How to Find Roofs That Need Replacing: A Contractor's Guide to the Software That Actually Works

Every roofing company owner I talk to is chasing the same thing in two different costumes. Costume one: "I need more leads." Costume two: "I need better leads." Strip both away and the actual job is this — figure out which specific houses, on which specific streets, have a roof that is genuinely at the end of its life or got worked over by a storm, and get a crew in front of those doors before three other companies do.

That is a targeting problem, not a marketing problem. And targeting is exactly where software earns or loses its keep.

The trouble is that most "roofing software" being pitched to contractors doesn't actually find roofs that need replacing. It finds addresses inside a polygon a storm crossed, or it finds homeowners who filled out a form, or it finds property records you could have pulled from the county yourself. Those are useful pieces. They are not the same thing as knowing which roof is due.

This guide walks through the real problem, the categories of tools that claim to solve it, how the underlying data is actually produced (so you can tell signal from sales deck), a workflow you can run starting tomorrow, the mistakes that quietly cost crews their week, and a frank section on where address-level roof-age and storm-modeled targeting data fits — including its honest limits. No pricing, no fake "3x your close rate" numbers, no magic. Just how this works when you do it for a living.

What "a roof that needs replacing" actually means

Before you can shop for software, get precise about what you're hunting. A roof becomes a real replacement opportunity for one of three reasons, and they don't look the same on a map.

1. Age and wear (the slow burn). An asphalt shingle roof in normal conditions runs a usable life that depends heavily on the product and the climate. A builder-grade 3-tab from the early 2000s is a different animal than a laminated architectural shingle installed in 2015. Manufacturers publish long warranty numbers, but field life is shorter — granule loss, mat embrittlement, thermal cycling, and UV degradation do their work on a curve, not a cliff. The National Roofing Contractors Association is blunt about this in its consumer guidance: warranty length is not the same as service life, and condition has to be assessed on the actual roof. The practical takeaway for targeting: a roof installed 18–24 years ago in a hot, sunny climate is a far better knock than the same product installed 8 years ago two streets over.

2. Storm wear (the fast burn). Hail bruises the mat and knocks granules off; wind lifts, creases, and tears tabs; debris punctures. The damage that matters for a replacement is often not the dramatic stuff a homeowner notices. It's functional damage — bruising you feel with your thumb, mat fractures, creased tabs at the lift line — that shortens the remaining life and, depending on the insurer and policy, may be claimable. The Insurance Institute for Business & Home Safety (IBHS) has done extensive impact research showing hail damage severity depends on stone size, density, hardness, and impact angle, plus the age and brittleness of the shingle itself. An old roof and a fresh roof in the same hailstorm do not come out the same.

3. Failure and defect (the sudden burn). Active leaks, ice-dam damage, ponding on low-slope sections, failed flashing, deck rot. These are real but harder to find from the curb or the sky, and they tend to come to you (the homeowner calls) rather than you finding them.

Good targeting software helps most with #1 and #2, because those are the ones where data at the property level can point you at a door before the homeowner has picked up the phone. Keep that framing as you evaluate anything: Does this tool help me find age-and-wear roofs, storm-wear roofs, or both — and how does it actually know?

Why the wear curve matters more than the warranty number

Contractors lose money targeting on warranty math. A homeowner with a "30-year" shingle assumes 30 years; a roofer who knows the field understands the usable curve bends down well before the warranty expires, and that the bend is steeper in some conditions than others. Three variables move that curve hard:

• Climate and sun exposure. UV and heat are the dominant agers of asphalt. The same shingle on a roof in a high-UV, high-heat region degrades faster than one in a mild, cloudy climate. Within a single market, south- and west-facing slopes age faster than north-facing ones because they bake longer each day. That's why you'll sometimes condemn one slope of a roof and find the opposite slope has years left.
• Ventilation and deck temperature. Poorly vented attics cook the underside of the deck and the shingles, accelerating embrittlement. Two identical roofs installed the same week can be years apart in real condition because one attic runs 20–30 degrees hotter.
• Product tier. Builder-grade 3-tab, mid-tier architectural, and premium designer or impact-rated shingles age on different curves. The cheap stuff on a 2003 spec home is a far better age target than a premium laminate on a custom home of the same vintage.

The operational point: a tool that only knows "year built" is blind to all three of these. A tool that reads the actual roof surface — fade, granule loss, streaking, slope-by-slope differences — is reading the wear curve directly, which is what you actually care about.

