How to Turn Roofing Job Photos Into Structured Claim Data

A roof inspection generates between 40 and 200 photos. Most of them die in a phone's camera roll. Six weeks later somebody on your team is squinting at a blurry shot of a shingle, trying to remember which slope it came from, which address it belonged to, and whether that dark streak is hail bruising or a shadow. That is the gap between taking pictures and producing documentation. The first is free and worthless. The second is the entire foundation of a clean repair estimate that holds up when an adjuster, a homeowner, or your own production manager looks at it.

Turning job photos into structured claim data is a process problem, not a camera problem. The phone in your installer's pocket already shoots 12 megapixels and stamps every frame with a timestamp and GPS coordinates. The failure is never resolution. It is that the photos arrive as an undifferentiated pile with no slope, no orientation, no measurement reference, no defect classification, and no link to a line item. "Structured" means each photo can answer four questions on its own: Where on the roof is this? What am I looking at? How big or how many? What repair line does it justify? When every photo answers those four questions, you have data. When it doesn't, you have a screensaver.

What follows is the workflow my best crews actually run, the field naming system that survives contact with a wet roof and cold hands, the metadata that does work for you automatically, and the exact way photos map onto an estimate. It also covers the compliance line that keeps you on the right side of public-adjusting law, because the fastest way to turn good documentation into a liability is to use it to do something you are not licensed to do. We'll stay strictly on the side you are allowed to occupy: documenting thoroughly, writing an accurate scope, and handing it to the homeowner.

What "structured claim data" actually means

Let's define the target before we build toward it. A claim file that an adjuster can process without three rounds of follow-up, and that your office can turn into an estimate in an hour instead of a day, has a specific shape. Strip away the software and the buzzwords and it comes down to five layers of structure sitting on top of raw pixels.

Layer 1 — Identity. Every photo is tied to a single property: address, claim or job number, and date of loss if a storm is involved. This sounds trivial until you run two inspections in the same subdivision on the same morning and the photos interleave in the camera roll.

Layer 2 — Location on the roof. Each photo is assigned to a slope (front, rear, left, right, or a labeled facet on a complex roof), and ideally to a feature within that slope (field, ridge, hip, valley, eave, rake, penetration, flashing). Without slope mapping, a pile of damage photos can't be tied to the specific facets you're scoping for replacement.

Layer 3 — Subject classification. The photo is labeled by what it documents: overview, test square, hail strike, mechanical damage, wind/creased shingle, manufacturing defect, prior repair, code item, interior leak evidence, or accessory (vent, pipe boot, drip edge). This is the layer that lets you count and total.

Layer 4 — Measurement reference. Damage photos include scale (a chalk circle, a tape, a coin, a hail gauge, a square outline) so the size and density are objective rather than "trust me." Test-square photos show a defined area, usually a 10-by-10-foot square, with strikes marked and counted inside it.

Layer 5 — Estimate linkage. Each documented condition maps to a line item: the slope's square footage, the accessory count, the flashing linear feet, the code-driven items (ice-and-water shield, drip edge, ventilation) that the repair requires. This is where photos stop being evidence and become the basis for a number.

When people say a claim got "kicked back," almost always one of these layers was missing. The hail was real but there was no scale, so density couldn't be assessed. The damage was on the front slope but no photo proved which slope. The test square showed eight strikes but there was no overview to prove the square was representative. Structure is what turns a true story into a provable one.

Why this matters even when there's no storm

Not every roof job is a storm claim, and the legal cautions later in this piece apply specifically to insurance work. But the structured-data discipline pays off on every job type. A retail re-roof with clean before/during/after documentation cuts callbacks, settles disputes, and protects your warranty position. A repair you did under contract that later gets blamed for an unrelated leak is defended in minutes if you have dated, slope-labeled photos of the original condition. The workflow is the same. The storm context just raises the stakes and adds rules.

The field-capture system: shoot like you're building a file, not a memory

The single biggest leverage point is the moment of capture. Everything downstream is cheaper if the field tech shoots in a deliberate sequence with a labeling convention, instead of wandering the roof tapping the shutter. You cannot fix bad capture in the office; you can only re-climb the roof, which costs you a truck roll and your credibility.

