own head and in every word your reps say at the door.

Timing the knock after a storm

Storm work is a footrace, and the timing details decide who wins:

• The first 24 to 72 hours are dominated by the loudest, fastest, and often the least scrupulous canvassers. If you have per-roof targeting, you can skip the chaos of blanketing and go straight to the streets that actually took the worst, which is both more efficient and more credible.
• Weeks 2 to 8 are where homeowners who got blanketed early but didn't act are now ready to talk, and where the storm-chasers have already moved on. Local, credible contractors with good targeting clean up here.
• Watch for the insurer's clock. Many policies require claims within a set window after a loss. You do not advise on claims, handle them, or promise outcomes, that is the homeowner's claim and the insurer's decision, but knowing the window helps you understand the homeowner's urgency.

A note on compliance, because it is easy to get wrong in storm season: document conditions and provide an honest estimate. Do not promise a "free roof," do not promise to cover or waive deductibles, and do not present a storm model or forecast as proof that a specific roof is damaged. The homeowner owns the claim, the insurer decides coverage, and you are simply the contractor who documents and repairs. Staying inside those lines is legally safer, and it also builds the kind of reputation that wins the weeks-2-to-8 work.

Method 4: Pick the right neighborhoods before you pick the right doors

Everything above operates at the address level. But your reps move in routes, and routes live in neighborhoods. So there is a layer above address-scoring: choosing which neighborhoods deserve a concentrated push this week.

The cohort heuristic

The fastest way to find a dense pocket of due-by-age roofs is to find a subdivision built in a tight window 18 to 28 years ago that has not substantially turned over. Tract developments roof in cohorts, so when the cohort ages out, you get whole streets of simultaneously-due roofs. The signal you are hunting:

• Build years clustered in, say, 1998 to 2003 (from parcel data).
• Imagery showing mostly original-looking roofs with age signs (streaking, color blotch), not a sea of bright new shingles.
• Few recent re-roof permits in the tract.

Find one of those and you have a target-rich environment where knocking density pays off, because the houses next to your appointment are also due.

The storm-overlay heuristic

The other neighborhood-picker is recent severe weather. Take your detailed storm history, find the pockets that took the highest hail energy or wind, and prioritize those, especially where they overlap with older cohorts. An old neighborhood that just took hard hail is the densest possible concentration of due roofs you will find.

A worked example

Suppose you run a company in a hail-exposed metro. You are planning next week. Here is the stack:

1. Storm layer: A hail event nine days ago. Detailed swath data shows the damaging core ran through three older subdivisions on the north side.
2. Cohort layer: Two of those three subdivisions were built 2000 to 2004 (parcel data) and show mostly original roofs (imagery).
3. Permit layer: Pull re-roof permits for those two subdivisions. Remove the ~15% of homes already re-roofed in the last decade.
4. Imagery triage: Score the remaining homes 1 to 3 on visible condition. Flag the 3s and 2s.
5. Route build: You now have, say, 280 high-probability addresses on tight streets. Your reps walk those, leading with an honest inspection offer, not 1,200 random doors.

That is the whole method in one paragraph: storm core, intersected with old cohort, minus recent re-roofs, refined by imagery, sequenced into a route. Each layer multiplies the hit rate of the next.

Method 5: Door-level canvassing data and CRM discipline

Finding the roofs is half the job. Not re-finding the same doors, tracking who said "come back in spring," and measuring which signals actually convert, that is the half that compounds over years.

Treat the territory as a living map

The companies that win long-term run a canvassing CRM (there are several in the roofing space) where every door has a status: not-home, not-interested, inspection-scheduled, sold, come-back date, recently-roofed-by-someone-else. The map gets smarter every time a rep touches it. Two years in, you are not re-knocking dead doors, you are working a curated list of warm and aging leads.

Key disciplines:

• Log the "come back" dates religiously. A homeowner who said "my roof has maybe two more years" in 2024 is a 2026 appointment. That is a lead you already paid for.
• Tag the reason for every no. "Just re-roofed" is permanent. "Spouse not home" is a callback. "Not interested" might be a price objection you can re-approach after a storm. The reason code is the asset.
• Close the loop on conversions. Tag which signal got each sold roof onto the route (age, storm, transaction). After a few hundred doors you will know which signals convert in your market and can reweight accordingly.

