How to Document Brittle or Non-Repairable Shingles

Every roofer who has spent a season doing repairs knows the moment. You lift a tab to slide a new shingle under the course above, and the tab snaps off in your hand like a stale cracker. You try the next one over, same thing. The mat has gone hard, the asphalt has lost its oils, the granules are sheeting off in your palm, and the repair you quoted as a four-shingle patch has just told you it cannot be done without tearing the field apart. That roof is brittle. It is, for practical purposes, non-repairable in spots — and whether the homeowner is paying cash or filing on a storm, the thing that decides what happens next is not your opinion. It is your documentation.

Brittleness is one of the most contested, most poorly documented conditions in residential roofing. Adjusters push back on it because "old and worn out" is not a covered peril. Homeowners doubt it because the roof "looks fine from the street." And honestly, a lot of contractors document it badly — a blurry photo of a broken shingle and the word "brittle" written on a proposal does not carry weight with anyone. The roofers who win these conversations are the ones who can show, in repeatable physical tests and dated photos, that the existing shingles cannot accept a repair without further damage. That is a documentation skill, and it is learnable.

Below is the full field workflow: what brittleness actually is at the material level, the bend test and the four other tests that hold up, exactly what to photograph and how, how to write the repairability finding into an estimate without crossing any legal lines, and the specific mistakes that get good findings thrown out. The point is to make the roof speak for itself in your file so that the homeowner, the cash customer, or the carrier's adjuster reaches the same conclusion you did standing on the ridge.

What "brittle" and "non-repairable" actually mean

Start with precision, because vague language is the first thing that gets a finding dismissed. "Brittle" is not a diagnosis by itself. It is a description of a measurable physical state, and "non-repairable" is a conclusion you draw from that state plus the repair you were trying to perform.

An asphalt shingle is a sandwich: a fiberglass or organic mat saturated and coated with asphalt, surfaced with mineral granules, with a strip of factory sealant on the back. When that shingle is healthy, the asphalt coating is pliable enough that you can flex a tab a few degrees and it returns. The sealant strip, when warmed by the sun, bonds each course to the one below so the field acts like a continuous membrane.

Three things degrade over time and turn a shingle brittle:

1. Oil loss in the asphalt. Ultraviolet exposure and heat cycling drive off the volatile oils that keep the coating flexible. As the asphalt hardens, the shingle loses its ability to flex without fracturing. This is the core of brittleness.
2. Mat embrittlement. The mat carries the tensile strength. Once the surrounding asphalt can no longer protect and flex with it, the mat itself fractures under any handling load.
3. Sealant failure. The factory adhesive loses tack with age and heat. Tabs that no longer seal lift in wind and, critically for repairs, will not re-bond after you disturb them.

"Non-repairable" is the operational consequence. A repair on an asphalt roof requires three physical things to happen: you have to lift the course above the damaged shingle to access the nail line, you have to remove the damaged shingle, and you have to seat and seal the replacement so it bonds into the surrounding field. On a brittle roof, lifting the course above breaks the course above. Removing one shingle cracks its neighbors. And the new shingle, even installed perfectly, sits in a field that will not seal to it because the surrounding sealant is dead. You cannot make a localized, watertight, wind-rated repair without damaging good material — which means the "repair" keeps growing until it is a section replacement or a slope.

That distinction matters legally and commercially. You are not documenting that the roof is old. Age is not a covered peril and "the roof is 22 years old" is not a finding anyone has to act on. You are documenting that the existing shingles will not accept a repair without sustaining additional damage — a present, physical, demonstrable condition. Keep that sentence in your head; it is the spine of every photo and note below.

Brittle vs. weathered vs. defective — don't blur them

Three conditions get lumped together and they are not the same. Documenting the wrong one undercuts your credibility.

A storm claim is decided on storm-caused damage. Brittleness is relevant to a storm claim only on the repairability question: once a covered peril damages some shingles, can those specific damaged shingles be replaced without damaging the rest? That is a legitimate, well-established part of a repair-versus-replace estimate. Keep brittleness in its lane — as the reason a spot repair of storm damage is not feasible — and your file stays clean.

The five tests that hold up (and how to run them)

You need physical evidence, not adjectives. These are the field tests that produce documentable, defensible results. Run them, photograph them as you go, and narrate what you are doing. A test the file can't see did not happen.

