How to Build a Roof Claim Evidence Index Your Carrier Can't Wave Away

Most roofing crews lose money on claims they already earned. The damage was real, the estimate was fair, and the carrier still paid less than the job costs — not because the contractor was wrong, but because the file was a mess. Forty photos in a phone roll with no captions. A measurement report that doesn't tie to anything. A repair estimate with line items nobody can connect to a slope, a course, or a single photo. When a desk adjuster opens a file like that, they don't see a roof. They see noise, and noise gets the low number.

A roof claim evidence index fixes that. It is the organized spine that ties every photo, every measurement, every code reference, and every estimate line item together so that a person who has never set foot on the property can follow your reasoning in order and verify it without calling you. Build it well and your accurate estimate gets reviewed on its merits. Build it badly and even a perfect estimate gets discounted.

Before we go further, one boundary that matters for every roofing company in the restoration space. A contractor's lane is to inspect the roof, document the condition thoroughly, and prepare an accurate repair estimate for their own scope of work — then hand that documentation to the homeowner. The homeowner files the claim. The insurer decides what is covered. You do not negotiate the claim for a fee, interpret the homeowner's policy, promise an approval or a payout, or tell anyone their deductible is going to disappear. We will come back to that line several times because the evidence index is exactly where contractors either stay on the right side of it or wander off. Everything below is built to make your documentation stronger while keeping you firmly in the contractor's role: document, estimate, hand it over.

What a roof claim evidence index actually is

Strip away the jargon and an evidence index is a numbered table of contents for a single property's claim file. Every piece of proof gets a stable ID. Every estimate line points back to one or more of those IDs. Every code citation points to the published source. The index itself is a short document — usually one to three pages — that lets a reader jump straight from "why is there a line for drip edge" to "photo 14, north eave, existing drip edge corroded, plus IRC R905.2.8.5" without hunting.

Think of it as three layers stacked on top of each other:

1. The raw evidence — photos, the aerial or hand measurement report, weather data, the original carrier paperwork, manufacturer specs, code text. These are the exhibits.
2. The index — the map that assigns each exhibit a number, a location, a date, and a one-line description, and groups them so a reader can navigate.
3. The estimate cross-reference — the column or table that connects each Xactimate (or comparable) line item to the exhibit numbers that justify it.

When those three layers exist and agree with each other, you have a packet that a desk reviewer can audit in ten minutes instead of bouncing back with "please provide documentation supporting the following line items." That bounce-back is where weeks disappear.

Why the index beats a pile of photos

A contractor with 200 photos thinks they have great documentation. A desk adjuster reviewing 30 files that day sees an unsorted pile and reaches for the parts they can quickly confirm — usually the obvious field damage — and writes the rest off as unsupported. The volume works against you. Forty captioned, located, sequenced photos that march cleanly through the roof beat 200 random ones every time, because the reviewer can follow them.

The index is also what protects you when memory fades. A supplement might get reviewed 60 or 90 days after inspection. The crew that climbed the roof has done 40 roofs since. Without an index, "why did we call out the pipe boots" becomes a guess. With one, it's photo 22 and 23, both penetrations, both with cracked collars, dated and time-stamped.

The principles that make evidence hold up

Before building anything, internalize five rules. Skip these and the prettiest index in the world still gets discounted.

Contemporaneous beats reconstructed. Evidence captured at the time of inspection, with original timestamps and location data intact, carries far more weight than anything assembled later from memory. Shoot it right the first time.

Metadata is evidence. The EXIF data inside a photo — capture time, GPS coordinates, device — is part of the proof. Do not strip it. Screenshotting a photo to "clean it up," emailing it in a way that recompresses it, or running it through a filter destroys the metadata and weakens the file. Keep originals.

Specific beats general. "Hail damage to roof" proves nothing. "Photo 9: south slope, field shingle, 1.25-inch impact mark with mat fracture and granule displacement, measured against chalk circle" proves something. Every caption should answer where, what, and how you know.

Every estimate line needs a parent. If a line item exists in your estimate, at least one exhibit must justify it. A line with no evidence parent is the first thing a reviewer strikes. Conversely, evidence with no line item is wasted work — useful for completeness, but it isn't earning anything.