The standard photo set every roof should get

Build a fixed sequence so nothing is forgotten and the order itself encodes meaning. My crews shoot this set, in this order, on every inspection. The consistency is the point: when every file follows the same order, anyone in the office can navigate it blind.

1. Address proof. The house number on the door or curb, and a street-level front elevation showing the whole house. This anchors identity and proves you were at the right property.
2. Four elevations. Front, right, left, rear of the structure from the ground. These establish stories, complexity, and the slope-naming reference you'll use for the rest of the set.
3. Roof overviews per slope. A wide shot of each slope from the eave looking up the plane. These are your "establishing shots" that prove which slope every close-up belongs to.
4. Accessory inventory. Each penetration and accessory: pipe boots, furnace/B-vent flashings, turtle vents, ridge vent, power vents, satellite mounts, chimney, skylights, valleys, step flashing, drip edge, gutters. One clear shot each. This inventory becomes your accessory line-item count.
5. Test squares. A defined 10-by-10 area on each slope (or each slope direction if the roof is large), with hail strikes circled and counted inside it. More on this below.
6. Individual damage close-ups. Each significant strike, crease, or defect with a scale reference, plus a context shot showing where on the slope it sits.
7. Edge and detail conditions. Flashing condition, sealant failure, deck exposure, prior repairs, layers at a cut-up.
8. Code and ventilation evidence. Existing ventilation type and count, existing underlayment/ice barrier at a cut-up if accessible, existing drip edge or lack of it.
9. Interior/attic if relevant. Active leak staining, daylight at penetrations, decking condition, insulation water marks. Only where access and safety allow.

That's roughly 40 to 90 frames on a simple roof. The discipline isn't shooting more; it's shooting the same set every time so the structure is born at capture.

The two-shot rule for every defect

The most common rookie error is the orphan close-up: a perfect macro of a single hail bruise with zero context. Nobody can tell which slope it's on or whether it's one of forty or the only one on the roof. Fix it with a rule every tech can remember: every defect gets two shots — a context shot and a detail shot. The context shot frames the defect within the slope so its location is obvious (shoot a recognizable landmark like a vent or ridge in frame). The detail shot is the close-up with scale. Shot in immediate succession, the timestamps keep them paired, and anyone reviewing the file reads them as a unit.

Scale references that make damage objective

"It looks like hail" is an opinion. "Here is a strike with a coin next to it inside a chalked 10-by-10 square that contained nine strikes" is data. Carry and use scale on every damage photo:

• Chalk or a lumber crayon to circle each strike. Circling does three things: it makes the strike visible in a photo where it would otherwise vanish, it lets you count strikes per square, and it shows you didn't miss the obvious ones.
• A coin or a dedicated hail gauge placed beside a representative strike to show diameter. A U.S. quarter is about 0.955 inch; it's a universally understood reference in a photo.
• A tape measure for linear conditions: valley length, flashing runs, ridge length, gutter footage.
• A 10-by-10 chalk square for test squares, the standard sampling area used to assess strike density across a slope.

None of this is about exaggerating damage. It is about removing ambiguity so that whatever the true condition is, it reads the same to you, the homeowner, and the adjuster.

The test square, done correctly

The test square is the backbone of hail documentation, and it's where sloppy work shows. A defensible test square has all of the following in the photo record:

• An overview showing the square's location on the slope (so it's clearly representative, not cherry-picked from the worst spot).
• The chalked outline of a roughly 10-by-10-foot area.
• Each strike circled inside the square.
• A count — written on the deck in chalk or recorded in your notes — of total strikes in the square.
• The slope direction noted, because hail damage is often directional; the slope facing the storm's approach takes more hits than the leeward slopes.
• A representative close-up of a strike from that square with scale.

Do one square per slope direction on a cut-up roof, or one per slope on a simple roof. The goal is to show density and distribution, not to find the single worst shingle on the building. Adjusters discount cherry-picked damage; representative squares survive scrutiny.

Photographing the conditions that get challenged most

Some conditions get scrutinized harder than others, and those are exactly the ones to over-document. Adjusters and reviewers have seen every shortcut, so the burden of proof sits heaviest on the items that are easy to fake or exaggerate. Knowing which conditions draw scrutiny lets you front-load the evidence.