Where the money is: re-roofs, not new construction

One more framing before tools. The replacement market a targeting tool serves is overwhelmingly the re-roof market — existing homes getting a second, third, or fourth roof — not new construction. That matters because the entire housing stock is constantly cycling through replacement, and the U.S. Census Bureau's American Housing Survey shows a large, aging owner-occupied single-family base. A huge share of those homes are now into the window where their original or first-replacement roof is aging out. Targeting software exists to find the specific houses inside that slow, continuous churn that have reached the front of the line — and to flag the ones a storm just shoved to the front early.

The five categories of software roofers are sold (and what each is really for)

When a vendor says "we help you find roofs that need replacing," they almost always belong to one of five buckets. Knowing the bucket tells you what the tool can and can't do.

1. Aerial measurement / report tools

These take an address and return a roof measurement report — squares, pitch, facets, ridge and hip and valley lengths, waste factor. Think of the established satellite/aerial measurement providers. What they're for: producing an accurate takeoff without climbing the roof, so you can estimate and order material faster and safer. What they are not: a prospecting tool. A measurement report tells you the size and shape of a roof. It tells you nothing about whether that roof is worn out. You point them at a house you already chose. They don't choose houses for you.

2. Property / parcel data platforms

These aggregate county assessor and deed records: year built, last sale, owner-occupancy, lot size, sometimes permit history. Useful for filtering — owner-occupied, single-family, built before year X — and for skip-tracing a mailing list. The limit is that "year built" is the year the house was built, not the year the roof was last replaced. A 1992 house can easily have a 2019 roof. Parcel data gives you a coarse prior, not the answer. It's a sieve, not a sensor.

3. Storm / hail map and notification tools

This is the big, crowded category. These tools ingest weather data — radar-derived hail estimates, wind reports, storm tracks — and draw swaths on a map so you know where a storm hit. Several are well known in the restoration world. They answer "where did it hail last night, and how big?" reasonably well, and many will hand you a list of every address inside the swath. The thing to understand is what that swath actually represents, which we'll dig into below. The short version: a hail swath is a probability surface for the region, not a damage verdict for each house. Treating "inside the polygon" as "has a damaged roof" is the single most expensive mistake in storm restoration, and it's the default behavior most of these tools encourage.

4. Lead marketplaces and shared-lead services

These sell you contact info for homeowners who expressed interest — form fills, ad responses, sometimes call transfers. They are a demand-capture product, not a roof-finding product. Someone already decided they might want a roof; you're buying a slice of their attention, usually shared with competitors. There's a place for this, but it is the opposite of the problem this guide is about. You're not finding worn-out roofs; you're bidding on people who raised their hand. Margins compress fast when the same lead lands in four inboxes.

5. Roof-age and storm-modeled targeting platforms

The newest category, and the one most directly aimed at "which specific roofs are due." Instead of starting from a storm and listing everyone underneath it, or starting from a county record, these start from the roof itself: they estimate roof condition and age signals from aerial and historical imagery, and (the good ones) model how a given storm would have loaded each individual roof rather than just whether the address sat inside a swath. RoofPredict sits in this bucket; we'll cover it specifically, and honestly, later. The category's promise is a ranked list of doors — "knock these first" — rather than a raw dump of every house in a county or a polygon.

Most mature operations end up using two or three of these together: a targeting layer to choose doors, a measurement tool to estimate the ones that bite, and a CRM to run the pipeline. The expensive error is buying a tool from category 3 or 4 and believing it's doing the job of category 5.

A side-by-side: what each category answers

Read that table the way a buyer should: four of the five categories answer a question that is adjacent to "which roofs need replacing" without answering it directly. Only the last one is built to answer the question itself — and even it produces a ranked likelihood, not a confirmed verdict. Knowing this stops you from buying a measurement tool or a hail-map subscription and then being confused about why your reps still don't know which doors to knock.

How the data is actually made (so you can read the sales deck)

You cannot evaluate targeting software without understanding where its claims come from. Two data sources matter most: roof imagery and storm data. Here's how each is really produced, and where the soft spots are.