You document; you do not adjudicate. Your captions describe what you observed and what your scope requires. They do not say "this is covered," "the carrier owes," or "this will be approved." State facts about the roof and your repair scope. Coverage is the insurer's call, filed by the homeowner. Keep your language on the observable and the estimable.

The field capture workflow that feeds the index

The index is only as good as what comes off the roof. A disciplined capture sequence is what separates a packet that builds itself from a packet you fight with for hours at the office. Here is a sequence that works whether you have a four-person production team or you're a solo owner-operator.

Step 1: Establish the property baseline before you climb

Shoot the address and the house from the street so there is no question which property this file belongs to. Capture all four elevations from the ground. These wide shots set scale and orientation, and they give the reviewer a mental model of the building before you take them up close. A reviewer who is oriented trusts the close-ups more.

While you're on the ground, photograph collateral and ground-level indicators: dented gutters, downspouts, fascia, gutter screens, AC condenser fins, soft metals like mailbox and vent caps, window screens, and any spatter on painted surfaces. Collateral on soft metals is some of the most persuasive storm evidence you can capture because it's hard to fake and easy to verify. It belongs in the index even though it isn't roof surface, because it corroborates that an event occurred.

Step 2: Orient by slope and shoot in a fixed rotation

Decide on a consistent slope-naming convention and use it on every roof, forever. North/South/East/West is the cleanest because it's unambiguous and ties to aerial reports. Front/Right/Rear/Left works too as long as you anchor it ("right as you face the front door"). Pick one. Inconsistency between roofs is how supplement managers lose their minds.

For each slope, shoot in the same order every time:

• A slope overview showing the whole plane.
• Mid-range shots working across the slope so coverage is continuous, not random.
• Close-ups of each damage indicator, ideally with a reference for scale — a chalk circle, a quarter, a measuring device, or a gauge.
• The transitions and details for that slope: ridge, hip, valley, eave, rake, penetrations, flashing.

Doing the same rotation on every slope means your index practically writes itself, because the photos already arrive grouped and in order.

Step 3: Establish scale and prove impact, don't just assert it

A bare close-up of a mark is weak. The same mark photographed next to a coin, a hail gauge, or a chalk-circled test square is strong because the reviewer can see the size and see that you measured it rather than eyeballed it. For hail specifically, the persuasive elements are: a measurable impact mark, granule displacement exposing the mat, a fracture or bruise you can feel and often see, and a pattern consistent with a storm rather than scattered mechanical scuffs. Document the mark, the size reference, and a wider shot showing the density across the slope.

A test square — a marked area, often four feet by four feet or ten feet by ten feet depending on what your region's adjusters expect — with the impacts circled is a standard, well-understood way to show damage density per slope. Photograph each test square so the slope and the chalk are both visible. This is documentation of condition. It is not a verdict on coverage.

Step 4: Capture the things that drive scope, beyond the damage

The damage justifies the claim. The roof's existing construction justifies the scope. These are different photo sets and both belong in the index.

Scope-driving conditions to capture on every inspection:

• Layers. Photograph a cut or edge that reveals how many shingle layers exist. Two layers change tear-off labor and disposal.
• Decking condition where visible.
• Existing flashing at walls, chimneys, and skylights, and its condition. Reused flashing on a full replacement is a common scope gap.
• Drip edge presence and condition at eaves and rakes.
• Ventilation — ridge vents, box vents, turbines, power vents — counted and photographed.
• Pipe boots and penetrations, each one, with collar condition.
• Valley type (open metal, closed-cut, woven) because it dictates the line item.
• Underlayment and ice-and-water evidence where exposed.
• Steep or two-story access conditions that affect labor.
• Pitch, measured, because it changes the price modifier.

Each of these maps directly to an estimate line. A reviewer who can see the existing drip edge corroded in a photo is far less likely to strike your drip-edge line.

Step 5: Note local code triggers at the property

Many legitimate estimate lines exist because code requires the item when you re-roof, not because the storm damaged that specific component. Drip edge, ice barrier in cold climates, deck attachment, and underlayment requirements commonly come from the adopted building code. Your job at the property is to photograph the existing condition and note which code-driven items apply. The code citation gets attached in the index. You are documenting that the code applies and that the existing condition doesn't meet it — not arguing coverage.

Building the index itself, field by field

Now assemble the layer that ties it together. An evidence index entry should carry, at minimum, these fields. Keep it boring and consistent.

