How Roofers Can Tell If a Roof Has Already Been Replaced

Walk any street and you will pass two kinds of houses that look identical from the sidewalk: one wearing its original 22-year-old roof, and one that got a full tear-off four years ago. To a homeowner they read the same. To a roofer who knows what to look for, they could not be more different. One is a job. The other is a polite goodbye at the door and a wasted half hour.

Knowing the difference before you commit time, gas, payroll, or a quote is one of the most underrated skills in residential roofing. Get it wrong in the optimistic direction and you knock doors that closed years ago. Get it wrong in the pessimistic direction and you skip a 20-year-old roof that was due for a conversation. The good news is that a replaced roof leaves fingerprints everywhere: on the surface, in the attic, in the gutters, on paper at the county, and in the aerial record. You just have to know which fingerprints to dust.

What follows is a working field guide for telling whether a roof has already been replaced, written for the person who has to make that call fast and be right more often than not. We will move from the cheapest signals (what you can read from the truck) to the most definitive (permits and dated imagery), and cover the edge cases that fool even experienced crews: roof-overs, partial replacements, repaired-not-replaced roofs, and the homeowner who genuinely does not know.

Why "has it been replaced?" is the wrong-but-right first question

Notice the real question underneath the obvious one. You do not actually care whether a roof was replaced for its own sake. You care because replacement resets the clock. A roof that was torn off and redone five years ago is a roof with 15 to 20 years of service life left, an intact warranty, and an owner who is not in the market. A roof that has never been touched and is pushing the end of its rated life is the opposite: a near-term decision, a potential storm claim, and a homeowner who may not yet know they have a problem.

So "has it been replaced" is really three questions wearing one coat:

1. Is this roof old enough to be worth a conversation? (age and remaining life)
2. Has the clock already been reset by a recent re-roof? (replacement history)
3. Did something happen to it that the owner has not addressed? (storm and wear)

The inspection skills below answer all three, but keep the framing in mind. A roof that was replaced is not a failure to detect a job; it is a correct read that saves you a knock. The point is not to catch every re-roof for trivia's sake. The point is to spend your hours on the houses that can actually buy a roof.

The cost of guessing

Put numbers on it, because the numbers are why this matters. Say a sales rep can productively work 40 to 60 doors in a day of canvassing, and a direct-mail piece runs somewhere in the range of $0.50 to $1.00 all-in once you count printing, list, and postage. If a fifth of the homes you target already have a four-year-old roof, you are not losing a little efficiency. You are spending a fifth of your gas, your stamps, and your rep's daylight on houses that physically cannot say yes.

Now flip it. One residential re-roof is worth thousands of dollars in revenue. The math of qualifying is brutal in your favor: every hour you do not spend on already-replaced roofs is an hour pointed at roofs that can buy. The skill of reading replacement is, in plain terms, a margin skill.

The full picture: what "replaced" can actually mean

Before the signals, get the categories straight, because "the roof was replaced" hides at least five different situations, and they do not all reset the clock the same way.

The lesson: "new-looking" and "recently replaced" are not synonyms, and "replaced" and "reset" are not synonyms either. A roof-over can look fresh and still be near the end of its useful life, with code problems baked in. A partial job can read as new on the slope you can see and be 20 years old on the one you cannot. Hold these categories in your head as you work the signals, and you will avoid the two classic mistakes: writing off a roof-over as a recent tear-off, and writing off a whole house because one visible slope got patched.

Reading the roof from the ground (the truck-window pass)

The fastest, cheapest read happens before you leave the vehicle. A trained eye can sort most homes into "old enough to talk to" versus "probably recent" from the street in well under a minute. None of these signals is conclusive alone. Stacked together, they get you to a confident bet.

1. Granule loss and surface texture

Asphalt shingles are protected by mineral granules embedded in the asphalt. Those granules are the roof's sunscreen, and they wear off over time. A young roof has a deep, even, almost velvety color and crisp granule coverage. An old roof looks thin, faded, and shiny in patches where the asphalt is showing through. Where you can see slick black or gray asphalt peeking between the granules, that field is well into its second half.

• New / recently replaced: rich, saturated color; uniform texture; granules tight and full.
• Mid-life: color holding but starting to lighten; minor granule thinning at the most sun-exposed slope.
• End-of-life: bald spots, shine, visible asphalt, color washed out; often worst on the south- and west-facing slopes that take the most sun.