Soft metals tell the truest hail story. Hail leaves dents in soft metals — gutters, downspouts, gutter aprons, valley metal, roof vents, drip edge, fascia wraps, even HVAC condenser fins on the ground. These dents are hard to fake and harder to argue away than shingle bruising, because metal doesn't shed granules or self-heal in a photo. Shoot the soft-metal hits deliberately: a context shot showing the run, then a detail shot with scale. A roof with consistent denting on the wind-facing metals and matching strike density on the shingles tells one coherent story. When the shingle damage is subtle, the soft metals are often what make the documentation conclusive.

Directionality should be visible across the whole file. Wind and hail are directional. The slope facing the storm's approach takes more and harder hits; opposite and leeward slopes take fewer. When your per-slope test-square counts line up with the storm's known approach direction — west-facing slope high, east-facing low — the file reads as physically consistent. When counts are random or implausibly uniform across all slopes, a reviewer's antenna goes up. Document every slope's count even on the slopes with little damage; the low counts on the protected slopes are part of what proves the high counts on the exposed slope are real.

Distinguish mechanical and cosmetic from storm damage honestly. Foot traffic, a dropped tool, tree-limb abrasion, blistering, and manufacturing defects can all mimic or muddy storm damage. The disciplined move is to photograph and label them as what they are. Misrepresenting mechanical or pre-existing wear as fresh storm damage is how a file gets discredited entirely — once a reviewer catches one mislabeled condition, they distrust the whole set. Honest labeling of the non-storm conditions actually strengthens the credibility of the storm conditions you document beside them.

Date-of-loss consistency. For storm work, the condition you photograph should be consistent with weathering since the date of loss, not years of accumulated aging. You don't interpret coverage — that's the insurer's job — but you do document facts: fresh fractures show a different granule pattern than long-weathered surface loss. Capture the surface honestly and let the facts sit.

Aerial and drone imagery as part of the file

Ladders and roof walks are still the backbone, but aerial measurement reports and drone photos earn their place in a structured file when used correctly. An aerial measurement report gives you slope-by-slope square footage, ridge/hip/valley/eave/rake linear feet, and pitch without a tape, and those measured planes are what your slope-labeled overview photos attach to. Drone imagery adds safe overviews of steep or fragile roofs and a top-down perspective that ground shots can't capture.

The limits are real and worth stating. Aerial measurements are estimates from imagery and should be spot-verified against field reality, especially on additions, dormers, and complex cut-ups where automated tracing can miss a facet. Drones rarely resolve individual hail strikes well enough to replace a hands-on test square — they show distribution and overall condition, not the fine granule fracture you need for density. And if you fly commercially, the FAA's Part 107 rules apply; the pilot needs the certificate and must follow the airspace rules. Treat aerial imagery as a layer that strengthens the geometry and overview portion of the file, not a substitute for the close-up, scaled, hands-on documentation that the damage conditions require.

Naming and tagging: the convention that survives the field

Here's the uncomfortable truth: a field tech on a hot roof with gloves on is not going to type a careful filename for every photo. So your structure has to come from one of two places — a naming convention simple enough to actually follow, or software that captures the structure automatically so the tech never types anything. Most strong shops use both: software for the bulk, and a fallback manual convention for when the app fails or a subcontractor shoots on a personal phone.

A manual naming convention that works

If you're capturing to a plain camera roll and organizing later, adopt a slope-and-subject convention and shoot a slate. A "slate" is an old film-set trick: before each slope's photos, shoot one frame of a small whiteboard or a phone note that says the slope and address. That single divider frame lets anyone segment the camera roll into slopes after the fact, even if no other photo is named.

For files you rename in the office, use a sortable, machine-readable pattern:

[JOB]_[SLOPE]_[SUBJECT]_[SEQ]
1042_FrontSlope_TestSquare_01
1042_FrontSlope_HailDetail_03
1042_Rear_PipeBoot_01
1042_LeftRake_StepFlashing_02

Rules that make this hold up:

• Lead with the job number, not the date, so all of a property's photos sort together regardless of when shots were added.
• Use a controlled vocabulary for slope and subject — pick your terms once and never improvise. Front/Rear/Left/Right plus numbered facets for complex roofs. Subjects from a fixed list (Overview, TestSquare, HailDetail, Crease, PipeBoot, Flashing, Vent, Valley, DripEdge, Interior, Address).
• Zero-pad sequence numbers (01, 02 … 10) so they sort correctly.
• No spaces in filenames; underscores or hyphens only, so the names survive uploads, exports, and cloud sync.