The sequential ID is the load-bearing part. Once a photo is P-014, it is P-014 in the index, in the estimate cross-reference, and in any letter. Never renumber. If you delete a bad photo, leave a gap rather than shifting everyone down — a renumber after the fact breaks every reference downstream and is how packets quietly fall apart.

Grouping and ordering

Group the index the way a person walks a roof, not the order your phone happened to capture things. A reviewer-friendly order:

1. Property baseline (address, elevations)
2. Ground-level collateral
3. Slope-by-slope, in your fixed rotation, each slope's overview then damage then details
4. Roof-wide systems: ventilation, penetrations, flashing
5. Scope-driver close-ups: layers, decking, drip edge, pitch
6. Measurement report
7. Weather and storm-date evidence
8. Code references
9. Original carrier paperwork

Within a slope, overview first so the reader is oriented, then the detail shots. The reviewer should be able to read your index top to bottom and reconstruct the inspection in their head.

The one-line description is where pros separate from amateurs

Write descriptions in a fixed grammar: Location — feature — observed condition — basis.

• Weak: "hail damage"
• Strong: "South slope — field shingle — 1.25" impact with mat fracture and granule loss — measured against chalk circle, photo P-031"

The basis clause ("measured against," "counted," "compared to manufacturer spec") is what turns an assertion into a documented observation. It also keeps you honest: if you can't write the basis, you don't actually have the evidence yet, and you should go get it.

Connecting evidence to estimate line items

This is the cross-reference layer, and it's where most documented estimates are won or lost. A desk reviewer's core question on every line is "what supports this?" Your cross-reference answers it before they ask.

Build a simple table that runs the other direction from the index — line item first, exhibits second:

Three disciplines make this table do its job:

No orphan lines. Every line has at least one exhibit in the support column. If a line is genuinely there for a code reason rather than a photographed condition, the code citation IS its support — but then the code column must be filled, and the index must contain the code text or a link to it.

Quantities trace to a source. A quantity that matches the measurement report is verifiable in seconds. A quantity that doesn't match anything invites the reviewer to substitute their own number. If you're claiming more than the measurement shows — waste factor, starter, ridge cap — state the basis in the quantity column.

Specs are attached, not asserted. If a line depends on a manufacturer requirement (for example, a specific underlayment or a ventilation ratio), put the spec sheet in the index as a document exhibit and cite it. "Manufacturer requires" with nothing attached is an invitation to strike the line.

Worked example: the drip edge line

Walk one line end to end because the pattern repeats for everything.

• The estimate line: R&R drip edge at eaves and rakes, quantity in linear feet from the measurement report.
• The photographed condition: P-014 (north eave, existing drip edge corroded), P-018 (east rake, no drip edge present), P-021 (south eave, drip edge undersized).
• The code basis: the adopted code's drip-edge requirement for asphalt shingle roofs.
• The index entries: each photo carries its slope, the observed condition, and the linked line.
• The cross-reference row: line item, LF basis from measurement, the three photos, the code citation.

A reviewer reading that can confirm in under a minute that drip edge is missing or failed, that code requires it on a re-roof, and that the quantity matches the measured perimeter. There's nothing to argue with. Compare that to a bare "drip edge — 180 LF" with no photo and no code — same line, completely different outcome.

The document chain: keeping the carrier's own paperwork in the file

An evidence index isn't only your photos. The original claim paperwork the homeowner received is part of the record, and tracking it cleanly is half of supplement work.

Keep, dated and labeled, every version of:

• The carrier's estimate or scope sheet, including each revision.
• Any engineer or third-party inspection report.
• The homeowner's policy declarations page (for understanding deductible amounts and coverage type — you read these to scope and invoice correctly, not to interpret coverage for the homeowner).
• Correspondence and decision letters.
• Your own estimate, each version.
• Photo reports from any prior inspection.

Why keep the carrier's documents in your index? Because a supplement is a comparison. The cleanest supplement is a line-by-line diff between the carrier's scope and a complete, code-compliant scope, with every difference tied to an exhibit. When you can show "the carrier's estimate omits drip edge; here is the existing failed drip edge in P-014; here is the code requirement," you've made a factual, documentable case. You have not interpreted policy. You have not promised the difference will be paid. You've shown a scope gap and supported it.