Roof age and condition from imagery

Nobody can read a literal install date off a photo of your roof. There is no metadata stamped on shingles. So when a tool claims to know "roof age," what it's really doing is one or more of these:

• Historical imagery differencing. Aerial and satellite imagery of the same address exists going back years. If a roof's appearance changed — color, reflectivity, texture — between a 2014 capture and a 2017 capture, that's evidence of a re-roof in that window. This is genuinely powerful: it can bracket when a roof was last replaced into a range, even when county permit records are missing (and permits are missing constantly, because plenty of re-roofs happen without one).
• Condition signals from current imagery. Streaking, granule loss patterns, patched sections, color fade, moss, and visible wear can be scored from high-resolution imagery. This estimates condition, which correlates with remaining life.
• Material and parcel priors. Combine the imagery read with "year built" and regional product norms to sharpen the estimate.

The honest output of all this is a range and a probability, not a date. "This roof was most likely re-covered between 2006 and 2010, and shows moderate wear" is a true and useful statement. "This roof is exactly 17 years old" is a fake-precision statement, and if a vendor talks that way, distrust the rest of their deck. A good targeting tool gives you a roof-age range and is upfront that it's an estimate. The value isn't pinpoint accuracy on one house — it's that across a whole farm area, the ranking is right often enough that your knocking time concentrates on genuinely older, worn roofs instead of being sprinkled randomly.

Storm data: what a hail swath really is

This is where most contractors get fooled, so it's worth being precise.

When you see a hail map, the colored swath is almost always derived from radar. The most common product underneath is MESH — Maximum Estimated Size of Hail — a radar-derived estimate. NOAA's National Severe Storms Laboratory developed the algorithms behind these products; they translate radar reflectivity aloft into an estimated maximum hail size at the surface. Two things about MESH that the colored polygon doesn't tell you:

1. It's an estimate of the maximum hail the storm could produce in that area, aloft — not a measurement of what actually hit each roof. Hail melts, gets blown sideways by wind, and falls unevenly. A radar pixel covers a big chunk of ground; within it, one street can get pelted and the next can get drizzled.
2. Ground truth is sparse. The actual confirmed hail reports that feed NOAA's Storm Prediction Center and the Storm Events Database come from spotters, the public, and trained observers — and there simply aren't many of them relative to the number of homes. So the swath is a model fed by thin verification.

None of this makes hail maps useless — they're a great first filter. It makes them dangerous when treated as a verdict. "This address is in a 1.5-inch MESH swath" means "a storm capable of damaging hail probably passed over this neighborhood." It does not mean "this specific roof is damaged." The difference between those two statements is the difference between a productive canvass and a week of crews getting doors slammed because three out of four roofs they knocked were fine.

Per-roof storm modeling: the meaningful upgrade

The better targeting approach doesn't stop at "is this address in the swath?" It asks: given this storm's hail size, density, direction, and timing, and given this specific roof's slope, orientation, age, and material — how hard was this particular roof actually loaded?

A few examples of why that matters:

• A steep south-facing slope and a low north-facing slope in the same hailstorm take very different impact energy because of the angle hail strikes them.
• A brittle 20-year-old 3-tab and a fresh impact-rated shingle in the same storm come out completely differently — IBHS impact testing shows shingle age and type strongly affect damage threshold.
• Wind-driven hail concentrates on the windward elevations; the leeward side of the same house may be untouched.

This is the line we draw at RoofPredict: we model the storm on each roof, not only where it passed. That's the difference between a swath-membership list and a genuinely ranked target list. It's also, importantly, still a probability — a roof that modeled high is a roof more likely to have functional damage, which means it's a better door to knock and inspect. It is not proof of damage. Nothing modeled from the sky is. The inspection on the roof is what documents condition; the model just tells you which ladders to set up first.

Hail vs. wind: two different damage signatures, two different targets

It's worth separating the two main storm forces, because they wear roofs differently and a good model treats them differently.

Hail damages by impact. The functional damage that shortens roof life is bruising — fractures in the shingle mat — and granule displacement that exposes the asphalt to UV and speeds aging. Severity scales with stone size, density, hardness, and the angle of impact, and it interacts with shingle age: IBHS impact work consistently shows older, more brittle shingles fracture at lower impact energy than fresh ones. So in a single hailstorm, the old roofs are both more likely to be functionally damaged and more likely to be genuinely due anyway. That's exactly the overlap a combined age-plus-storm model is built to surface.

Wind damages by uplift and flexion. As wind flows over a roof it creates lift on the leading edges, eaves, rakes, and ridges; tabs flutter, the sealant strip fails, and shingles crease or tear. The National Weather Service notes that damaging straight-line winds and the wind field around severe storms can exceed the design tolerances of aging or poorly fastened shingles. Wind damage concentrates on the windward elevations and the edges, which is the opposite distribution from where you'd expect — and it means a roof can look fine from the street while the windward field is compromised.