Watch the south and west exposures first. The sun does the most damage there, so on a roof that has not been replaced, those slopes age fastest. If the south slope looks tired and the north slope looks decent, that is a normal, un-replaced roof aging the way physics says it should. If every slope looks uniformly fresh, that points toward a recent full replacement.

2. Granules in the gutters and at the downspout splash

You do not need to be on the roof to see what the roof is shedding. Look at the gutters, and look at the ground where the downspouts drain. A roof actively losing its surface deposits granules that look like coarse black or gray sand. Fresh-looking piles of granule grit at the downspout outlet are a strong sign of an aging roof shedding its protection. A recently replaced roof sheds a small amount of loose manufacturing granules in the first months and then very little.

So: heavy granule accumulation at the splash blocks = old roof. Clean splash blocks under an otherwise tired-looking roof = look closer, something does not add up (could be a recent clean, could be a roof newer than it looks).

3. Edge and ridge lines

The silhouette tells you a lot. Run your eye along the ridges, hips, rakes, and eaves.

• Curling, cupping, and lifting at shingle edges is an age and heat signal. Tabs that curl up at the corners or cup in the middle are old. A recently replaced roof lies flat.
• Sagging ridge or wavy planes suggest deck or structural age, not a recent re-deck.
• The double-thick edge of a roof-over. This is the single most useful curb tell for a layered roof. At the rake and eave, a roof-over often shows an obviously taller, two-course-thick edge and a heavier shadow line because there are literally two roofs stacked there. If the edges look chunky and the field looks slightly lumpy or uneven, suspect a layer-over rather than a clean tear-off.

4. Flashing, vents, and accessories

The roof itself can be re-shingled while the metal stays old, or everything can be replaced together. The accessories are a tell either way.

• Bright, unweathered flashing around chimneys, in valleys, and at wall lines, paired with fresh shingles, points to a complete recent job.
• Old, rusted, or paint-peeling flashing sitting against fresh-looking shingles is a flag: either a cheap re-shingle that reused old metal, or a roof-over, or a partial job. None of those is a clean recent tear-off, and all of them mean the roof has problems a careful eye should note.
• Vent and pipe boot condition. Plastic and rubber pipe boots crack and dry-rot in roughly 8 to 12 years of sun. A roof with several cracked, sun-rotted boots has been up there a while. Bright, supple, uniform boots suggest a newer install.

5. Color and plane uniformity (the partial-job detector)

Stand where you can see two or more roof planes at once and compare them. A full replacement produces uniform color and weathering across every plane (allowing for the normal sun difference between exposures). A partial replacement, a repair, or a re-shingle done in stages produces a mismatch: one plane noticeably newer than another, or a crisp rectangular seam where new shingles meet old. That seam is the signature of a partial. When you see it, the correct read is not "new roof" but "part of this roof is old and somebody already had a reason to touch it."

6. Drip edge, ridge caps, and starter detail

Subtle, but they separate a real recent job from a tired one. Modern complete replacements almost always include a metal drip edge at the eaves and rakes, factory ridge-cap shingles or a ridge-vent system, and a proper starter course. Older original roofs frequently lack a crisp drip edge or use cut three-tab tabs for ridge caps instead of dedicated cap shingles. A clean drip edge plus a ridge vent plus dimensional shingles is a strong modern-replacement signature. The absence of those details on an older-looking roof says original or very old.

Putting the curb pass together: a quick scoring habit

No single sign decides it. Build a fast mental tally. Here is a workable rule of thumb you can run in 45 seconds per house:

Three or more signals pointing the same way is a confident read. A split tally is your cue to use a sharper tool (binoculars, the attic, the permit record, or dated aerial imagery) rather than guess.

Estimating roof age from the ground (without the homeowner's help)

Most of the time the homeowner cannot tell you when the roof was done, especially if they bought the house used. So you learn to estimate. Estimating age is the inverse of detecting replacement: if you can credibly put a roof at 18 to 22 years, you have effectively confirmed it was not recently replaced.

Anchor to the typical service life of the material

Different roofing materials have very different expected lifespans, and knowing the material narrows your age guess. Industry and manufacturer guidance puts typical service life roughly here (real-world conditions shorten the top end; sun, hail, and poor ventilation shorten it a lot):

The single biggest lever here is telling 3-tab from architectural at a glance, because it changes the whole age math. A faded 3-tab roof at 18 years is at the end of the line. An architectural roof at 18 years may have a decade left. Get the material right first.