A note on classifying and reading the documents

Claim paperwork arrives as a mess — PDFs with different layouts per carrier, photos from three sources, estimates with hundreds of line items. The grunt work is sorting it: which document is the carrier estimate, which is the engineer report, which line items appear where, what the deductible figure is. Doing this by hand for every file is slow, and slow is what kills recoverable depreciation and supplement deadlines. We'll come back to how to compress this.

Storm-date and weather evidence: useful, with limits

If the claim is storm-related, the date and the event belong in the index. Pull the verifiable record for the property's location and the loss date from authoritative weather sources, and include it as a weather exhibit with the source named.

What weather data does and doesn't do, honestly:

• It corroborates that a hail or wind event occurred near the property around the date of loss. That supports the plausibility of storm damage.
• It does not prove that this specific roof was damaged by that specific storm. Hail swaths are patchy; a confirmed regional report doesn't guarantee impact at one address, and the absence of a report doesn't prove nothing happened. Your physical evidence — the impact marks, the directional pattern, the collateral on soft metals — is what shows damage at the property. Weather data supports the timeline; the roof shows the damage.

State it that way in the index. "NWS storm report confirms hail in the area on the date of loss" is a defensible caption. "This storm caused this damage" overstates what the weather record shows and is the kind of leap a reviewer is trained to challenge. Let the physical evidence carry the damage finding and let the weather data carry the timeline.

Where RoofPredict does the heavy lifting

Everything above is a manual workflow, and a sharp supplement manager can run it on a spreadsheet. The problem is volume and consistency — doing it the same way on the 200th roof as the 1st, and not letting a depreciation deadline slip because someone was buried in PDFs. RoofClaim, the claim revenue-cycle side of RoofPredict, is built to compress exactly the grunt work in the index without touching the part that has to stay in your hands.

Here is what a contractor actually does with it on a claim file:

Intake linked to the home. The claim attaches to the specific property, so the photos, measurement, and original carrier documents live in one place tied to that address rather than scattered across a phone, an email thread, and a drive folder. The index has a home.

Upload, auto-classify, and OCR the claim docs. Drop in the carrier estimate, the engineer report, the photos, the denial letter, the invoices. The system classifies what each document is and runs OCR so the line items and figures inside the carrier's PDFs become text you can compare against — instead of you retyping a 200-line estimate to build a diff. That is the document-chain step from the previous section, done in minutes instead of an afternoon.

Opportunity detection that maps line items to a roofing knowledge base. This is the part that earns its keep. The carrier's estimate gets mapped against a roofing scope knowledge base, and the system flags where scope appears missing, where a code-required item isn't present, and where a supplement was likely missed — each flag carrying an evidence anchor and a pricing reference. So instead of a supplement manager remembering from experience that this carrier always forgets drip edge, the gap surfaces with the anchor to attach. You still review every flag and decide what's defensible. It surfaces candidates; you make the call. That keeps the judgment — and the responsibility — with you.

Packet-completeness scoring and supplement cadence. Before you send, the file gets a completeness check: are the supporting exhibits attached, are the code citations present, are the quantities sourced. A packet that scores low is a packet a desk reviewer will bounce. The supplement aging and follow-up cadence keeps each open item from going quiet — supplements die from neglect more than from rejection, and a tracked cadence is what keeps them moving.

Recoverable-depreciation autopilot and deductible tracking. The depreciation piece has a checklist of its own — completion evidence and a final invoice — and a clock. The autopilot tracks the completion-evidence and final-invoice requirements so the recoverable depreciation actually gets released instead of being left on the table because the paperwork lapsed. Deductible tracking keeps the homeowner's deductible visible and correctly invoiced, which is a billing function on your side of the line.

Locked, UPPA-gated templates. The outputs — supplement packets, depreciation-release letters, deductible invoices, missing-docs letters, audit reports — generate from templates that are locked to contractor-documentation language. They produce documentation of scope and condition, not claim negotiation. That gating is deliberate: it makes it harder to accidentally cross from "here is our documented estimate" into "we'll handle your claim," which is the line that gets contractors in trouble.