Why this matters for software: a tool that only knows "hail swath" misses wind-driven losses entirely, and a tool that treats the whole roof as one surface misses that damage clusters on particular slopes and edges. Per-roof modeling that accounts for orientation and the storm's direction is what tells a rep which side of the house to inspect first — a small thing that makes the on-roof time faster and the documentation cleaner.

The honest limits of every roof-finding tool

If you remember one thing from this guide, make it this: all of this software produces a prioritized list of where to look, not a confirmed list of who needs a roof. The roof itself, inspected by a competent person, is the source of truth. Software that pretends otherwise is selling you a liability.

Concretely, here's what no targeting tool can do, no matter the marketing:

• It can't confirm damage. Functional hail damage is verified by hands-and-eyes on the roof — test squares, chalk, feeling for bruises, checking soft metals and mats. A map cannot do this.
• It can't decide a claim. Whether storm damage is covered is the insurance carrier's call, based on the policy and the adjuster's inspection. The contractor's job is to document conditions and provide an honest estimate of the work. The homeowner owns their claim and their decision. Software that implies it can get a claim approved, promise a "free roof," or guarantee a deductible outcome is steering you toward conduct that regulators and the better state contractor and insurance rules frown on hard. Don't.
• It can't replace the conversation. Age and storm data get you to the right door. A homeowner still has to want an inspection and trust the person on their porch.

Used correctly, the value is enormous: you spend your finite knocking and driving hours on the roofs most likely to be real, instead of carpet-bombing a ZIP code. Used as a crutch — "the software said this house, so it must need a roof" — it'll get your reps tossed off porches and, worse, put your company on the wrong side of unfair-claims rules.

A field workflow: from county-wide to the right doorbell

Here's a concrete, repeatable process that combines the tool categories above into an actual day's work. Adapt the specifics to your market, but the funnel logic holds.

Step 1 — Define the farm, not the storm

Many contractors start with a storm and react. Better operators start with a farm: a defined geography where they want to build density, reputation, and referral compounding. A farm might be three or four subdivisions built in a similar era with aging roofs. Why farm-first? Because density wins — clustered jobs cut windshield time, multiply yard-sign and referral effects, and let you dominate a neighborhood instead of scattering one job per ZIP. Pull parcel data to understand the housing stock: build years, owner-occupancy, single-family share.

Step 2 — Layer roof-age signals

Onto your farm, layer roof-age-range data. Now you're not only looking at "houses built before 2005"; you're looking at "houses whose roof was most likely last replaced 15–22 years ago and shows moderate-to-heavy wear." This is the difference between a list of 4,000 houses and a ranked shortlist of the 600 most likely to be due on age alone. Sort descending by age-and-wear score.

Step 3 — Overlay storm exposure (when relevant)

If a real storm has come through, overlay it — but use per-roof modeling, not raw swath membership, if you have it. The roofs that score high on both age/wear and storm loading are your A-list. A 19-year-old worn roof that took a windward face of 1.5-inch hail is a dramatically better knock than a 6-year-old roof on the leeward side of the same street, even though a basic hail map paints them the same color.

Step 4 — Build the route

Turn the A-list into an efficient walking/driving route. Tight clusters first. A rep who knocks 80 well-chosen doors in a compact area beats a rep who drives 40 miles to knock 80 scattered ones. Many canvassing CRMs do this routing; the point is that good targeting upstream makes the routing downstream actually pay off.

Step 5 — Knock with a real reason

The pitch changes when your targeting is good. Instead of "we're in the neighborhood, any storm damage?" — which every homeowner has heard fifty times — your rep can lead with something specific and honest: "We've been inspecting roofs on this street that are in the age range where wear starts showing up, and a number of them had issues from the storm earlier this month. We'd be glad to take a look and tell you straight whether yours is fine or not." That's truthful, it's specific, and it respects the homeowner. Train reps to under-promise: the inspection's job is to find out, not to assume.

Step 6 — Inspect, document, hand off

On the roof, document conditions properly — photos, test squares, measurements. This is where the measurement tool and a clean inspection app earn their place. If there's storm damage and the homeowner wants to pursue a claim, the contractor documents and estimates; the homeowner files; the carrier's adjuster decides. Keep those lanes clean. It protects the homeowner, and it protects you.