1. The bend test (the bread and butter)

This is the single most persuasive demonstration of brittleness, because it isolates the exact property at issue: the shingle's ability to flex without fracturing.

How to run it:

• Pick a representative shingle on the slope you are scoping — not the worst one you can find, a representative one. (Documenting your worst example as if it were typical is how a finding gets called cherry-picked. More on that later.)
• Where you can, use a loose tab or the rake/eave edge so you are testing in place. If you must remove a piece, note that you removed it and from where.
• Flex the tab upward to roughly the bend it would have to take during a normal repair — the flex needed to slide it out or seat a replacement under the course above. You are not trying to fold it in half; you are replicating repair handling.
• Document what happens: does it flex and return, does it crack at the bend, or does it snap clean off?

A healthy shingle bends and recovers. A brittle one cracks audibly and fails at the bend line. Video is far stronger than a photo here — a five-second clip of the tab cracking under controlled bend is close to irrefutable, where a still of a broken shingle could be anything. Narrate it on the clip: address, slope, "bending to repair flex, fractures on contact."

2. The lift / unseal test

This tests the actual first step of any repair: can you lift the course above the target shingle to reach the nails?

• Choose a shingle you would have to lift to perform a repair below it.
• Break the sealant bond and lift the tab the way you would to expose the nail line of the course beneath.
• Document the result. On a repairable roof, the tab lifts, the sealant releases, and the shingle stays intact. On a brittle roof, the tab either tears at the sealant line, fractures across the tab, or breaks the adjacent shingle as it lifts.

This test is powerful because it maps one-to-one onto repair mechanics. You are not making an abstract argument about age; you are showing that the literal first motion of a repair destroys good material.

3. The sealant / re-bond test

Even if you got a shingle out, the repair has to seal back into the field. This test shows whether it can.

• Find a tab that should be sealed and check whether the factory adhesive still has tack. Press a lifted tab back down on a warm day and see if it grabs.
• Document tabs that lift with no resistance and do not re-bond. Note the surface temperature and time of day, because sealant behavior is temperature-dependent and you want to preempt the "it just needs to warm up and re-seal" rebuttal.

A field where the sealant is uniformly dead means a replacement shingle will not bond to its neighbors, leaving a wind-vulnerable patch. That is a real, documentable reason a spot repair is not durable.

4. The granule and mat exposure check

This is supporting context, not the headline. Heavy granule loss exposes the asphalt to more UV, accelerating the embrittlement you are already demonstrating.

• Photograph granule accumulation in valleys, at downspout outlets, and at gutter aprons.
• Photograph bare asphalt areas where the coating is fully exposed.
• Where the mat is visible through worn coating, document it — that is end-of-coating-life made literal.

Tie it back: granule loss is why the asphalt has hardened, which is why the bend test fails. It is the mechanism, and it makes the bend-test result make sense to a non-roofer reading the file.

5. The thermal-shock / handling check

Brittleness is worst when shingles are cold, and most repairs are not done in July. If you are inspecting in cold weather, document it, because cold makes the non-repairability case stronger and an adjuster who inspects on a warm afternoon will see a more forgiving roof.

• Note ambient and surface temperature.
• Document that cold-weather handling fractures shingles that might survive warm handling — and that real repairs happen across all seasons, including cold ones.

You are not faking a condition; you are making the record honest about the conditions a repair would actually face.

A note on the "torch test" and other gimmicks

You will see videos of contractors taking a torch or heat gun to a shingle to show it crumbles, or soaking and freezing a sample. Skip the theater. Anything that adds heat, flame, or artificial stress you did not encounter in the normal course of a repair invites the accusation that you damaged the shingle to manufacture a finding. Stick to tests that replicate ordinary repair handling. The bend test and lift test win because they are exactly what a real repair would do — nothing more.

Reading the results: what counts as a pass and a fail

A test is only useful if you know what outcome means what. Roofers run the bend test all the time and then describe the result so loosely that the file doesn't carry a clear conclusion. Tighten it up.

For the bend test, there are really three outcomes, and you should be able to say which one you got:

• Flexes and recovers. The tab bends to repair flex and springs back with no cracking. This shingle is repairable. Honest documentation includes the roofs that pass — if you only ever document failures, your files start to look like a foregone conclusion.
• Flexes, then surface-cracks. The coating crazes or the surface cracks but the mat holds. This is a roof on the edge. Note it as marginal and lean on the other tests (sealant, lift) to decide whether a durable repair is realistic.
• Fractures or snaps. The tab cracks across the bend or breaks off. This is the clear non-repairable result and the one your video should capture cleanly.