The controlled vocabulary is the part people skip and regret. If one tech writes "FrontSlope," another writes "front," and a third writes "Slope1," you can't filter or batch anything. Publish the term list, laminate it, tape it inside the truck.

Tag-based structure beats folder-based structure

Folders force a photo into one bucket. A hail strike on the front slope at a pipe boot belongs to three categories at once — front slope, hail, penetration — and folders make you pick one. Tags don't. If your tool supports tagging (most field apps do), tag each photo with slope, subject, and condition, and let folders just hold the job. Then you can pull "every front-slope hail photo across the file" or "every flashing photo on the property" in one filter. That filtering ability is exactly what makes an estimate fast: you ask the file a question and it answers.

Metadata: the structure you get for free

Every digital photo carries embedded EXIF metadata, and most phones add GPS. This is free structured data that you are probably throwing away. Used well, it does three jobs no manual process can match.

Timestamps prove sequence and date

The capture time on each frame is a tamper-evident record of when you documented a condition. For storm work, the relationship between your inspection date and the date of loss matters; for warranty defense, a dated before-photo is decisive. Make sure every device's clock is set correctly and the time zone is right — a phone stuck on the wrong zone after travel can stamp photos an hour off and create needless confusion. Verify this once per device and forget it.

GPS coordinates confirm location

Geotags embed latitude and longitude in the photo. With location services on, every shot proves it was taken at the property, not pulled from a stock folder. This is quietly powerful for credibility: a file where the geotags cluster on the insured address reads very differently from one where they're scattered or absent. Turn location services on for the camera, confirm geotagging is enabled, and you get this for nothing. (Be aware geotags also mean these files carry the homeowner's address coordinates — handle and store them like the customer data they are.)

Don't strip metadata on export

Here's a trap: many sharing and compression steps strip EXIF. Texting a photo, dropping it into some chat apps, or running it through an aggressive image compressor can remove the timestamp and GPS. When you move photos into your claim file or estimate, use a path that preserves metadata — direct upload from the field app, or a transfer method that keeps originals intact. Keep the untouched originals archived even after you export working copies. If the date or location of a photo is ever questioned, the original with intact EXIF is your answer.

A quick metadata hygiene checklist

• Device clocks set to the correct local time and zone, verified per device.
• Location services enabled for the camera/field app; geotagging on.
• Originals archived before any compression or sharing.
• Export/transfer method confirmed to preserve EXIF (spot-check one exported file's properties).
• Storage that retains files for the length of your warranty and any statute-of-limitations window in your state.

From photos to an estimate: the linkage layer

This is the step that earns the money. You have a structured photo set; now it becomes a repair estimate. The estimate itself almost always lives in Xactimate for insurance-related work, because that's the platform most carriers price against, but the principle is platform-agnostic: every line item must trace back to photographic evidence, and every documented condition must produce a line item. Nothing floats unsupported in either direction.

Map measurements to slopes

Start with the roof geometry. Whether you measure by hand, walk it with a wheel, or pull an aerial measurement report, you need square footage per slope, ridge and hip linear feet, valley and rake and eave linear feet, and pitch. Your slope-labeled overview photos are what tie those measured planes to the documented damage — the measurement report says "front slope = 14.2 squares," and your photos prove that 14.2 squares is the one with the representative hail density.

Build the line items from the documented conditions

Work condition by condition. For each, the photo evidence drives a specific line:

The count and footage come straight from your accessory inventory and measurement photos. Because you shot one frame per accessory, you literally count the photos to get the unit quantity. Because you shot the four elevations and per-slope overviews, you can defend the squares. This is the entire payoff of structured capture: the estimate assembles from the file rather than from memory.

Code-driven items are documentation, not opinion

A large share of estimate disputes are over code-required items — ice-and-water shield in cold climates, drip edge, ventilation minimums, fastener counts, decking attachment. These aren't negotiable add-ons; they're what the International Residential Code (and your locally adopted amendments) require when a roof is replaced. Document the existing condition (no drip edge, inadequate ventilation, no ice barrier) and cite the adopted code section. That converts a code line from "upsell" to "required to bring the repair to current code," which is a documentation question you can answer factually. Know your local code adoption; jurisdictions amend the IRC, so verify the version your AHJ enforces rather than assuming.