Step 7 — Feed results back into targeting

The loop that separates pros from dabblers: track which targeted doors converted to inspections and jobs, by age band and storm score, and feed it back. Over a season you learn which signals predict real roofs in your market and tune accordingly.

Working the same farm twice: the maintenance pass

A farm isn't a one-and-done. The smartest operators run a defined geography on a cadence — a heavy first pass, then lighter return passes as roofs age into the window and as storms roll through. Because age data shifts predictably (a roof that was a "maybe" at 13 years is a strong knock at 16), and because storms re-shuffle the deck overnight, the same neighborhood produces fresh A-lists over time. The maintenance pass is where the targeting investment compounds: you already have brand presence and yard signs there from the first pass, so the second pass converts warmer. This is also why farm-first beats storm-chasing for a company trying to build durable density rather than chase the next county.

Staffing the funnel: who does what

A targeting program only pays off if the handoffs are clean. A workable division of labor for a small-to-mid crew:

• Sales manager / owner owns the targeting layer: defines farms, pulls the A-lists, sets the week's routes. This is a thirty-minute desk job that determines whether forty rep-hours are spent well or wasted.
• Canvassers / reps work the routes, knock honestly, and book inspections. Their job is to get on roofs, not to diagnose from the porch.
• Inspectors / estimators verify condition on the roof, document, and produce the estimate. This is the source-of-truth step.
• Office / CRM admin tags every result back against the age band and storm score, closing the feedback loop.

When one person tries to do all four, the feedback loop is the first thing that gets dropped, and the program slowly drifts back into random knocking.

A worked example: 4,000 houses down to a 9-door morning

To make this concrete, walk through a simplified version of what the funnel does to the numbers. (These are illustrative figures to show the mechanics, not a guarantee — your market will differ.)

Say your farm contains 4,000 single-family homes.

• Parcel filter — owner-occupied, single-family, built before 2008: knocks it to roughly 2,600 plausible candidates. You've removed rentals, new builds, and non-targets.
• Roof-age-range layer — keep only roofs most likely last replaced 14+ years ago with visible wear: down to about 700. You've removed the houses that got a new roof five years ago and would just waste a knock.
• Storm-loading model (a real storm came through) — keep roofs that modeled meaningful impact on at least one major slope: down to about 180.
• Route clustering — your rep takes the densest cluster for the morning: about 30 contiguous A-list doors.
• Contact + interest — a third are home and open to an inspection: roughly 9–10 real inspection opportunities before lunch, every one of them a roof with a genuine reason to look.

Now compare that to the default storm-tool workflow: dump every address in the hail swath — call it 2,400 addresses — and have reps knock them in whatever order, with no idea which roofs are old, which faces took the hail, or which got a new roof last year. Same hours, a fraction of the hit rate, and a lot more slammed doors souring the neighborhood on your brand.

The software didn't replace the rep, the ladder, or the inspection. It just made sure the rep spent the morning on 9 roofs worth inspecting instead of 30 roofs worth driving past.

The economics, without the hype

Let's put rough mechanics on why this matters to the P&L, using cost-per-real-inspection as the unit (the only number that actually maps to revenue). Suppose two reps each have a 6-hour knocking day and can knock about 10 doors an hour.

• Untargeted (knock the swath): Of those 60 doors, maybe a third answer, and of those, a slice are even plausibly due — many are recent roofs or undamaged. You might net 2–4 genuine inspection opportunities for the day, and you've soured a bunch of doors with a generic pitch.
• Targeted (knock the ranked A-list): The same 60 doors, but every one is an older or storm-loaded roof. The answer rate is similar, but a far higher share of the people who answer have a roof worth inspecting. Netting 8–12 genuine inspection opportunities from the same hours is realistic.

The rep cost is identical either way. The data cost is the variable. So the real question isn't "how much does the software cost" — it's "what does each real inspection cost me, all-in, with vs. without it?" When you measure that way, a targeting layer that doubles the share of productive knocks is usually cheaper per real inspection than the rep hours it saves from being wasted, and far cheaper than shared leads on a cost-per-inspection basis. That's the comparison to run in a trial, and it's the one most contractors skip because they're anchored on list price instead of unit economics.

How to evaluate a roof-finding tool: a buyer's checklist

When a vendor demos, run these questions. The answers separate real targeting from a pretty map.