For the lift test, the question is binary in practice: did lifting the course above to reach the nail line destroy material that a repair needs to keep intact? If the course above tore, fractured, or cracked its neighbors, the answer is yes and the spot repair is not feasible. If it lifted clean and reseated, the roof can take a repair and you should document that honestly.

For the sealant test, you're looking for whether a replacement will bond into the field. Tabs that lift with no resistance and refuse to re-tack on a warm surface mean a patch will sit unsealed and wind-vulnerable. Note the surface temperature every time, because the single most common rebuttal to a dead-sealant finding is "it'll re-seal once it's warm" — and a record showing the sealant was dead at 90 degrees ends that argument.

How many samples, and from where

One test on one shingle is an anecdote. A defensible file shows a pattern. A reasonable practice for a typical slope:

• Run the bend test in several spread-out locations on each slope you're scoping — not clustered in one bad patch. Three to five representative points on a slope gives you a pattern rather than a single data point.
• If a slope passes where another fails, document each slope on its own. A roof is not uniform; a shaded north slope can be a decade behind a south slope in the same field. Saying "the front slope fails, the back slope is marginal, here's each" is more credible than a blanket claim, and it's the truth.
• Tie sample locations to your slope overview photos so anyone can see your tests were spread across the slope, not picked from the worst corner.

Your documentation kit and the photo workflow

Good documentation is not about owning expensive gear. It is about a repeatable sequence so nothing gets missed and every photo is provable.

The kit

• A phone or camera that timestamps and, ideally, geotags. Confirm the date and time are set correctly before you climb — a wrong timestamp poisons the whole file.
• Chalk (for marking, and for hail test squares if there is a storm component).
• A tape measure or a six-inch scale card for every close-up, so size is provable.
• A soft brush or your hand for clearing granules off a test area.
• Something for a wide-shot reference: a placard or even a piece of paper with the address and date, photographed once per slope.
• A way to shoot short video clips. Your phone is fine.

The photo sequence that survives scrutiny

Shoot in this order, every roof, so the file tells a story from the curb to the cracked tab:

1. Establishing shots. The full elevation from the ground, four sides. These prove which house and let anyone orient the close-ups.
2. Slope overviews. Each slope from the roof, wide enough to show overall condition and to place your test locations.
3. Context-to-close progression. For every defect, shoot a sequence: a wide shot showing where on the slope, a medium shot, then the macro. A lone macro of a cracked shingle with no context is the photo adjusters and homeowners distrust most, because it could be from any roof, any year.
4. Scale in the close-ups. Tape or scale card in frame on the macros.
5. The tests, as video. Bend test, lift test, sealant test — captured live, narrated.
6. Collateral and corroboration. If there is a storm element, document soft-metal damage (gutters, vents, flashing, downspouts, AC fins, any wood you can reach) the same way. Hail does not read on shingles alone.
7. A continuity frame. Where practical, one shot that ties a close-up back to its slope — same chimney, same vent in both frames — so no one can claim the macro came from somewhere else.

Why metadata is your friend

Date, time, and location embedded in the file do quiet, heavy work. They establish when you documented the condition, which matters if a storm date is in play, and they establish that the photos are from this roof, this visit. Do not strip metadata when you export. If you annotate photos, keep an unannotated original — a clean copy plus a clearly-labeled marked-up copy is far stronger than a single edited image that looks like it could hide something.

Video beats stills for the things that move

A still photo proves a state. A video proves a behavior. Brittleness is a behavior — it is what the shingle does when you handle it. The clips that win these conversations are short, narrated, and continuous: you say the address and slope, you bend the tab, it cracks, you keep rolling. No cuts. A continuous clip can't be accused of hiding the part where you pre-broke the shingle.

Writing the repairability finding into the estimate

Here is where most of the legal exposure lives, and where careful wording protects you. Your estimate documents a physical condition and prices a scope of work. It does not interpret anyone's insurance policy, promise any outcome, or argue the homeowner's case to a carrier. Stay on the document-and-estimate side of the line and you are doing exactly what a roofer is supposed to do.