Worked example: a directional hail slope

Make it concrete. A 24-square hip roof, storm approached from the west. Your file contains:

• Four elevations and four slope overviews. The west slope overview shows obvious surface granule loss.
• A west-slope test square: chalked 10-by-10, eleven strikes circled and counted, a quarter beside a representative strike showing roughly 1-inch diameter, slope direction noted.
• East-slope test square: same method, two strikes. South and north: four and five. The directionality is documented, not asserted.
• Accessory inventory: 3 pipe boots, 2 turtle vents, 1 power vent, ridge vent across the hips, 40 linear feet of valley, no existing drip edge at the rakes.
• A cut-up photo at the eave showing one layer of organic-felt underlayment, no ice barrier, on a roof in a cold-climate county.

From that file the estimate writes itself: replace the slopes the documentation supports, ridge cap by measured linear feet, valley metal by the 40 feet shown, three boots and three vents by count, drip edge added per code with the "existing: none" photo as justification, ice-and-water at eaves per the local cold-climate code with the cut-up photo and county code reference as support. Every number points to a frame. If an adjuster questions the west-slope replacement, you show eleven representative strikes with scale, not a cherry-picked macro. That is structured claim data doing its job.

The compliance line you cannot cross

Everything above is squarely in your lane: you may inspect, document, and prepare an accurate estimate to repair your own work, and you may state facts about your scope to the carrier. Where contractors get into real legal trouble is crossing from documentation into adjusting the claim, which in most states is unlicensed public adjusting and is enforced by the state department of insurance. The line is bright, and it's worth teaching your whole sales team explicitly.

What you may do:

• Inspect the roof and document conditions thoroughly with photos.
• Write an accurate, Xactimate-aligned estimate for the repair scope.
• State facts about your scope and your findings to the carrier.
• Hand the documentation and estimate to the homeowner so the homeowner can file and pursue their own claim.

What you may NOT do for a fee (this is the do-not-say and do-not-do list — print it for your reps):

• Negotiate, adjust, or "handle" the claim on the homeowner's behalf.
• Interpret the homeowner's policy or tell them what is or isn't covered.
• Promise a specific payout, approval, or that the claim "will get approved."
• Promise the deductible will be waived, absorbed, eaten, or made to disappear. (Absorbing or rebating a deductible is illegal in many states and is insurance fraud regardless.)
• Advertise or imply a "free roof."
• Represent the homeowner against the insurer.

The safe frame is simple to say and easy to live by: you document thoroughly, you write an accurate repair estimate, and you hand it to the homeowner; the homeowner files and the insurer decides coverage. Your photos and scope are facts about the roof and your work. The moment you start interpreting the policy or promising an outcome, you've stepped into licensed territory. Keep your value where it's legal and where it's actually strongest — being the contractor with the most thorough, most defensible documentation in the market. Thorough documentation is persuasive on its own; it doesn't need an illegal promise attached.

Software: what to look for, and the honest limits

Manual conventions work, but at volume you'll want field-documentation software to capture the structure automatically. The right tool removes typing from the field tech entirely: it knows the property, prompts the standard photo set, tags slope and subject with a tap, preserves metadata, and exports a clean PDF or pushes the structured set into your estimate and CRM. When you evaluate options, weigh these capabilities:

• Guided/structured capture — the app prompts the standard sequence so nothing is missed and every photo is born tagged.
• Slope and subject tagging by tap, with your controlled vocabulary, not free text.
• Metadata preservation — timestamps and geotags survive into the report.
• Test-square and annotation tools — circle strikes, drop scale, mark counts on the photo itself.
• Aerial measurement integration so slope square footage flows straight to line items.
• Estimate/Xactimate export or hand-off so the documented conditions become a sketch and scope, not a re-keyed list.
• Offline capture — roofs have bad signal; the app must work without bars and sync later.
• CRM/job linkage so the file attaches to the right property automatically.