On roof age and condition

• Do you give roof age as a range with a confidence level, or a single fake-precise date? (Range = honest.)
• What's your data source — historical imagery differencing, condition scoring, parcel priors, or a blend? Can you show me the imagery you're reasoning from?
• How recent is your imagery in my market? Coverage and freshness vary wildly by region.
• How do you handle roofs replaced without a permit? (If they lean only on permit records, that's a big blind spot.)

On storm data

• Is your storm layer raw MESH/swath membership, or do you model loading per individual roof (slope, orientation, material, age)?
• Do you represent storm output as a probability/odds of damage, or do you imply it's proof of damage? (Proof claims are a red flag — legally and practically.)
• What's your ground-truth or verification story for hail size?

On the list it produces

• Does it hand me a ranked list (knock these first) or a raw dump of everyone in a polygon?
• Can I combine age and storm signals, or only one at a time?
• Does it export cleanly into my CRM / canvassing routing?

On honesty and compliance

• Does any of your marketing language promise approved claims, "free roofs," or covered deductibles? (If yes, walk away — that's how reps and companies get in trouble with state insurance and contractor rules.)
• Are you clear, in writing, that your output guides where to inspect, not who is damaged?

On economics

• How does the cost compare to what I'd spend on shared leads for the same number of real inspections? (Targeting tools and lead marketplaces are different spend; compare cost-per-real-inspection, not cost-per-list.)

Any vendor who gets cagey on the age-as-range, storm-as-probability, and "it's not proof of damage" questions is selling confidence, not data.

Where RoofPredict fits — and where it doesn't

I'll be direct about the tool we build, including what it won't do.

RoofPredict is a roof-age and storm-modeled targeting platform (category 5 above). It's built for the specific job this whole article is about: telling a roofing contractor which roofs, house by house, are due — so your crews knock the roofs a storm actually wore out and the roofs aging out, instead of carpet-bombing a swath or buying shared leads.

Two things it does that matter:

1. Roof-age range per address, from aerial imagery. Not a fake install date — an honest range ("most likely last re-covered in this window") with wear signals, derived from historical and current imagery. Across a farm, that turns a flat list into a ranked one: the genuinely older, more worn roofs rise to the top of your route.
2. Storm physics modeled per roof. When a storm comes through, we don't just check whether an address sat inside a hail polygon. We model how that storm — its hail size, density, direction — would have loaded each individual roof, accounting for slope and orientation. That's the "we model the storm on each roof, not only where it passed" idea, and it's why the A-list it produces is tighter and truer than swath membership.

The output is a ranked list of doors and routes: knock these first, here's why each one made the list.

Now the honest limits, because you should hear them from us:

• It is not proof of damage, and we don't pretend it is. A high storm score means a roof was more likely loaded hard enough to have functional damage — it's odds, not a verdict. The inspection on the roof is what documents condition. Roof age is a range, not a date.
• It does not touch claims. RoofPredict doesn't file, handle, approve, or guarantee anything about an insurance claim, a deductible, or a "free roof." That's not our lane and it's not yours to promise. The contractor documents conditions and estimates; the homeowner owns and files the claim; the carrier's adjuster decides coverage. We keep those lanes clean on purpose — it protects the homeowner and keeps you on the right side of unfair-claims rules.
• It doesn't knock the doors or close the deal. It gets your reps to the right porches with a real reason. The conversation, the inspection, and the craftsmanship are still yours.
• It depends on imagery coverage and freshness, which vary by region. We're straight about where our data is strong and where it's thinner.

If you're tired of paying for the same shared lead as three competitors, or watching crews burn a week knocking a hail swath where half the roofs are six years old, that's the gap this is built for. It's a targeting layer, not a miracle.

Common mistakes that quietly wreck a targeting program

Even with good software, here's what I see sink contractors. Most of these cost you weeks before you notice.

Treating swath membership as damage. Already covered, but it's the number-one killer, so it's first. "They're in the polygon" is a reason to look, not a reason to assume. Reps who knock with assumed damage get caught flat-footed when the roof is fine, and they sound like every other storm chaser.

Confusing house age with roof age. Year-built is not roof-age. A whole subdivision built in 1998 will have a wide spread of actual roof ages by now — original roofs, post-storm replacements from 2011, recent flips. Targeting on house age alone wastes a third of your knocks on roofs that were already done.

Ignoring imagery freshness. If a tool's newest imagery for your market is four years old, a roof replaced last summer still looks old to it. Always ask how fresh the data is where you work, and sanity-check the first few houses on a route by eye before committing a crew.