Don't confuse activity with targeting

A common failure mode: a company buys a canvassing app, tracks thousands of knocks, generates beautiful heatmaps of where they knocked, and never improves which doors they knock. Activity tracking is necessary but it is not targeting. The CRM tells you what you did. The signals (age, storm, condition) tell you what you should do. You need both, and you should not let the dashboard convince you that knocking more is the same as knocking smarter.

Method 6: Turn a flagged roof into an inspection, then a contract

Targeting gets the right rep to the right door. What happens in the next ninety seconds decides whether all that data turns into revenue. This is where a lot of well-targeted programs leak, because the people who built a sharp knock list never built an equally sharp door script and inspection process.

The doorstep opener for an age-flagged roof

When the flag is age, not storm, the opener should sound like a neighbor with expertise, not a pitch. Something close to: "Hi, I'm with [company], we're a local roofing company and we've been doing inspections on this street, a lot of these homes were built around the same time and the original roofs are reaching the age where they start to go. I'd be happy to do a free, no-obligation look at yours and just tell you honestly where it stands, you might have years left, you might not."

Three things make that work. It is honest (you genuinely don't know their roof's condition yet). It uses the cohort signal as social proof ("a lot of these homes"). And it offers information, not a sale, which lowers the homeowner's defenses. Reps who lead with "your roof needs replacing" before climbing it have nothing to back the claim and homeowners know it.

The doorstep opener for a storm-flagged roof

When the flag is a storm, the line is different but the discipline is identical: offer to look, don't diagnose from the sidewalk. "We had that hail come through about ten days ago and we're inspecting roofs on the streets that took the worst of it. I can get up there and document what I find and give you an honest estimate, no pressure either way." Notice what is absent: no promise the roof is damaged, no mention of a "free roof," no deductible talk, no claim that a storm model proves anything about their specific roof. You are offering documentation and an estimate. That is the compliant, credible posture, and it is also what separates you from the chaser who told them their roof was "totaled" from the curb.

What a real inspection documents

The inspection is where probability becomes evidence. A thorough roofing inspection that supports an honest replacement recommendation typically covers:

• Field of the roof: granule loss, mat exposure, shingle brittleness, curling, cupping, and for hail, the bruising and fractured-mat hits that only show up hands-on.
• Penetrations and flashing: chimneys, vents, pipe boots (a top leak source as the rubber dries and cracks), step and counter flashing, valleys.
• Edges and ridges: drip edge, rake, ridge caps, and any lifted or missing pieces from wind.
• Decking and structure: soft spots underfoot, sagging planes, sheathing condition where visible.
• Attic side when accessible: daylight, staining, active or past leaks, ventilation adequacy, and insulation moisture.
• Documentation: date-stamped photos of every finding, measurements, and a written summary the homeowner keeps.

That documentation does double duty. It substantiates your honest recommendation to the homeowner, and if the homeowner chooses to file a claim, it is the factual record they and their insurer rely on. You are not filing the claim, advising on coverage, or promising an outcome, you are the contractor who documented conditions and provided an estimate. That distinction is the entire compliant model, and it is worth drilling into every rep until it is reflexive.

Edge cases that trip up inspections

A few situations where a flagged roof needs careful judgment rather than a reflex recommendation:

• Tile roofs with good tile but failed underlayment. The tiles can look fine from the air and the curb while the underlayment beneath has reached end of life. This is a legitimate replacement (lift-and-relay or full) that age-of-tile thinking misses, and it is common in tile-heavy Southwestern and Florida markets.
• Recent partial repairs. A roof with a fresh patch may have a localized problem, not a whole-roof failure. Don't over-recommend, inspect the rest before deciding.
• Layovers (two or more shingle layers). If a roof already carries multiple layers, codes in many jurisdictions require a full tear-off rather than another layover, which changes the scope and price you quote. Spot it during inspection, not after the contract.
• Mixed-age roofs. Additions and prior partial replacements mean one address can have two different roof ages on it. Inspect each plane.

Building your data stack without overspending

You do not need an enterprise data budget to run everything above. Here is how the layers map to what they cost, so you can assemble a stack that fits your size.