Be honest about limits. No software shoots the photos for you — guided capture still depends on a tech who climbs the whole roof and frames the shots. No tool can manufacture scale or a test square that wasn't actually performed. And no app changes the compliance line: a slicker report does not authorize you to adjust the claim. Software amplifies a disciplined field process; it does not replace one. A bad process with great software just produces a well-organized pile of incomplete photos.

Where the roof selection itself fits: RoofPredict

Everything so far assumes you're already standing on the right roof. The earlier question — which roofs are even worth inspecting — is where most storm-restoration teams waste their season, knocking blocks that the storm barely touched or that have ten years of life left. This is the part RoofPredict is built for, and it sits upstream of the documentation workflow rather than inside it.

RoofPredict reads aerial imagery to estimate a roof-age range for each address — a range, not a precise install date, because imagery infers age, it doesn't read a birth certificate off the shingles — and models storm physics per roof, so you get the odds that a specific roof was worn out by the wind and hail that actually passed over it. The output is a ranking of the doors, routes, and lists most likely to hold roofs that are due: aging out on their own, or knocked down early by the storm. It also enriches your own CRM or mailing list with those roof-age and storm signals, so you're prioritizing the addresses you already own rather than buying leads.

The honest framing matters and we hold to it: a roof-age range is a probability, not a guarantee, and a storm model gives you odds a roof qualifies, never proof. RoofPredict tells you where the worn-out and aging-out roofs most likely are; it does not tell you what an adjuster will decide, and it doesn't document anything — that's still your boots, your camera, and the structured workflow above. Used together, the sequence is clean: RoofPredict ranks the roofs most likely to be due, your crew inspects the high-probability addresses, and the field-capture system turns those inspections into structured, estimate-ready claim data. You spend your inspection hours on roofs that are actually likely to need work, and every roof you do inspect produces a defensible file.

Common ways pros get this wrong

After reviewing thousands of claim files, the same failures repeat. Each one is cheap to avoid once you've named it.

1. No establishing shots. Stacks of close-ups with no overview to prove which slope they're on. Fix with the two-shot rule and per-slope overviews.
2. No scale on damage. Beautiful macros that prove nothing about size or density. Fix with chalk circles, a coin or gauge, and test squares.
3. Cherry-picked test squares. A square placed on the single worst spot reads as manipulation and gets discounted. Place it representatively and shoot the overview that proves it.
4. Improvised vocabulary. Three techs, three sets of slope names, nothing filterable. Fix with a laminated controlled term list.
5. Stripped metadata. Photos texted around until the timestamp and GPS are gone. Fix with metadata-preserving transfer and archived originals.
6. Missing accessory inventory. Boots and vents not photographed individually, so the count is a guess and the estimate is short. Fix with the one-photo-per-accessory rule.
7. No code-condition documentation. Code items written as line numbers with no "existing condition" photo to justify them. Fix by shooting the cut-up and the existing ventilation/drip edge.
8. Documentation drifting into adjusting. A rep promising approval or a waived deductible. Fix with the printed do-not-say list and training that keeps everyone on the documentation side.
9. Re-keying instead of linking. Photos and estimates living in separate systems, re-typed by hand. Fix with a tool that pushes the structured set into the estimate.
10. One person who "knows the system." A workflow living only in a veteran's head dies when they quit. Fix by writing the standard set, the vocabulary, and the checklist down so any new hire produces the same file.

Quality control before the file leaves your office

The last gate is a review pass that catches gaps while they're still cheap to fix — before the file goes to the homeowner and before anyone has to re-climb the roof. Build a five-minute office QC step into every job. The reviewer isn't second-guessing the tech's judgment; they're checking that the file is complete and self-explanatory to someone who wasn't on the roof.

Run the file against a pass/fail checklist:

If any row fails, it's flagged before the file ships. A failed metadata row gets re-exported. A failed scale or context row that can't be fixed from existing photos means a re-climb — annoying, but far cheaper now than after the estimate is built on missing evidence. The compliance row is non-negotiable: scrub any promise language from notes, texts, or the report before anything goes to the homeowner.

A simple maturity ladder for your documentation process

Most shops can place themselves on a four-rung ladder, and knowing your rung tells you the next move.