The honest limit: none of this decides coverage, files the claim, or talks to the carrier for you. It organizes your evidence, surfaces likely gaps with anchors, and keeps the cadence and deadlines from slipping. The homeowner still files. The insurer still decides. You still review and own every line. What you save is the hours of sorting, retyping, and chasing — the work that has nothing to do with your expertise and everything to do with whether the file gets done right and on time.

A full worked file, start to finish

Abstract rules land better against a real sequence. Walk one ordinary hail file the whole way through so the pattern is concrete.

The property. A two-story home, architectural shingles, roughly 28 squares, four primary slopes plus a garage, reported hail date six weeks back. The homeowner called because a neighbor's roof was approved and they want to know if theirs qualifies. Your job is to inspect, document, and produce an accurate estimate for them to file — not to tell them they'll be approved.

Ground work. You shoot the street view and address (P-001), the four elevations (P-002 through P-005), and the collateral: dented gutter run on the north side (P-006), bruised downspout (P-007), dimpled AC condenser fins (P-008), spatter marks on the south-facing fascia (P-009), and a soft-metal vent cap with clear impact (P-010). Right there you have a corroboration set that an event with hail-sized impact energy reached this property — independent of the roof surface.

On the roof. You name slopes North/South/East/West and shoot each in the fixed rotation. North slope: overview (P-011), three mid-range passes (P-012 to P-014), four damage close-ups with a hail gauge for scale (P-015 to P-018), and a chalk test square with seven circled impacts (P-019). You repeat the rotation on the other three slopes and the garage. By the time you climb down you have roughly 60 photos that already sort themselves, because every slope went overview-to-detail in the same order.

Scope drivers. You cut a shingle at the eave to reveal a single layer (P-044), photograph the existing undersized drip edge at the north eave (P-045) and its absence at the east rake (P-046), count and shoot five penetrations and their pipe boots (P-047 to P-051), capture the wall flashing at the chimney showing rust-through (P-052), and measure the pitch at 8/12 (P-053). Each of these is a future estimate line you can now point a reviewer at.

Office assembly. You pull originals off the phone with metadata intact, assign sequential IDs, and write each index entry in the location-feature-condition-basis grammar. You import the measurement report as M-001, the NWS storm report for the loss date as W-001 (labeled as timeline corroboration, not causation), and the carrier's original scope sheet as D-001. Then you build the cross-reference: shingle R&R tied to the measurement and the slope overviews; drip edge tied to P-045 and P-046 plus the code citation; pipe jacks tied to the five penetration photos; steep charge tied to the 8/12 pitch photo; chimney flashing tied to P-052.

The supplement comparison. D-001, the carrier's scope, omits drip edge and the chimney flashing and prices the pitch as standard. Your index already holds the photos and the code citation for each gap. The supplement writes itself as a factual diff: here is what a complete, code-compliant scope includes, here is the photo of the existing failed component, here is the code requirement. You hand it to the homeowner. You do not tell them it will be paid, and you do not tell them what their policy covers. You've documented the gap and supported it; the decision sits with the insurer.

That is the entire arc. Notice what's absent: no guesswork, no scrambling to remember why a line exists, no captions that predict an outcome. The discipline up front is what makes the back end fast and defensible.

File management and naming so nothing rots

The best index in the world fails if the underlying files are a swamp. A few habits keep the raw evidence usable months later.

One folder per claim, tied to the address. Everything for a property lives in a single place: photos, measurement, carrier docs, your estimate, correspondence. Scattering across a phone roll, an email thread, and three peoples' drives is how originals get lost and metadata gets stripped in transit.

Filename carries the ID and the location. A file named for its exhibit ID and slope — something like P-019-north-testsquare — is self-describing. When a reviewer or a teammate opens the folder, the filenames already tell the story. Random camera filenames force everyone back into the index to decode what they're looking at.

Originals are sacred; work on copies. If you need to annotate, circle, or brighten a photo for clarity, do it on a copy and keep the original untouched in the folder. The original with intact metadata is the evidence; the annotated version is a presentation aid. Never let the annotated copy replace the original.

Version the documents. When the carrier issues a revised scope, keep both the old and the new, dated. A supplement often turns on what changed between revisions, and you can't show a change if you overwrote the earlier version.

Retention has a clock. Claim files can matter long after the job closes — recoverable depreciation, a later dispute, a warranty question. Keep the full file for years, not weeks. Storage is cheap; a reconstructed file from memory is worthless.