What an estimate is allowed to say

A roofing contractor may inspect a roof, document its condition, and prepare an accurate estimate to repair or replace it. You can state facts about the roof and facts about your scope. You can write, plainly, that the existing shingles are brittle and will not accept a repair without sustaining further damage, and you can price the scope that condition requires.

What you do not do — for a fee or otherwise as part of representing the homeowner — is negotiate, adjust, or "handle" their insurance claim, interpret what their policy covers, promise a specific payout or approval, say anything about their deductible being waived or absorbed, advertise a "free roof," or position yourself as their representative against the insurer. That last category is unlicensed public adjusting in most states, and even calling yourself an "insurance claim specialist" has gotten roofers in legal trouble. Your job is to document thoroughly and estimate accurately, then hand that to the homeowner. The homeowner files. The insurer decides coverage. You stay in your lane and your documentation does the talking.

The structure of a defensible repairability finding

Write it as a short, factual narrative attached to your line items. A workable template:

Existing roof condition: The existing shingles on the [front/rear/etc.] slopes are an asphalt three-tab/architectural shingle showing [granule loss exposing asphalt coating / mat exposure / loss of flexibility]. On hands-on inspection performed [date], representative shingles fractured when flexed to the degree required for repair handling (see bend-test video, file refs ___) and adjacent shingles fractured when the course above was lifted to access the nail line (see lift-test photos/video, file refs ___). Factory sealant on the [slope] tested is no longer bonding (see ___).

Repairability finding: Because the existing shingles fracture on the handling a repair requires, individual damaged shingles cannot be replaced without damaging surrounding shingles. A localized repair is therefore not feasible without progressive damage to adjacent material. The scope below reflects [section/slope/full] replacement as the minimum feasible repair.

Notice what that does and does not do. It states what you observed, with file references. It draws a repairability conclusion from physical facts. It scopes the minimum work the condition requires. It says nothing about coverage, nothing about a deductible, nothing about what the carrier will or should pay. If there is a storm involved, you document the storm damage separately and let the repairability finding explain why the storm-damaged shingles can't be patched — but you never tell anyone the claim will be approved.

Tying the scope to the finding

Your line items should map cleanly to the finding so the logic is visible:

• If the finding is that one slope can't be spot-repaired, the scope is that slope (plus the standard accessory items a slope replacement actually needs).
• If you're matching into adjacent slopes for a wind-rated, watertight result and the existing product is discontinued or won't blend, document the match problem as its own finding with its own photos — discontinued color, profile mismatch, exposure difference — because non-matchability is a separate, real reason a partial repair fails to produce a sound roof.
• Price what the work requires by code and by manufacturer instructions, and keep your unit logic consistent from estimate to estimate. Inconsistent pricing across similar roofs is the fastest way to look like you're scoping to a number instead of to the roof.

The do-not-say list, kept on a card

Print this and keep it where you write estimates. None of these belong in your documents, your texts, or your front-door pitch:

• "We'll get your claim approved." (You don't decide coverage.)
• "We'll handle / manage / negotiate the claim for you." (Public adjusting.)
• "We'll maximize your settlement" / "fight the adjuster" / "recover every dollar." (Adjusting / advocacy.)
• "Your deductible is covered / waived / absorbed / we'll eat it." (Insurance fraud in many states.)
• "Free roof." (Implies the deductible disappears; same problem.)
• "We're insurance claim specialists." (Self-labeling that has triggered UPPA findings.)

What you can say is simpler and truer: "We'll document the roof's condition thoroughly and give you an accurate estimate. You file with your carrier; they decide coverage. We'll repair our scope to code." That sentence keeps you compliant and still does everything a homeowner needs from you.

A worked example: the 14-shingle "repair" that wasn't

Concrete beats abstract. Here is how the whole workflow plays out on a real-shaped job.

A homeowner calls about a few wind-lifted shingles on the front slope after a storm came through. On the phone it sounds like a fourteen-shingle repair — replace the lifted ones, reseal, done. You quote nothing on the phone. You go look.