Cross-reference the neighborhood and the house

Roofs in a subdivision were often built and later re-roofed in clusters. Storms hit whole streets, and re-roof waves follow. If you are looking at a 1999-era subdivision and most roofs show the same tired 3-tab, the few that look crisp and dimensional are your recent replacements, and the tired ones are very likely original or close to it. The houses tell on each other.

Also use the house's own clues: the year built (public record, more on this below), the style of the home, and whether the roof material matches the era. A 25-year-old house with an original 3-tab roof is a textbook end-of-life candidate. The same house with a fresh architectural roof is a re-roof you can date roughly by how new it looks.

The hard limit of eyeballing age

Be honest about what the ground read can and cannot do. From the curb you can confidently sort roofs into broad bands: clearly new (0-5), clearly mid-life, clearly old (likely 18+). What you cannot do from the street is read an exact install date. A roof is best described as an age range, not a date, and anyone who tells you they can pin a curb-side roof to a specific year is overselling. The honest output of a ground inspection is a range with a confidence level, and that is genuinely enough to make the targeting decision.

What the attic and the up-close inspection reveal

When you earn a closer look, whether on an inspection appointment or once you have permission, the attic and the roof surface tell you things the curb cannot. This is where you confirm a replacement and, just as importantly, find evidence of a roof-over or a deck problem.

Count the layers

The most direct way to know whether a roof was torn off or layered over is to count the courses of roofing material. You can do this at the eave edge, at a rake, or by lifting a shingle near a vent or valley.

• One layer over the deck and underlayment = at some point this was torn off to the deck (or it is the original single layer).
• Two or more layers = a roof-over happened. The visible top layer may be only a few years old while the system as a whole carries the problems of the old roof underneath: trapped heat, telegraphed unevenness, and no inspection of the deck when the new layer went on.

Most building codes, following the International Residential Code, limit a roof to a maximum of two layers of asphalt shingles and require a tear-off rather than another overlay beyond that, or where the existing roof is water-soaked or deteriorated. So a roof you find carrying two layers is at the end of the overlay road: its next roof must be a full tear-off. That is useful to know and useful to explain.

Look at the deck from inside

In the attic, the underside of the roof deck records history. Look for:

• New, light-colored, clean sheathing versus old, darkened, water-stained wood. Patches of new plywood among old boards mean deck was replaced, which only happens during a real tear-off or a repair.
• Multiple rows of old nail holes in the rafters or sheathing. Each re-roof leaves its own nailing pattern. Several generations of nail holes mean several roofs over the years.
• Daylight, active leaks, or rot. These tell you condition regardless of age, and they are the things that turn an aging roof into an urgent one.

Read the up-close surface

Close up, granule embedment, sealant strip bonding, and brittleness tell age. Old shingles are stiff and crack when bent; you can sometimes snap a tab corner that should flex. The self-seal adhesive strips on old shingles lose their bond, which is why old roofs lose shingles in wind. A roof whose tabs lift easily by hand and whose corners crack is old, full stop, no matter how recently the surface was hosed off to look presentable.

The paper trail: permits, public records, and what they really tell you

The surface and the attic give you a physical read. The paper trail can sometimes give you a date. Used well, public records turn a guess into a fact. Used carelessly, they mislead, because the most commonly cited record is the one most likely to be wrong about the roof.

The Zillow / county "year built" trap

The most common mistake roofers and homeowners both make: treating the home's year built as the roof's age. Real-estate listings and county assessor records prominently show the year the house was constructed. That number tells you when the original roof went on, and nothing about whether it has been replaced since. A 1995 house may have had two roofs since then. The year built is a ceiling on roof age (the roof cannot be older than the house), never the roof age itself. Re-roofs are invisible to these records. Anchor on year built for context, never as an answer.

Building permits: the gold standard, with caveats

A re-roof in most jurisdictions requires a permit. Where permits are filed and digitized, a property's permit history is the single most reliable public evidence that a roof was replaced and when. Many county and city building departments publish permit lookups online; you search the address and read the permit type, date, contractor, and sometimes the final inspection sign-off.

A roofing or re-roof permit dated four years ago is about as close to proof of replacement as you will get without climbing up. Here is how to actually use them, and where they fall short.