How supplements and depreciation actually move

Two mechanics inside a claim file trip up contractors more than the inspection itself, because they're administrative rather than technical. Worth being precise about both, strictly on the contractor's side.

Supplements are a documented diff, not a negotiation. A supplement is what you submit when the complete, code-compliant scope exceeds the carrier's initial scope. Your role is to document the gap — the missing or under-scoped items — with exhibits and citations, and provide it so it can be reviewed. You are not negotiating a number for a fee or arguing coverage. You're showing, factually, that the roof requires more scope than the first estimate reflected, and supporting each item. The packet-completeness discipline matters most here: a supplement with photos and citations attached gets reviewed; one without gets bounced or ignored.

Recoverable depreciation is a checklist with a deadline. On a replacement-cost claim, a portion is often initially withheld as depreciation and released after the work is complete and documented. The contractor's side is mechanical: provide completion evidence and a final invoice that matches the approved scope, within the carrier's timeframe. This is pure documentation and billing for work you actually performed — not adjusting. The money gets left on the table constantly, not because it was denied, but because the completion paperwork lapsed or the deadline passed while everyone moved on to the next roof. A checklist and a clock fix it.

Deductibles are the homeowner's obligation, tracked as billing. The homeowner owes their deductible. You track it so you invoice correctly and collect what's due for the work. What you never do is tell a homeowner the deductible will be waived, absorbed, or made to disappear, or imply the roof is free — that's both a misrepresentation and, in many places, illegal. Track the figure, invoice it, collect it. That's the whole of the contractor's role on the deductible.

Keeping these three straight — supplement as documented diff, depreciation as a completion checklist, deductible as straightforward billing — is what separates a roofing company that reliably collects the revenue it earned from one that does great roofs and quietly loses money on paperwork.

The compliance line, stated plainly so nobody crosses it

The evidence index makes your documentation strong, which is exactly when it's tempting to overreach. Here is the do-not-say list, because teaching it is part of doing this right.

You MAY:

• Inspect the roof and document its condition thoroughly.
• Write an accurate, Xactimate-aligned repair estimate for your own scope of work.
• State facts about your scope and the roof's condition to anyone, including the carrier, regarding the work you will perform.
• Hand your documentation and estimate to the homeowner so they can file.
• Track your own deductible billing and recoverable depreciation as part of getting paid for completed work.

You MAY NOT, when you are not a licensed public adjuster:

• Negotiate, adjust, or "handle" the claim for the homeowner for a fee.
• Interpret the homeowner's policy or tell them what their coverage means.
• Promise a specific payout, an approval, or that a supplement "will" be paid.
• Tell a homeowner their deductible will be waived, absorbed, eaten, or made to disappear.
• Advertise a "free roof" or imply the insurance will fully cover everything at no cost.
• Represent the homeowner against the insurer.

The evidence index lives entirely inside the "may" column. It documents condition and supports an estimate. The moment your captions or letters start saying "the carrier owes," "this is covered," "you'll pay nothing," or "we'll get this approved," you've left documentation and entered adjusting. Keep the language factual and observable, keep the decision with the insurer, keep the filing with the homeowner, and the index protects you instead of exposing you.

A quick language test for any caption or letter: does the sentence describe something you observed or something you'll do (safe), or does it predict what the insurer will pay or decide (not safe)? "Existing drip edge corroded, code requires replacement on re-roof" passes. "Carrier will approve drip edge" fails. Run every line through that test.

A repeatable packet-assembly checklist

Turn the whole thing into a checklist your team runs the same way every time. Consistency is the entire point.

Before the climb

• Address and four elevations from the ground
• Ground-level collateral (gutters, soft metals, screens, AC fins)
• Confirm slope-naming convention for this roof

On the roof, per slope

• Slope overview
• Continuous mid-range coverage across the plane
• Damage close-ups with scale reference
• Test square with circled impacts, photographed with slope visible
• Transitions and details: ridge, hip, valley, eave, rake

Roof-wide scope drivers

• Ventilation counted and photographed
• Each penetration and pipe boot
• Existing flashing at walls, chimney, skylights
• Drip edge presence and condition
• Layers revealed at a cut or edge
• Decking where visible
• Pitch measured
• Access and steep conditions