On the roof, the front (south-facing, full-sun) slope is a 20-year three-tab with heavy granule loss; granules are piled two inches deep at both downspout outlets. You run the sequence:

• Establishing + slope shots: four elevations from the ground, the front slope wide from the ridge.
• Bend test, representative tab, mid-slope: you flex a tab to repair flex on a 41°F morning. It cracks across the bend on the first try. Five-second narrated clip: address, "front slope, bending to repair flex," crack.
• Lift test on a storm-damaged shingle: you lift the course above one of the actually-lifted shingles to reach its nails. The tab you're lifting tears at the sealant line and the shingle to its right cracks. Photo sequence: wide (which shingle), medium, macro with the tape showing the fracture, plus a clip of the lift.
• Sealant test: three tabs nearby lift with zero resistance and won't re-bond when pressed back down. You note surface temp and time.
• Storm collateral: you document dented gutter aprons and a bent vent cap from the same event, since the homeowner raised a storm.

The math is now obvious in the file. To replace each of the fourteen storm-damaged shingles you have to lift the course above, and that course fractures, which means each fourteen-shingle "repair" cascades into its neighbors. The slope's sealant is dead, so any replacement won't bond. Your repairability finding writes itself: individual shingles can't be replaced without progressive damage; the minimum feasible repair on the front slope is a slope replacement, and you document the match problem if the front slope can't blend to the others.

You hand the homeowner three things: the photo-and-video file, a plain-English explanation of why the patch they imagined isn't possible, and an accurate estimate for the slope (or roof) the condition requires. Then you tell them the truth about the rest: "File with your carrier. They'll send someone out and decide what they cover. I've documented everything so the facts are all there. What they pay is between you and them." You priced the work. You documented the roof. You did not handle a claim. That is the whole job, done right.

Anticipating the rebuttals before you write the file

The point of documentation is to settle questions before they're asked. If you know the four or five pushbacks a brittleness finding draws, you can document against each one while you're still on the roof. Build the file like you're going to have to defend it, because sometimes you will.

"That's age and wear, not storm damage." This is the big one when a storm is involved, and it's a fair question — age and wear are not a covered peril. The answer is to keep brittleness in its lane. You are not claiming the brittleness is storm damage. You are documenting storm damage separately (the actual lifted, bruised, or punctured shingles, plus soft-metal collateral) and using the brittleness only to explain why those storm-damaged shingles can't be spot-replaced. Two findings, cleanly separated: here's the storm damage, and here's why a patch of that damage isn't feasible.

"It'll re-seal when it warms up." Pre-empted by recording surface temperature on your sealant test. If the sealant was dead at high surface temperature, warmth isn't the issue. If you tested cold, go back warm and re-test rather than leaving the rebuttal open.

"You damaged that shingle yourself." This is why continuous, narrated video of the test matters more than a still of a broken shingle, and why you never use heat, flame, or any stress beyond ordinary repair handling. The clip shows the shingle whole, then shows it failing under the exact handling a repair requires, with no cut where you could have pre-broken it.

"You only photographed the worst spot." Defeated by spread-out samples and slope-by-slope documentation, and by saying in writing that the shingle shown is representative and giving the sample locations.

"A repair is cheaper, just do that." This is where the lift test and the match documentation do their work. The honest answer is that a "repair" on this roof isn't actually a repair — lifting the course above breaks it, the patch won't seal, and a discontinued product won't blend — so the cheap repair doesn't produce a sound, watertight, wind-rated roof. You're not upselling; you're documenting that the smaller scope doesn't restore the roof.

Writing with these in mind changes what you shoot. You'll capture the temperature, spread your samples, roll continuous video, separate storm damage from wear, and document the match problem — not because someone told you to, but because you already know which questions are coming.

What credible documentation looks like to an outside reviewer

It helps to step out of the roofer's seat for a minute and look at your file the way a homeowner's spouse, a skeptical adjuster, or a reviewer who never saw the roof will look at it. They are pattern-matching for trust signals. The files they believe share a few traits, and you can build all of them in deliberately.

• It's internally consistent. The wide shot, the medium shot, and the macro obviously come from the same roof — same chimney, same vent, same ridge line visible across frames. Nothing feels imported.
• It shows behavior, not only states. Video of the shingle failing under handling beats a gallery of broken shingles, because behavior is hard to fake and a state is easy to dispute.
• It includes passes, not only failures. A file that documents the marginal back slope honestly reads as an assessment. A file that's nothing but worst-case macros reads as a sales pitch.
• It has scale and metadata. Tape in the close-ups, correct dates and locations baked into the files, originals preserved. The absence of these is what makes a file feel like it's hiding something.
• It separates findings. Storm damage here, brittleness there, match problem over there — each with its own evidence. Blurring them together is what lets someone dismiss the whole thing.
• It stops at the roof. No coverage opinions, no deductible talk, no promises about the claim. A clean condition-and-estimate file with no advocacy language is one nobody can accuse of overreach.