• Rung 1 — Camera roll. Photos live on personal phones, no convention, no review. Output is unreliable and undefendable. Next move: adopt the standard photo set and the two-shot rule.
• Rung 2 — Convention. A written standard set, controlled vocabulary, and the slate/naming pattern, executed manually. Files are usable but office labor is high. Next move: add the office QC checklist and metadata hygiene.
• Rung 3 — Tooling. Field-documentation software handles guided capture and tagging; metadata is preserved; export flows toward the estimate. Office labor drops sharply. Next move: integrate measurement and estimate hand-off so conditions become line items without re-keying.
• Rung 4 — Linked system. Capture, measurement, estimate, and CRM are connected; targeting upstream is data-driven; QC is routine. Files are fast, consistent, and defensible, and any new hire produces the same output. This is where documentation stops being a bottleneck and becomes an advantage.

You don't skip rungs. A shop on Rung 1 that buys Rung 3 software still produces Rung 1 files, because the discipline didn't come with the license. Climb in order.

Rolling the workflow out to a crew that won't read a manual

A documentation standard that lives in a binder is a standard nobody follows. Field crews learn by doing and by accountability, not by reading. A few moves make the rollout actually stick.

Make the standard set muscle memory with one ride-along each. Spend one inspection beside each tech walking the standard sequence out loud. After one guided run, most techs internalize the order. The sequence itself is the teaching tool — because it's always the same, it becomes a rhythm rather than a list to remember.

Laminate the controlled vocabulary and the checklist; put them in the truck. A one-page term list and the field checklist taped where the tech grabs the ladder removes the excuse of forgetting. The cost is a sheet of paper.

Review the first few files together. For each tech's first handful of jobs, run the QC checklist with them present so they see what "complete" looks like and what a kicked-back file costs. After that, the QC step runs without them, but the early shared reviews calibrate the standard.

Tie it to a number they feel. A re-climb costs a truck roll and a half-day. A kicked-back file costs the rep momentum and the homeowner's trust. When techs see that complete files mean fewer return trips and faster-closing jobs, the standard sells itself. Recognition for clean files beats nagging about messy ones.

Write it down so it outlives the veteran. The single biggest fragility in most shops is that the whole system lives in one experienced person's head. Document the standard set, the vocabulary, the QC checklist, and the compliance line in plain language so the day that person is out sick — or leaves — the next inspection still produces the same file.

A one-page field checklist

Hand this to every tech. It's the whole workflow compressed to what fits on a clipboard.

Before climbing:

• Confirm address; shoot house number + front elevation.
• Confirm device clock and time zone correct; location services on.

On the roof, every slope:

• Overview shot up the plane (establishing shot).
• Test square: chalk 10x10, circle and count strikes, scale on a representative strike, note slope direction.
• Every defect: context shot + detail shot with scale.

Accessories (one clear photo each):

• Pipe boots, B-vent/furnace flashings, turtle/ridge/power vents, satellite mounts, chimney, skylights.
• Valleys (with length), step/counter flashing, drip edge (or note absence), gutters.

Conditions:

• Cut-up showing layers, underlayment, ice barrier (cold climates).
• Existing ventilation type and count.

Interior (if safe and accessible):

• Leak staining, daylight at penetrations, deck/insulation water marks.

Before leaving:

• Quick scroll: is every slope covered? Does every damage close-up have a context shot and scale? Any missing accessory?

Back at the truck:

• Sync/upload via the metadata-preserving path. Confirm the file attached to the right job.

Putting it together

Structured claim data isn't a software feature you buy; it's a habit your field team builds. The phone already gives you resolution, a timestamp, and a location. Your job is to add the four layers that pixels can't supply on their own — where on the roof, what it is, how big or how many, and which line it justifies — and to do it the same way every time so any person in your shop can read the file. Shoot the standard set. Use the two-shot rule. Put scale on everything. Run representative test squares. Keep a controlled vocabulary. Preserve the metadata. Link every condition to a line item and every line item to a photo.

Do that and three things happen: your estimates assemble in an hour instead of a day, your files survive scrutiny instead of getting kicked back, and your warranty and dispute exposure drops because you can always prove what the roof looked like and when. Pair the discipline with smart targeting upstream — inspecting the roofs most likely to be due rather than every door on the block — and you spend your season producing defensible documentation on roofs that actually need work. Stay on the documentation side of the line, hand the homeowner an accurate estimate, and let the structured file speak for itself. It's more persuasive than any promise you're not allowed to make.