At the office

• Pull and verify original capture metadata; keep originals
• Assign sequential exhibit IDs; never renumber
• Write each index entry: location, feature, condition, basis
• Group the index in walk-the-roof order
• Build the line-item cross-reference; confirm no orphan lines
• Trace every quantity to the measurement report or a stated basis
• Attach manufacturer specs as document exhibits where cited
• Attach code citations for code-driven lines
• Add weather exhibit with source named (timeline only, not causation)
• Include carrier's original estimate and all revisions
• Run the language test on every caption and letter
• Score packet completeness before sending

Common mistakes that quietly cost money

Renumbering after the fact. Someone deletes three blurry photos and renumbers the rest. Now the cross-reference points at the wrong images and the index lies. Leave gaps instead.

Captions that assert instead of document. "Storm damage" everywhere. It tells the reviewer nothing and signals that you didn't look closely. Replace with location-feature-condition-basis.

Stripped metadata. Photos screenshotted, filtered, or recompressed on the way through email. The capture time and location — real evidence — are gone. Move originals, not copies that pass through processing.

Orphan line items. Lines with no exhibit behind them. They're the first struck and they make the reviewer distrust the lines that ARE supported.

Evidence with no line item. Forty photos of a beautiful test square and no corresponding scope. Documentation that earns nothing is wasted effort — tie it to a line or know why it's there.

Mismatched quantities. The estimate says 200 LF of ridge; the measurement report says 160. The reviewer trusts the report and substitutes the number, and now your whole file looks loose.

Reused flashing on a full replacement. Estimating around existing flashing instead of replacing it. It's a scope gap that's easy to document with a photo and easy to lose by forgetting.

Letting supplements go quiet. An open supplement with no follow-up cadence ages out of attention. Most lost supplements weren't rejected; they were forgotten. A tracked cadence fixes it.

Crossing the compliance line in the packet language. A caption or letter that predicts the payout or promises approval. It can taint an otherwise clean file and put your license and reputation at risk. Run the language test.

Scaling this past the first ten roofs

A solo owner can run this on a spreadsheet and a phone. The trouble starts when you're doing volume and the documentation depends on whoever happened to climb the roof that day. The fixes are organizational, not heroic.

Standardize the rotation so the index is automatic. When every inspector shoots the same sequence, the photos arrive pre-grouped and the index entries write themselves. Variation between people is what creates office hours.

Template the captions. Give inspectors a fixed caption grammar and a short pick-list of conditions. "South / field shingle / impact with mat fracture / chalk-circled" is faster to write and more consistent than free text.

Separate capture from assembly. The person on the roof captures; a supplement specialist assembles the index and cross-reference. The specialist gets fast at the pattern and the field crew isn't slowed down.

Score before you send. A completeness check — exhibits attached, citations present, quantities sourced, no orphan lines — catches the bounce-back before the carrier does. One avoided bounce-back is often a week saved.

Track the cadence centrally. Open supplements and pending depreciation need a clock and an owner, or they slip. This is the single highest-return discipline once you're at volume, because the work is already done — you just have to not drop it.

This is the gap between a contractor who's great on a roof and a roofing company that reliably gets paid for the scope it earned. The roof skill gets the damage right. The evidence index, run the same way every time, is what gets it documented well enough that an accurate estimate gets reviewed on its merits.

Bringing it together

A roof claim evidence index is not paperwork for its own sake. It's the difference between handing a homeowner a documented, defensible estimate and handing them a pile of photos that gets discounted on a busy adjuster's desk. Build the three layers — clean contemporaneous evidence, a numbered index in walk-the-roof order, and a line-item cross-reference with no orphans — and your accurate estimate gets read the way you meant it. Stay inside the contractor's lane the entire time: document condition, write the estimate, hand it to the homeowner, let them file, let the insurer decide. That's the whole job, done right.

If you're tired of rebuilding this from scratch on every file and watching depreciation and supplements slip because someone was buried in PDFs, RoofClaim inside RoofPredict runs the index, the line-item mapping with evidence anchors and pricing, the completeness scoring, and the depreciation and supplement cadence on locked, contractor-documentation templates — so the grunt work compresses and the judgment stays yours. Document thoroughly, estimate accurately, hand it over. The index is how you make that hold up.

FAQ

What exactly goes into a roof claim evidence index?