How to pull a permit:

1. Find the jurisdiction that has authority (city building department if inside city limits, otherwise the county).
2. Use their online permit portal, or call the building department, or submit a public-records request. Search by the property address or parcel number.
3. Look for permit type containing "roof," "re-roof," "reroof," or "tear-off," with an issue date and ideally a final inspection date.
4. Note the contractor of record and the valuation; both add context.

Where permits mislead:

• No permit does not mean no roof. Plenty of re-roofs, especially insurance-driven storm jobs and cash jobs, get done without a pulled permit. An empty permit history is weak evidence of an original roof, not proof of one. Many areas have spotty enforcement, and some jurisdictions historically did not require permits for like-for-like re-roofs.
• A permit pulled does not always mean the work finished. Look for the final inspection.
• Records vary wildly by jurisdiction. Some counties have clean digital portals back 20 years; others are paper-only or do not publish online. Rural areas are often thin.
• Permit lag. A very recent job may not be in the system yet.

So permits are powerful when they exist and say "re-roof, 2021, final inspection passed." They are much weaker as a way to prove a roof has NOT been replaced, because the absence of a record proves little.

Insurance claim history and prior storm dates

If a homeowner will talk, ask whether they have filed a roof claim before, and when their last hail or wind event was. A prior approved claim usually means a replacement followed. You will not have direct access to their claim file, and you should not pretend to, but the conversation surfaces the history. Pair it with the regional storm record (more on that next) and you can often reconstruct whether a recent storm already drove a re-roof on that street.

Storm history: the signal that connects age to opportunity

Replacement and storms are tightly linked, because hail and high wind are what drive most re-roof waves in much of the country. Understanding the storm history of a specific address does two things at once: it helps you understand whether a recent storm already caused a replacement, and it helps you understand whether an old, un-replaced roof took damage that the owner has not yet acted on.

National sources let you reconstruct severe-weather history at a regional level. The National Weather Service and the Storm Prediction Center publish severe-storm and hail reports, and NOAA's Storm Events Database records significant hail and wind events by date and area. With those, you can answer: did this area get hit, and when?

But here is the honest limitation that every storm map shares, and it is a big one: regional storm data tells you where it hailed, not which individual roofs the storm actually wore out. Hail falls unevenly. Wind funnels and shadows. Two houses on the same block can take very different beatings from the same cell, depending on slope orientation, exposure, tree cover, and the storm's exact track. A hail-swath map shows you a county or a polygon. It does not show you the roof.

That gap, between "where the storm passed" and "which roofs it actually damaged," is the gap between a generic storm map and a usable target list. Closing it is where modeling per roof, rather than per region, earns its keep.

Doing this at scale: from one-by-one to a sorted street

Everything above works house by house, and you should know how to do all of it by hand, because the field instinct is what keeps you honest. But a sales manager covering thousands of addresses cannot eyeball every roof or pull every permit one at a time. The skill scales only if the data does. This is where aerial imagery and per-roof modeling change the economics of qualifying.

What aerial imagery can and cannot tell you about replacement

High-resolution aerial and satellite imagery, especially when you can compare images from different dates, is a genuinely powerful replacement detector at scale. The idea is simple: if you have an image of a roof from 2017 and another from 2023, a change in color, granule pattern, or shingle layout between the two dates is direct evidence that the roof was replaced in that window. A dated image archive turns "is it new?" into "it changed between these two dates," which is far stronger.

What imagery does well:

• Detect surface change over time (a re-roof shows up as a color and texture shift between dated images).
• Read uniformity, plane mismatch, and the broad weathering state of the surface across an entire area at once.
• Cover thousands of roofs in the time it would take to drive a few streets.

What imagery cannot do, and where honesty matters:

• It cannot read a specific install date off a single undated image any more than your eyes can from the curb. The output is an age range, not a certificate.
• It cannot see the deck, count layers reliably, or confirm a roof-over the way an attic can.
• It can be fooled by cleaning, by lighting and image-capture differences between dates, and by partial work.

Used well, aerial imagery does at scale exactly what your trained eye does at the curb: it sorts roofs into confident bands of likely age and flags the ones that changed. It does not replace a real inspection, and any honest tool says so.

Where RoofPredict fits

This is the part of the job RoofPredict was built for: doing the curb-and-storm read across every house in your area, automatically, so your crew spends the day on roofs that can actually buy. It pulls aerial imagery to estimate a roof-age range per address, and it models storm physics, hail and wind, on each individual roof rather than just telling you a region got hit. The output is a ranked view of a street or area: which roofs are old enough to be due, and which ones the storms actually wore out, with the recently re-roofed houses pushed down the list where they belong.