Free or near-free: County assessor and parcel data (build years, ownership), most municipal building-permit portals, consumer mapping tools for basic aerial imagery, and public storm-report databases from NOAA and the National Weather Service. A disciplined sales manager can run Methods 1 through 4 on this stack alone for a handful of neighborhoods a week. The cost is time, not money.

Modest subscriptions: Aerial roof-measurement platforms (for squares, facets, and pitch on bids), a canvassing CRM for the door-level discipline in Method 5, and higher-resolution or historical imagery providers. These pay for themselves quickly because they cut measurement time and stop you re-knocking dead doors.

Targeting data layers: Tools that produce the two hard signals at territory scale, a per-address age range and per-roof storm modeling. This is the tier that replaces dozens of manual desk-hours and lets you run the whole playbook across thousands of homes instead of a few neighborhoods. The decision to buy here should follow the same logic as any tool: it multiplies a process you already understand. Run the manual version first so you can judge the output.

The sequencing advice is the same as elsewhere in this guide: start free, prove the process on one neighborhood, learn what the signals look like in your specific market and material mix, and only then spend up the stack. A company that has hand-scored two hundred doors is a vastly smarter buyer of automation than one that hasn't.

Where RoofPredict fits

Everything above can be assembled by hand. Plenty of good companies do exactly that, with a sales manager who lives in parcel data, mapping tools, and permit portals. The problem is that the manual version is slow, it does not scale past a few neighborhoods a week, and the two hardest signals, a defensible roof-age range per address and per-roof storm energy, are precisely the ones that are most painful to produce by hand.

That is the specific gap RoofPredict is built for. It is a data layer for roofing contractors that scores roofs house-by-house across your service area on the two signals that matter most:

• A roof-age range per address, derived from aerial imagery (and the cohort/condition cues described above), expressed as a range rather than a false-precision date. This is the same age triage you would do manually with imagery and parcel data, run at the scale of an entire territory instead of one neighborhood at a time.
• Storm modeled per roof, rather than only where it passed. Instead of a county or zip alert, RoofPredict models the storm against each individual roof, so you get the addresses that took the highest energy in an event rather than a 30-mile box. It ranks the doors and routes so crews knock the roofs the storm actually wore out, plus the roofs aging out, first.

The honest framing matters here, so let me be precise about what it is and is not. RoofPredict ranks probability: which roofs are most likely due by age and most likely to have taken storm energy. The age output is a range, not an install date. The storm output is odds, not proof of damage, you still send a rep, and the inspection makes the call. It is not a lead-buying service that sells you the same contacts as ten other companies; it is a targeting layer that tells your own team which of your territory's doors to work first. And it stays inside the compliance lines that keep you out of trouble: you document conditions and provide estimates, the insurer decides coverage, the homeowner owns the claim.

In practice it slots into the workflows in this guide as the engine under Methods 1 through 4: it does the imagery-based age triage and the per-roof storm modeling at territory scale, and hands your CRM and your reps a ranked list. You still run the canvassing discipline of Method 5. You still get on the roof. What changes is that the 5%-likely territory has been re-sorted into routes where the odds are stacked toward roofs that are genuinely due.

If the manual playbook above is something you already do well and you just want it to run across 20,000 homes instead of 200, that is the natural fit. If you are not doing it at all yet, start manual on one neighborhood this week using Methods 1 to 4, learn what the signals look like in your market, and you will be a far sharper judge of whether a data layer is worth it.

Putting it all together: a 7-day operating cadence

Here is how the pieces fit into a week, for a company that wants a repeatable rhythm instead of ad-hoc hustle.

Monday, target selection (sales manager, ~2 hours). Review storm activity from the prior week. Pick this week's two or three target neighborhoods using the cohort heuristic and any storm overlay. Pull parcel build-years and recent re-roof permits for those areas.

Tuesday, imagery triage (manager or junior rep, ~3 hours). Run the imagery scoring on the target neighborhoods. Remove recently re-roofed homes. Produce a ranked knock list of 3s and 2s, sequenced into tight routes. (If you use a data layer like RoofPredict, this step is mostly automated and you are reviewing a ranked list instead of building it.)