If you build for that reviewer — even when there is no claim and the customer is paying cash — your documentation does the persuading for you, and you spend less time arguing and more time roofing.

The mistakes that get good findings thrown out

Most contractors who lose the brittleness argument don't lose it on the roof. They lose it in the file. Here are the errors that show up over and over.

Cherry-picking the worst shingle. You find the single most degraded tab on the roof, snap it in half on camera, and present it as the condition of the field. The moment anyone realizes the rest of the slope is in better shape, your whole file is suspect. Document a representative shingle and say explicitly that it's representative. If conditions vary across slopes, document each slope separately rather than letting your worst example stand in for everything.

Pre-breaking and presenting the result. A macro of an already-broken shingle proves nothing about why it's broken or who broke it. Continuous, narrated video of the break happening under normal repair handling is the fix. If your evidence is a still of a fractured shingle with no clip of the fracture, assume it carries little weight.

No scale, no context. A close-up with nothing in frame for size, and no wider shot tying it to a slope, is the single most distrusted kind of roofing photo. Always: wide, medium, macro-with-scale, plus a continuity frame.

Confusing "old" with "non-repairable." Writing "roof is 22 years old, recommend replacement" is not a repairability finding and gives an adjuster an easy "age and wear aren't covered." Age is context. The finding is that the shingles fracture on repair handling — a present physical condition you demonstrated. Lead with the demonstration, not the date.

Adjective-only documentation. "Brittle, dried out, shot." Those are conclusions with no evidence under them. Every adjective needs a test or a photo behind it or it's just your opinion against theirs.

Stripped or edited metadata. Exporting through an app that strips EXIF, or sending only annotated images, makes a clean file look like it has something to hide. Keep originals with metadata intact and label any markups clearly.

Crossing the legal line in writing. The fastest way to turn a strong condition file into a liability is to staple claims language to it — "we'll get this approved," anything about the deductible, anything that reads as you representing the homeowner against the carrier. Document the roof. Estimate the work. Stop there.

Inconsistent scope logic. If the same condition produces a slope replacement on one house and a full tear-off on the next with no documented reason, you look like you're scoping to a payout, not to a roof. Tie every scope decision to a documented finding and keep your unit pricing consistent.

Building a repeatable system so it happens every time

Documentation quality collapses when it depends on whoever happened to climb the roof that day. The roofers whose files consistently hold up have turned this into a checklist their whole crew runs.

The brittleness documentation checklist

Hand this to every estimator and inspector:

• Confirm device date/time correct before climbing.
• Four ground elevations.
• Wide shot of each slope from the roof.
• Bend test on a representative shingle per slope — video, narrated with address/slope.
• Lift test on a shingle a repair would require lifting — video + photo sequence.
• Sealant re-bond test — photo + note on surface temp and time.
• Granule-loss documentation (valleys, downspout outlets, gutters).
• Mat-exposure shots where present.
• For every defect: wide → medium → macro-with-scale → continuity frame.
• Storm collateral (soft metals) if a storm is in play, documented same way.
• Match problem documented if partial replacement won't blend (discontinued color/profile/exposure).
• Note ambient + surface temperature for the visit.
• Repairability finding written from the physical facts, with file references.
• No claims/coverage/deductible language anywhere in the file.

Standardize the narration

Train the crew to say the same three things on every test clip: the address, the slope, and what they're doing ("bending to repair flex"). It feels repetitive on the roof and it makes the file bulletproof off the roof. Consistent narration across hundreds of clips also makes your company's documentation recognizable — anyone who sees one of your files knows what to expect from all of them.

Build the photo file the same way every time

Organize by address, then by slope, then context-to-close within each defect. A reviewer who can find the wide shot, the test video, and the macro for any given finding in ten seconds trusts the file more than one who has to dig through 90 unsorted images. Naming and folder discipline is unglamorous and it's worth more than a better camera.

Train the conversation, not only the climb

The strongest documentation still fails if the person who took it can't explain it at the kitchen table without overpromising. Teach reps to walk a homeowner through the bend-test clip — "watch what happens when I flex this the way a repair would" — and to say the compliant version of the claims conversation by reflex: you file, they decide, we documented everything, we'll fix our scope to code. The documentation and the script have to match, or a rep undoes a great file with one sentence about getting the claim approved.