In the terms used above, RoofPredict is a scaled-up version of the two skills already covered. It estimates age (so it down-ranks the four-year-old re-roof you would otherwise knock) and it models the storm on each roof (so it separates the houses a hail cell actually battered from the ones it merely passed near). "A hail map shows where it hailed; we show which roofs it actually wore out" is the whole point, and it is the gap regional storm data leaves open.

Be clear about the honest limits, because a tool that oversells is a tool that burns trust. RoofPredict gives a roof-age range, not an exact install date, and the storm modeling gives odds, not proof, that a given roof took damage. It does not pull permits for you, it does not climb into the attic to count layers, and it does not replace the inspection that confirms a job. It is a targeting tool: it tells you which doors are worth knocking and which to skip, so the physical inspection skills in the rest of this piece get pointed at the right houses. On the storm side, the documentation it produces is exactly that, documentation that supports a homeowner's conversation with their carrier. The roofer documents conditions and writes the estimate; the insurer decides coverage; the homeowner owns the claim. RoofPredict does not file, handle, or negotiate anything, and it makes no promises about what any carrier will or will not do.

If you want a gut check, the fair test is the one any skeptic should run: hand it a few roofs on a street you already know cold, the recent re-roof at the corner, the tired original two doors down, and see whether the ranking matches what your own eyes already told you. That is the right way to trust any data: make it agree with what you can verify before you let it point your crew.

Worked example: qualifying a single street

Let us run the whole method on a hypothetical block so the workflow is concrete. Say you are looking at a 14-house street in a subdivision built in 2001. Your goal: figure out which roofs are likely replaced and which are due.

Step 1: Establish the baseline age. The subdivision is from 2001. Original roofs, if 3-tab, are roughly 24 years old today and well past end-of-life. So the default assumption for any untouched roof here is "old, due, original." That assumption is your starting bet for every house, to be overturned by evidence of replacement.

Step 2: Truck-window pass. You drive the street slowly and tally curb signals per house.

Already the street is sorting itself. Houses 1 and 3 are old. House 2 is a recent tear-off you skip. House 5 is the classic trap: it looks newish but the chunky double edge and old flashing under a fresh field say roof-over, which means it is closer to due than its color suggests and its next roof must be a full tear-off. House 4 needs a sharper tool.

Step 3: Check the paper trail on the ambiguous ones. For House 4 you pull the county permit portal. There is a re-roof permit dated 2019 with a passed final inspection. That settles it: House 4 was replaced six years ago, not due. For House 2 you find a permit from three years ago confirming your curb read. For House 5 there is no permit at all, consistent with an un-permitted overlay, and your curb read stands.

Step 4: Layer in storm history. You check the regional severe-weather record and find a significant hail event crossed this area about 18 months ago. That reframes the old roofs: Houses 1 and 3, already due by age, may also carry storm damage from that event that the owners have not addressed. Those move to the top of your list. But you also note the limit: the regional record tells you the cell crossed the area, not which of these specific roofs it actually battered. Orientation and exposure differ house to house.

Step 5: Prioritize. Your worked street now sorts cleanly:

• Knock first: Houses 1 and 3 (old by age, possible recent storm damage, owners likely unaware).
• Knock / investigate: House 5 (roof-over near end of life; honest conversation about the next roof needing a full tear-off).
• Skip: Houses 2 and 4 (confirmed recent replacements; clock reset).

That is the entire discipline in miniature. Notice how many of the 14 houses you would have wasted effort on if you had knocked them all blind, and how the recently replaced ones revealed themselves through a stack of small, cheap signals plus one permit check. Now multiply that street by a few thousand addresses, and you can see why scaling the read with aerial age estimates and per-roof storm modeling is the difference between a rep working 50 random doors and a rep working 50 doors that can say yes.

The mistakes that fool good crews

Even experienced people get these wrong. Knowing the traps is half the skill.

Mistaking a roof-over for a recent tear-off

The single most expensive misread. A layer-over can look fresh and color-rich while the system underneath is old and the whole assembly is near the end of its allowed life. Crews that only read field color write these off as "new" and miss a near-term job, or worse, a homeowner who needs to know their roof cannot be layered again. Always check the edge thickness and, when you can, count the layers. A chunky double edge plus old flashing under a fresh field is a roof-over until proven otherwise.