Three layers: the raw evidence (photos, the measurement report, weather data, the carrier's original paperwork, manufacturer specs, and code text); the index itself, which assigns every exhibit a sequential ID, a date, a location, and a one-line description and groups them in walk-the-roof order; and a line-item cross-reference that connects each estimate line to the exhibits that support it. When all three exist and agree, a reviewer who never saw the property can verify your reasoning without calling you.

How many photos do I actually need for a roof claim file?

Enough to cover the roof continuously and prove each scope item, not a fixed count. Forty captioned, located, sequenced photos that march cleanly through the roof beat 200 random ones, because a desk reviewer can follow them. Cover property baseline, ground collateral, each slope (overview, continuous mid-range, damage close-ups with scale), and the scope drivers: ventilation, penetrations, flashing, drip edge, layers, decking, and pitch.

Why does photo metadata matter for a claim?

The EXIF data inside a photo — capture time, GPS location, device — is part of the proof that the evidence is contemporaneous, meaning captured at the time of inspection rather than reconstructed later. Screenshotting, filtering, or recompressing a photo on the way through email strips that metadata and weakens the file. Always move and keep the original images, not processed copies.

How do I connect an estimate line item to my evidence?

Build a cross-reference table running line-item first: the estimate line, the quantity basis, the supporting exhibit IDs, and any code or manufacturer-spec basis. Follow two rules: no orphan lines (every line needs at least one exhibit or a cited code reason behind it), and every quantity traces to the measurement report or a stated basis. A reviewer's core question on each line is 'what supports this' — the cross-reference answers it before they ask.

Can a contractor put weather data in the evidence index?

Yes, as a timeline exhibit with the source named. Authoritative weather records corroborate that a hail or wind event occurred near the property around the date of loss. They do not prove that this specific roof was damaged by that storm — hail swaths are patchy. Let the physical evidence (impact marks, directional pattern, collateral on soft metals) carry the damage finding, and let the weather data carry only the timeline. Captions that claim 'this storm caused this damage' overstate what the record shows.

What's the difference between damage photos and scope photos?

Damage photos justify that a claim exists — impact marks, fractures, granule loss, collateral. Scope photos justify what the repair must include — existing layers, decking, flashing condition, drip edge, ventilation count, valley type, and measured pitch. Both belong in the index. A reviewer who can see the existing corroded drip edge in a photo is far less likely to strike your drip-edge line, even though the storm didn't damage the drip edge directly.

Why are some estimate lines justified by code instead of damage?

Many legitimate lines exist because the adopted building code requires the item when you re-roof, not because the storm damaged that component. Drip edge, ice barrier in cold climates, deck attachment, and underlayment requirements are common examples. For those lines, the code citation is the support, so the index must contain the code text or a link to it. You document that the code applies and that the existing condition doesn't meet it — you don't argue coverage.

Where is the line between documenting a claim and illegally handling it?

A contractor may inspect, document condition, write an accurate repair estimate for their own scope, state facts about that scope, and hand the documentation to the homeowner to file. A contractor may not, for a fee, negotiate or 'handle' the claim, interpret the policy, promise a specific payout or approval, tell a homeowner the deductible will disappear, advertise a 'free roof,' or represent the homeowner against the insurer — that is unlicensed public adjusting. The evidence index stays entirely on the documentation side. Run every caption through one test: does it describe what you observed or will do (safe), or predict what the insurer will pay (not safe)?

How does RoofPredict help build the evidence index?

RoofClaim inside RoofPredict links the claim to the home, then lets you upload and auto-classify the carrier's documents with OCR so their line items become text you can compare against. Its opportunity detection maps the carrier estimate to a roofing knowledge base and flags likely scope gaps, missing code-required items, and missed supplements with evidence anchors and pricing — which you review and decide on. It scores packet completeness before you send, and runs recoverable-depreciation and supplement cadence so deadlines don't slip. It organizes evidence and surfaces candidates; it does not decide coverage, file the claim, or talk to the carrier.

Why do supplements get lost even when the damage was real?

Most lost supplements weren't rejected — they were forgotten. An open supplement with no follow-up cadence ages out of attention while the office moves on to new roofs. The same happens to recoverable depreciation when the completion evidence and final invoice lapse. The fix is a tracked cadence with a clock and an owner for every open item, so work that's already done actually gets billed and released instead of left on the table.