Where targeting the right roofs comes in

Everything above is what you do once you're standing on the roof. The harder business problem is which roofs to climb in the first place. Brittle, non-repairable roofs aren't randomly distributed — they cluster by age and by what the weather has actually done to them. A south-facing 20-year three-tab in a sun-baked subdivision is far likelier to fail the bend test than a 9-year architectural roof two streets over, and a slope that's eaten several hail and high-wind events has lost flexibility faster than an identical roof that the storms missed. If you could see, before you ever knock, which streets are full of roofs old enough and weathered enough to be at the brittle stage, you'd spend your inspection time where the documentation work actually turns into jobs — and skip the doors where the shingles will just bend and recover.

That's the gap RoofPredict fills. It scores the roofs in an area by roof age — as a range, estimated from aerial imagery, not an exact install date — and by the storms each individual roof has actually taken, modeling hail and wind impact house by house rather than just showing where a storm passed. It can also enrich a list you already own — your old estimates, your past customers, a mailing list — by appending those roof-age and storm signals to each address, so you can see which of your contacts are sitting on roofs that have aged into the brittle, repair-resistant zone. The honest limits matter and we say them plainly: an age range is a range, not a birth certificate, and a storm model gives you the odds a roof was worn out, not proof — the bend test on the ridge is still what proves it. What the data does is point your trucks at the right streets and the right addresses so more of your inspections land on roofs that genuinely need the work, instead of burning a morning on roofs that don't.

Used that way, the targeting and the documentation reinforce each other. You knock the doors most likely to have a non-repairable roof, you run the five tests, you build the file, and you hand the homeowner an honest estimate — and a far higher share of your climbs end in a real job, because you started by aiming at the roofs the age and the weather already wore out.

Edge cases worth knowing

Architectural vs. three-tab. Laminated architectural shingles are heavier and the bottom lamination can stay intact while the top fractures, so the bend test reads a little differently — flex the exposed laminate and watch for the top layer cracking and granule release. Three-tabs tend to fail cleaner and more dramatically. Document the product type so your test is read in the right context.

Cold-weather inspections. Cold makes everything more brittle, which strengthens the non-repairability case but invites the rebuttal that the roof would be fine in summer. Pre-empt it: note the temperature, and note that real repairs happen in cold months too, so a roof that can't be handled at 40 degrees can't be reliably repaired in winter.

Hot-weather inspections. The opposite problem. On a 95-degree afternoon, dead sealant may temporarily re-tack and a marginal shingle may flex further than it would cold. If a roof fails the tests even when warm, that's a strong file. Note the temperature so the conditions are on record.

Discontinued or unmatchable product. Sometimes the shingles would technically take a repair, but the product is discontinued or the color/profile won't blend, so a partial repair produces a visibly patched, non-uniform, sometimes wind-mismatched roof. That's a separate, legitimate finding — document the match problem with its own photos and treat non-matchability as its own reason a partial repair fails to restore a sound roof.

Layovers (two-layer roofs). A second layer over an old roof complicates repairs and brittleness reads across both. Document the layers and note how that affects repair feasibility — you often can't do a clean partial on a layover even before brittleness enters the picture.

Tile, slate, metal, wood. This whole workflow is written for asphalt shingles, where brittleness and repair handling are the core issue. Tile, slate, metal, and wood each have their own failure and repairability tests — don't borrow the asphalt bend test for them. The documentation principles (representative samples, context-to-close photos, video of behavior, scale, clean metadata, no claims language) carry across; the specific tests don't.

Putting it all together

Brittleness is a real, physical, demonstrable condition, and "non-repairable" is the honest conclusion when the existing shingles fracture on the handling a repair requires. The contractors who win these conversations — with cash customers and with adjusters — aren't the ones with the strongest opinions. They're the ones whose files make the roof prove the point: a representative bend test on video, a lift test that shows the first move of a repair breaking good material, a dead sealant test, the granule and mat context that explains why, and a photo sequence that ties every macro back to a real slope on a real house with the date baked in.

Write the repairability finding from those facts, scope the minimum work the condition requires, and stop at the line — document and estimate, never handle the claim, never touch the deductible, never promise an outcome. The homeowner files; the carrier decides; you fixed your scope to code and gave everyone the facts.