Over-targeting the same handful of streets. If five companies all bought the same hail list, those homeowners are getting knocked into the ground. Farm-first thinking and unique targeting signals (age + per-roof storm load) get you to doors the swath-list crowd isn't all hitting at once.

No feedback loop. Running targeting and never tracking which signals actually converted in your market means you never get smarter. Tag every inspection with the age band and storm score that sent you there. After a season you'll know which signals to trust locally.

Letting the software write checks the inspection can't cash. If your reps lead with "you have damage" or "this'll be a free roof," you've turned a good targeting tool into a compliance problem. Lead with "let's find out," and keep claim decisions with the carrier and the homeowner.

Buying a list product when you needed a targeting product. Shared leads and parcel dumps have uses, but if your goal is finding worn-out roofs before the homeowner calls, neither one does that job. Match the tool to the actual problem.

Staying clean: compliance, claims, and how you talk about damage

Targeting software puts your reps in front of more storm-exposed roofs, which means it also raises your exposure to the rules that govern how roofers and homeowners interact with insurance. This isn't legal advice, but a few principles keep good companies out of trouble — and they map directly onto how you should and shouldn't use the data.

Keep the three lanes separate. The contractor documents conditions and provides an honest estimate of the work. The homeowner owns the claim and decides whether to file. The insurance carrier's adjuster decides coverage. State insurance regulators — Texas's TDI publishes plain-language guidance on storm and roof claims, and most states have similar — take a dim view of contractors blurring those lanes. Your software's job is to tell you where to inspect. It is not, and should never be marketed as, a tool that decides or guarantees a claim outcome.

Don't let the model become a damage claim. A high storm score is a reason to inspect, expressed as odds. It is not evidence of damage and shouldn't be presented to a homeowner or an adjuster as if it were. "Our data flagged your roof's age and the storm exposure, so we'd like to inspect" is honest. "Our software shows your roof is damaged" is not — you haven't been on it yet. The FTC's general guidance on truthful, substantiated advertising applies to roofers like anyone else: don't claim what you can't back up.

Avoid the promises that get licenses pulled. No "free roof." No guaranteed deductible outcome. No "we'll get your claim approved." These show up in storm-chaser pitches and they're exactly the language that draws regulatory attention and erodes homeowner trust. Targeting more roofs is no excuse to pitch sloppier.

Mind safety once you're targeting volume. More inspections means more ladder time and more roof time. OSHA's fall-protection requirements in construction are the baseline, and a targeting program that floods your inspectors with roofs shouldn't outrun your safety discipline. Faster targeting upstream is only a win if the on-roof work stays safe and documented.

Used this way, the data is a pure asset: it gets you to the right roofs and keeps you honest about what you've actually verified. Used to manufacture urgency or imply damage you haven't confirmed, it's a liability with a subscription fee.

What separates operators who make this work from those who don't

After watching a lot of companies adopt targeting tools, the ones who get real lift share a few habits, and the ones who churn out of the tools share a few failures.

Winners treat the tool as a prioritizer, not an oracle. They use the ranked list to decide knocking order, then trust the inspection for truth. They never tell a rep "this house has damage" — they tell the rep "this house is worth looking at, here's why."

Winners run a tight farm. They pick a geography and own it across passes, rather than scattering knocks across whatever county got hail. Density compounds; scatter doesn't.

Winners close the loop. They tag every outcome back to the signal that produced it, so after one season they know whether 15-year or 17-year age bands convert better in their market, and whether their region's storms reward hail scoring or wind scoring more.

Winners sanity-check the data against their own eyes. Before committing a crew to a route, they eyeball the top few houses in imagery or in person. They know imagery freshness varies and a recent re-roof can fool an old capture, so they spot-check.

The companies that churn do the opposite of each: they believe the list literally, they chase storms instead of farming, they never measure which signals worked, and they trust the screen over the roof. The tool is the same in both hands. The discipline around it is the variable.

Tying it together: the stack that actually finds roofs

For most growing roofing companies, the working stack looks like this:

Notice what's doing the actual finding: the targeting layer. Everything else either filters before it or executes after it. If you only upgrade one part of your operation this year, upgrade the part that decides which doors are worth your crew's time. That's the lever. Knocking faster doesn't help if you're knocking the wrong roofs.