Trusting a clean field on a roof that was just washed or recently cleaned

A pressure-washed or recently cleaned roof can look years younger than it is for a season. Cleaning does not reset the clock. Confirm color reads against the harder evidence: brittleness up close, granule loss in the gutters, cracked boots, and the permit record. If the surface looks suspiciously fresh but the boots are cracked and the flashing is rusted, believe the boots.

Letting a partial job hide an old roof

A new slope facing the street can make a whole house read as replaced when the back two slopes are 22 years old. Always view multiple planes. The seam where new meets old is the tell. A homeowner who replaced one storm-damaged slope still has an aging roof everywhere else, and that is a legitimate conversation.

Reading year built as roof age

Covered above, but it bears repeating because it is the most common single error: the county's year-built figure is not the roof's age. Re-roofs are invisible to it. Use it as a ceiling and a context clue, never as the answer.

Treating a regional hail map as a per-roof damage list

A storm crossed the ZIP. That does not mean every roof under the polygon is damaged, and it does not mean any specific roof is. Hail and wind hit unevenly. Treating the swath as a target list wastes effort on undamaged roofs and skips damaged ones just outside the line. The fix is to model the storm on each roof, or at minimum to verify per-house with an actual inspection, rather than knocking a polygon blind.

Overstating certainty to the homeowner

You cannot read an exact install date off a roof from the ground, and you should not claim to. "This looks like it is in the 18-to-22-year range" is credible and defensible. "Your roof is exactly 19 years old" is not, and a sharp homeowner will catch you. Honest ranges build more trust than false precision, and they hold up better when the homeowner checks.

Field checklists you can use today

Two checklists. The first is the 60-second curb read. The second is the closer inspection when you have permission.

The 60-second curb qualification

• Identify the material (3-tab vs architectural vs other) and anchor the expected service life.
• Read field color and granule coverage; check the south/west slopes first.
• Scan shingle edges for curling, cupping, lifting.
• Look at downspout splash points for granule grit piles.
• Check flashing and valleys: bright vs rusted.
• Check pipe boots: supple vs cracked.
• Compare all visible planes for color match or a partial-job seam.
• Check edge thickness for a roof-over's double layer.
• Look for drip edge, ridge-cap shingles, and ridge vent (modern-replacement signature).
• Tally: three or more signals one direction = confident read; split = use a sharper tool.

The closer inspection (with permission)

• Count roofing layers at the eave or a lifted shingle (1 = tear-off/original; 2+ = roof-over).
• Test shingle flexibility and corner brittleness (brittle/cracking = old).
• Check self-seal strip bond (loose = old, wind-vulnerable).
• In the attic: look for new vs old sheathing, multiple nail-hole generations, stains, rot, daylight.
• Photograph flashing, valleys, penetrations, and any storm bruising for documentation.
• Note ventilation adequacy (poor ventilation shortens roof life and explains premature aging).
• Record findings as an age range with a condition note, not a single date.

The desk research (before or after the visit)

• Pull the building-permit history for the address; look for re-roof permits and final inspections.
• Note the year built as a context ceiling, never as roof age.
• Check the regional severe-weather record (NWS/SPC/NOAA Storm Events) for hail and wind dates.
• If using aerial imagery, compare two dated images for surface change between captures.
• Reconcile all sources; let the strongest evidence (a dated permit, a dated image change) win.

Putting it together

Telling whether a roof has been replaced is not one trick; it is a stack of cheap reads that get more definitive as you go. From the truck you read color, granules, edges, flashing, boots, plane uniformity, and edge thickness, and you sort most houses into confident bands in under a minute each. When you need more, the attic counts layers and the deck records history. When you need a date, permits and dated aerial imagery can sometimes supply one, with the standing caution that the absence of a permit proves little and the year-built figure proves nothing about the roof.

The through-line is honesty about precision. A roof is an age range, not a date, whether you read it from the curb or from a satellite. A storm map shows where it hailed, not which roofs it wore out. The roofers who win the qualification game are the ones who respect those limits, stack the signals, and point their hours at the houses the evidence says are actually due.