Do that consistently, point your trucks at the roofs the age and the weather already wore out, and the documentation stops being a chore you do to defend a job. It becomes the thing that earns the job in the first place.

FAQ

How do I document that a shingle is brittle and non-repairable rather than just old?

Demonstrate the physical behavior on the roof, don't assert the age. Run a bend test on a representative shingle and capture continuous, narrated video of it fracturing when flexed to the degree a repair requires. Add a lift test showing the course above breaking when you reach the nail line, and a sealant test showing tabs that won't re-bond. "22 years old" is context; "the shingles fracture on the handling a repair requires" is the finding, and the difference is what holds up with a skeptical homeowner or adjuster.

What is the bend test and how do I run it correctly?

Pick a representative shingle (not your single worst one), use a loose tab or the rake/eave edge where you can, and flex it to roughly the bend a repair would require to slide it out or seat a replacement. A healthy shingle flexes and recovers; a brittle one cracks at the bend or snaps off. Capture it on continuous video and narrate the address and slope. Run it in several spread-out spots per slope so you're showing a pattern, not one bad shingle.

Is a torch test or heat-gun test a good way to prove brittleness?

No. Anything that adds heat, flame, or stress beyond ordinary repair handling invites the accusation that you damaged the shingle to manufacture a finding. Stick to tests that replicate what a real repair does to the roof — the bend test and the lift test — because those are exactly the handling a shingle would face during an actual repair, which makes the result defensible.

Can I tell a homeowner their insurance will cover a brittle, non-repairable roof?

No. You document the roof's condition and write an accurate estimate; you do not interpret their policy, promise coverage or approval, say anything about their deductible, or handle/negotiate the claim — that crosses into unlicensed public adjusting in most states. The compliant message is simple: "We documented everything and gave you an accurate estimate. You file with your carrier, they decide coverage, and we'll repair our scope to code."

How many shingles should I test to make the documentation defensible?

Run the bend test in several spread-out locations on each slope you're scoping — three to five representative points gives you a pattern instead of an anecdote. Document each slope separately when conditions differ, since a shaded north slope can be a decade behind a sun-baked south slope. Tie your test locations to your slope-overview photos so it's clear the samples were spread across the slope, not picked from the worst corner.

Why does video work better than photos for documenting brittleness?

Brittleness is a behavior — it's what the shingle does when you handle it. A still photo proves a state and can be disputed ("how do we know you didn't pre-break it?"). Continuous, narrated video shows the shingle whole, then failing under the exact handling a repair requires, with no cut where it could have been damaged in advance. That's close to irrefutable where a lone macro of a broken shingle carries little weight.

How do I write the repairability finding into my estimate?

Write a short factual narrative attached to your line items: describe the existing shingle and condition, state that on hands-on inspection on a given date representative shingles fractured when flexed to repair handling (with file references to your test videos), and conclude that individual damaged shingles can't be replaced without progressive damage, so the minimum feasible repair is the section/slope/roof scoped below. Reference your photos, scope the work the condition requires, and include no coverage, deductible, or claim-handling language.

What's the difference between brittle, weathered, and a manufacturing defect?

Brittle means the shingle fractures on the handling a repair requires — shown with bend and lift tests, and it supports a repairability finding. Weathered means granule loss, exposed asphalt, or curling, which is useful-life context but not a peril by itself. A manufacturing defect is premature product failure (like blistering or delamination) shown by a pattern across the field, which points to a manufacturer or warranty conversation. Documenting the wrong one undercuts your credibility, so keep them separate.

Does cold or hot weather change how I should document brittleness?

Yes. Cold makes shingles more brittle, which strengthens the non-repairability case but invites the rebuttal that the roof would be fine in summer — so note the temperature and point out that real repairs happen in cold months too. On a hot day, dead sealant can temporarily re-tack and a marginal shingle may flex further, so a roof that still fails the tests when warm is a strong file. Record ambient and surface temperature on every visit either way.

What if the shingles could technically be repaired but the product is discontinued?

That's a separate, legitimate finding called a match problem. If the existing color, profile, or exposure is discontinued or won't blend, a partial repair produces a visibly patched, non-uniform, sometimes wind-mismatched roof rather than a sound one. Document the match problem with its own photos and treat non-matchability as its own reason a partial repair fails to restore a sound roof, independent of whether the shingles are brittle.