The goal was never "more software." It's spending your finite hours — the reps, the ladders, the daylight — on the specific roofs that are genuinely due. Age tells you which roofs are wearing out. Per-roof storm modeling tells you which roofs a storm actually worked over. Put those together, route them tight, knock them honest, and let the roof itself be the final word.

If you want to see what a ranked, roof-by-roof target list looks like for your own market — age ranges and per-roof storm modeling instead of a hail swath and a prayer — that's exactly what RoofPredict is built to show you. Book a demo and bring a neighborhood you already know; the fastest way to judge any targeting tool is to point it at streets where you already know which roofs are old.

FAQ

Can software really tell me which roofs need replacing before the homeowner calls?

It can tell you which roofs are most likely due, not which are confirmed bad. The best tools estimate a roof-age range from aerial imagery and model how a storm loaded each individual roof, then rank addresses so you knock the most likely candidates first. The roof still has to be inspected to confirm condition. Think of it as a high-quality prioritized list of where to look, not a verdict on who needs a roof.

How does software estimate roof age if there's no install date anywhere?

There's no date stamped on a roof, so tools infer it. The strongest method is differencing historical aerial imagery: if a roof's appearance changed between captures a few years apart, that brackets when it was re-covered. Tools also score visible wear (granule loss, streaking, patches) from current imagery and blend in parcel data like year built. The honest output is a range with a confidence level, not an exact year.

What's the difference between a hail map and per-roof storm modeling?

A hail map shows a swath, usually derived from radar estimates like MESH, telling you where damaging hail probably fell across a region. It treats every address inside the polygon the same. Per-roof modeling goes further and estimates how the storm loaded each specific roof based on its slope, orientation, material, and age, plus the storm's hail size and direction. The first answers 'where did it hail'; the second answers 'which roofs likely got hit hard.'

Is being inside a hail swath proof that a roof is damaged?

No. A swath is a probability surface for an area, based on radar estimates with sparse ground verification. Hail falls unevenly within it, and an address inside the polygon may have taken little or no impact. Functional damage is confirmed only by a hands-on roof inspection. Treating swath membership as proof of damage is the most common and costly mistake in storm restoration, and claiming damage you haven't verified creates real compliance risk.

How is this different from buying shared roofing leads?

Shared leads are homeowners who already raised their hand, usually sold to several contractors at once, so you're bidding on attention and competing on speed. Roof-finding software works the opposite direction: it identifies worn-out or storm-loaded roofs before the homeowner has reached out, and gives you doors competitors buying the same lead list aren't necessarily hitting. One is demand capture; the other is targeted prospecting.

Does aerial measurement software find roofs that need replacing?

No. Measurement tools return the size and shape of a roof you already chose, squares, pitch, facets, and lengths, so you can estimate and order material without climbing it. They don't tell you whether a roof is worn out or storm-damaged. They're an execution tool you use after targeting has picked the door, not a prospecting tool that picks doors for you.

Will targeting software handle insurance claims or get a claim approved?

No, and you should be wary of any tool that implies it can. The contractor documents roof conditions and provides an honest estimate; the homeowner owns and files the claim; the insurance carrier's adjuster decides coverage. Software that promises approved claims, covered deductibles, or a free roof is steering you toward conduct that state insurance and contractor rules treat harshly. Good targeting tools stay in the targeting lane.

How accurate is roof-age data, really?

On any single house it's an estimate expressed as a range, and it can be off, especially where aerial imagery is old or a roof was replaced without a permit. The value is statistical: across a whole farm area, the ranking is right often enough that your knocking hours concentrate on genuinely older, worn roofs instead of being spread randomly. Always check imagery freshness in your specific market and sanity-check the first few houses by eye.

What should I look for when evaluating roof-finding software?

Ask whether roof age comes as a range with confidence or a fake-precise date; whether storm data is raw swath membership or per-roof modeling; whether the output is a ranked list or a raw dump of everyone in a polygon; how fresh the imagery is in your market; whether it exports to your CRM; and whether any marketing promises approved claims or free roofs. Honest answers on age-as-range and storm-as-probability separate real targeting from a pretty map.

Can I just use county property records instead of paying for software?

County data is a useful coarse filter for owner-occupancy, single-family status, and year built, and you can pull it yourself. But year built is the age of the house, not the roof, and permits miss many re-roofs. Property records can't tell you a roof's condition or how a storm loaded it. They narrow the field; they don't find the worn-out roofs. Pair them with a targeting layer that actually reads the roof.