Wednesday through Friday, the knock. Reps work the routes, leading with an honest, free inspection offer. Every door gets logged in the CRM with a status and reason code. Inspections scheduled get booked. Reps document conditions with photos and provide estimates, no claims promises, no deductible talk, no "free roof" pitch.

Friday afternoon, the review (~1 hour). Look at the week's numbers: doors knocked, inspections set, contracts signed, and crucially, which signal each conversion came from. Tag come-back dates. Note any neighborhood that overperformed for next time.

Ongoing, the compounding. Every week the CRM map gets richer, the come-back pipeline fills, and your read on which signals convert in your specific market sharpens. The system gets better whether or not you sign a deal that week.

A simple scoring rubric you can copy

If you want a starting point for an address score, here is a transparent one you can build in a spreadsheet. Higher score, higher priority:

Knock the highest scores first. Recalibrate the weights after a few hundred doors based on what actually converted. This is a deliberately crude model, and a crude model you actually use beats a sophisticated one you don't.

Common mistakes that quietly kill targeting programs

A few failure modes show up again and again. Avoiding them is most of the battle.

Chasing zip-level storm alerts. As covered, county and zip alerts send everyone to the same 30-mile box. If your only storm signal is "a storm hit the area," you are competing on speed and volume against the loudest chasers in the business. Get to per-roof resolution or treat storm alerts as a weak flag only.

Treating age as a date instead of a range. Reps who say "your roof is 22 years old" when they really mean "probably 18 to 24" lose credibility the moment the homeowner says "actually we re-roofed it in 2014." Speak in ranges. It is both more honest and, oddly, more persuasive, because it sounds like expertise instead of a script.

Letting income data drive the route. Wealthy neighborhood does not equal due-roof neighborhood. Use ability-to-pay data for the close, never for the targeting.

Knocking newly-roofed homes. Nothing burns goodwill and rep morale like pitching a roof someone replaced last year. Permit data and imagery both flag these. Remove them. It is the cheapest improvement to a knock list available.

Confusing CRM activity with targeting. Heatmaps of where you knocked are not a list of where you should knock. Track activity, but let signals choose doors.

Making damage claims from imagery or weather. Imagery qualifies a door. A storm model gives odds. Neither proves a specific roof is damaged, only an inspection does. Reps who overclaim create both sales blowback and compliance exposure. Document, estimate, inspect, and let the homeowner and their insurer do their parts.

Over-investing in tools before learning the signals. Software multiplies a process. If you do not yet have a process, run the manual playbook on one neighborhood first. You will be a much smarter buyer afterward, and you will use whatever you buy far better.

Frequently asked, briefly

A few quick hits that come up constantly, expanded in the FAQ below: yes, you can estimate roof age from aerial imagery, but as a range, not a date. No, you should not lead with insurance or claims language at the door. Yes, permit data is worth the hassle, mostly for removing recently-roofed homes. And no, there is no single database that perfectly lists "roofs that need replacing," because "need replacing" is partly a judgment that only an inspection settles, what good tools give you is the ranked probability that gets the right rep to the right door.

The bottom line

Finding roofs due for replacement in an area is not about a magic list. It is about stacking signals that each shift the odds, roof age expressed as a range, severe-weather exposure measured per roof instead of per region, visible condition from imagery, and transaction and cohort cues, then sequencing the highest-probability addresses into tight routes and working them with disciplined CRM follow-up. Done by hand, this turns a 5%-likely territory into 20%-or-better routes without your reps working any harder. Done at scale with a data layer that produces the two hardest signals, a per-address age range and per-roof storm modeling, it turns your whole service area into a ranked map of who is due first.

Start this week. Pick one aging subdivision, pull build years and permits, triage the imagery, remove the new roofs, and send your best rep down the highest-scoring street with an honest inspection offer. That single disciplined route will teach you more about targeting than any amount of theory, and it is the foundation everything else builds on.

FAQ

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

You can estimate it as a range, not a precise install date. Imagery reveals age-correlated signs like granule-loss color blotching, algae streaking, prior patches, and missing shingles, and when you pair that with the home's build year from parcel data and any re-roof permits, you can land on a defensible range such as 'roughly 18 to 24 years.' Historical imagery, if your tool offers it, sharpens this further. What imagery cannot do is confirm hidden damage or give an exact date, so always speak in ranges and let an on-roof inspection make the final call.