That last part, doing this across thousands of addresses instead of one street at a time, is where targeting tools earn their place. RoofPredict estimates a roof-age range from aerial imagery and models hail and wind on each individual roof, then ranks a street so your crew knocks the roofs that are due and skips the ones that just got replaced. It will not climb into an attic, pull a permit, or promise you what an insurer will do, and the honest read is a range and a set of odds, not a certificate. But it does at scale exactly what your trained eye does at the curb, so the inspection skills above land on the right doors. If you want to see whether the ranking matches what you already know, the fair test is to hand it a street you know by heart and check its work against your own. You can see how that looks at RoofPredict.

FAQ

Can you really tell if a roof has been replaced just by looking from the ground?

Most of the time, yes, well enough to make a targeting decision. From the curb you read field color and granule coverage, shingle-edge curling, granule grit at the downspouts, flashing and pipe-boot condition, plane uniformity, and edge thickness. No single sign is conclusive, but three or more pointing the same direction is a confident read. What you cannot do from the ground is read an exact install date, only a likely age range.

What is the difference between a roof being replaced and the clock being reset?

A full tear-off and re-roof resets the clock and gives the roof a fresh service life. But a roof-over (a new layer atop the old) only partially resets it and can be near end-of-life despite looking fresh, since the old roof and its problems are still underneath. A partial replacement resets only the slope that was redone. So a roof can be partly replaced and still be an old roof everywhere else.

Why can't I just use the year the house was built to know the roof's age?

Because the year built only tells you when the original roof went on, not whether it has been replaced since. A 1995 house may have had two roofs already. Re-roofs are invisible to county assessor and real-estate records. Treat year built as a ceiling (the roof cannot be older than the house) and a context clue, never as the roof's actual age.

How do I check building permits to confirm a roof was replaced?

Find the jurisdiction with authority (city building department inside city limits, otherwise the county), use their online permit portal or call or file a public-records request, and search by address or parcel number for a permit type containing roof, re-roof, or tear-off, with an issue date and ideally a passed final inspection. A dated re-roof permit is strong evidence of replacement. The catch: a missing permit does not prove a roof was never replaced, since many cash and storm jobs go unpermitted.

How can I tell a roof-over from a recent tear-off?

Look at the edge thickness at the rakes and eaves. A roof-over often shows an obviously taller, double-thick edge and a heavier shadow line because there are two roofs stacked there, and the field may look lumpy or uneven. Old, rusted flashing sitting under a fresh-looking field is another flag. To confirm, count the roofing layers at the eave or by lifting a shingle: two or more layers means an overlay happened.

How accurate is roof age estimated from aerial imagery?

Aerial imagery is good at sorting roofs into confident age bands and, when you can compare images from two different dates, at detecting that a roof changed (was replaced) between those captures. It cannot read an exact install date off a single undated image any more than your eyes can from the curb. The honest output is an age range, not a certificate, and it does not replace an inspection that confirms condition or counts layers.

Does a regional hail map tell me which roofs were damaged?

No. A hail or storm map shows you where a storm passed, often a county or a polygon, not which individual roofs it actually damaged. Hail and wind fall unevenly, and two houses on the same block can take very different beatings depending on slope orientation, exposure, and the storm's exact track. To know which roofs were worn out, you either model the storm per roof or verify each house with an actual inspection.

What does RoofPredict do, and what are its limits?

RoofPredict estimates a roof-age range per address from aerial imagery and models hail and wind on each individual roof, then ranks a street or area so your crew knocks the roofs that are due and skips the recently replaced ones. Its honest limits: it gives an age range, not an exact date, and storm modeling gives odds, not proof, of damage. It does not pull permits, inspect attics, count layers, or file or handle insurance claims. It is a targeting tool that points your inspections at the right doors.

How long do different roofing materials usually last?

Rough service-life ranges: 3-tab asphalt about 15 to 20 years; architectural/dimensional asphalt about 22 to 30 years; wood shake about 25 to 30 years; standing-seam metal about 40 to 70 years; clay or concrete tile 50-plus years (with underlayment lasting 20 to 30); slate 75 to 100-plus years. Real-world conditions like intense sun, hail, and poor attic ventilation shorten the top end, sometimes significantly. Identifying the material first is the biggest lever in estimating age.

What is the most common mistake roofers make when judging if a roof is new?

Trusting field color alone. A roof that was recently pressure-washed, a roof-over with a fresh top layer, or a house with one newly replaced street-facing slope can all read as new while the roof is actually old or near end-of-life. The fix is to stack signals: check brittleness and granule loss, look at all planes for a partial-job seam, check edge thickness for an overlay, and confirm with a permit or dated imagery rather than judging by color from one angle.