FAQ

How many photos should a single roof inspection have?

A simple roof typically yields 40 to 90 deliberate frames; a complex cut-up roof can reach 150 or more. The number matters less than the coverage: every slope gets an overview, every defect gets a context shot plus a scaled detail shot, every accessory gets one clear photo, and each slope direction gets a representative test square. If those conditions are met, you have enough; if they aren't, a thousand random shots still leave gaps.

What is a test square and why does it matter?

A test square is a defined sampling area, conventionally about 10 by 10 feet, chalked onto a slope so you can assess hail-strike density objectively. You circle and count every strike inside the square, place it in a representative spot (not the worst), and shoot an overview proving its location. It converts 'this roof has hail' into a countable, defensible measurement of how much and where, which is far harder to dispute than a single dramatic close-up.

Does the photo metadata really matter, or is it just nice to have?

It matters. The embedded timestamp proves when you documented a condition, and the GPS geotag proves you were at the property. Both are valuable for storm-claim credibility and for warranty or dispute defense later. The catch is that texting, certain chat apps, and aggressive compression can strip this data, so move photos through a transfer path that preserves EXIF and keep the untouched originals archived.

What file naming convention works best in the field?

Lead with the job number, then slope, then subject, then a zero-padded sequence number, using a fixed controlled vocabulary and no spaces, for example 1042_FrontSlope_TestSquare_01. In the field, where typing is impractical, shoot a 'slate' frame (a whiteboard or phone note naming the slope and address) before each slope's photos so the camera roll can be segmented later, or use software that tags slope and subject with a tap so the tech never types at all.

How do roofing photos turn into an Xactimate estimate?

Each documented condition maps to a line item and each line item traces back to a photo. Slope square footage from your measurement comes from the slopes your overview photos identify; accessory counts come from your one-photo-per-accessory inventory; flashing and valley footage come from your scaled detail shots; code items like drip edge, ice barrier, and ventilation come from your 'existing condition' photos. Nothing in the estimate floats without supporting evidence, and no documented condition is left without a corresponding line.

Can a roofing contractor handle the insurance claim for the homeowner?

No. Negotiating, adjusting, or 'handling' a claim for a fee on the homeowner's behalf is unlicensed public adjusting in most states. A contractor may inspect, document, write an accurate repair estimate, and state facts about their own scope to the carrier, then hand the documentation to the homeowner. The homeowner files and the insurer decides coverage. Interpreting the policy or representing the homeowner against the insurer crosses the line.

Is it legal to offer a homeowner a 'free roof' or to waive their deductible?

No. Advertising a 'free roof' and absorbing, rebating, or waiving the insurance deductible are prohibited in many states and are treated as insurance fraud regardless of state. The deductible is the homeowner's contractual obligation. A contractor should never promise a waived deductible, a specific payout, or guaranteed claim approval. Keep your value in thorough documentation and an accurate estimate, which is both legal and more persuasive than a promise you can't make.

Do I need special software, or can I do this with a phone?

You can do it with a phone and a disciplined manual convention: standard photo set, two-shot rule, scale references, controlled vocabulary, and metadata-preserving uploads. Software helps at volume by prompting the standard sequence, tagging by tap, preserving metadata, and exporting the structured set into your estimate and CRM, which removes typing from the field. But software amplifies a good process rather than replacing it; without a tech who shoots the full set correctly, even the best app just organizes incomplete photos.

How does RoofPredict fit into a photo-documentation workflow?

RoofPredict sits upstream of documentation. It uses aerial imagery to estimate a roof-age range per address and models storm physics per roof, then ranks the doors and lists most likely to hold roofs that are due, so your crews inspect roofs that probably need work instead of every door on the block. It does not document anything itself, and its outputs are probabilities, a roof-age range and storm odds, not proof of damage or claim approval. Your boots, camera, and the structured field workflow still produce the actual claim file.

How long should I keep roof inspection photos?

Keep originals at least as long as your workmanship warranty and any applicable statute-of-limitations window in your state, since old photos are your defense in callbacks and disputes. Archive the untouched originals with intact metadata separately from working copies, and remember the files contain the homeowner's address and geotags, so store them securely as the customer data they are.