What's the fastest way to find a dense pocket of replacement-due roofs?

Look for a tract subdivision built 18 to 28 years ago that hasn't turned over. Tract homes get roofed in cohorts, so when one ages out, whole streets are simultaneously due. Confirm with parcel build-years clustered in a tight window, imagery showing mostly original-looking roofs rather than bright new ones, and few recent re-roof permits. If that cohort also sits inside a recent storm core, it's the densest concentration of due roofs you'll find.

Why are zip-code or county storm alerts a weak targeting signal?

Because hail and wind damage is street-by-street, even roof-by-roof, while county and zip alerts cover a 30-mile box. An alert tells you a storm touched somewhere in that box, not which specific roofs took damaging energy. Acting on it sends your whole team to canvass everything, alongside every other contractor who bought the same alert. You only learn which streets actually got hit by knocking all of them. Per-roof storm modeling, which estimates how much energy hit each individual roof, turns that region into an actual route.

Are building permit records worth the trouble for roof targeting?

Yes, primarily for removing homes rather than finding them. A re-roof permit from the last decade means that address is off your list for years, and pulling those addresses off a knock list saves rep time and avoids the goodwill damage of pitching someone who just bought a roof. Permits also help you spot holdouts: in a neighborhood where many homes re-roofed after a past storm, the ones without permits are still on aging or damaged roofs. The data is messy and coverage varies by jurisdiction, so treat it as a strong refining layer, not a complete census.

How long does a typical asphalt shingle roof last?

It depends on the shingle type and the climate, and it's always a range. Standard 3-tab asphalt shingles are commonly cited around 15 to 20 years, and architectural or dimensional asphalt shingles around 20 to 30 years. Climate, attic ventilation, slope, sun exposure, and install quality move those numbers significantly, a hot, poorly ventilated, south-facing roof ages faster than a shaded one. Use the material's service-life range as a starting clock, then refine with imagery condition and permit history for a given address.

Should my reps mention insurance or storm damage when they knock?

Lead with an honest, free inspection offer, not claims talk. You can note that a recent storm passed through and you're offering inspections, but you should not tell a homeowner their specific roof is damaged before inspecting it, promise a 'free roof,' offer to cover or waive deductibles, or present a storm model or forecast as proof of damage. The compliant lane is simple: you document conditions and provide an estimate, the insurer decides coverage, and the homeowner owns the claim. Staying in that lane is both legally safer and better for your reputation.

Is there a single database that lists every roof that needs replacing?

No, and be skeptical of anyone who claims there is. 'Needs replacing' is partly a judgment that only an on-roof inspection can settle, so no database can definitively list it. What good tools and good manual workflows give you is a ranked probability, which roofs are most likely due by age and most likely to have taken storm energy, so you send the right rep to the right door. The inspection still makes the final determination.

Does targeting wealthier neighborhoods help find replacement-due roofs?

Not for finding due roofs. Home value and income data tell you about ability to pay, which matters at the close, but they say almost nothing about whether a roof is worn out. Wealthier homes often have newer or premium materials. Drive your route off age, storm exposure, and visible condition, then use income and value data only to inform the sales conversation, never to choose which doors to knock.

How does RoofPredict help find roofs due for replacement?

RoofPredict is a data layer for roofing contractors that scores roofs house-by-house across your service area on the two hardest-to-produce signals: a roof-age range per address derived from aerial imagery, and storm energy modeled per individual roof rather than per zip or county. It ranks the doors and routes so your crews knock the roofs a storm actually wore out and the roofs aging out, first. It outputs probability, not certainty, age is a range, storm exposure is odds, not proof of damage, and you still send a rep to inspect. It's a targeting layer for your own territory, not a shared lead-buying list.

What's the single highest-leverage skill for finding due roofs cheaply?

Reading roofs from aerial imagery for triage. It's free or cheap, it scales to thousands of roofs from a desk, and it sorts a neighborhood into 'worth a knock,' 'maybe,' and 'skip' before anyone leaves the office. Pair imagery condition cues with parcel build-years and re-roof permits, score each address, and you've converted a blind territory into a ranked route. It doesn't replace getting on the roof, but it dramatically narrows where you spend boots-on-the-ground time.