Supplement Software vs Writing Supplements In House: What Actually Pays Off for a Roofing Company

Every roofing company that touches insurance work hits the same wall around its 30th or 40th job a month: the front end is producing approved jobs faster than the back end can document and bill them correctly. The estimate that came back from the carrier is short. There are line items the crew built that nobody captured. Profit is leaking out the bottom of jobs that looked great on the contract. And the owner is staring at three options to fix it: buy supplement software, hire and train someone in house, or ship the whole problem to an outsourced supplement service that takes a cut.

The pitch decks make this sound like a clean three-way choice. It isn't. The real decision is about where you want the labor, the expertise, and the liability to live, and how much volume you have to spread the cost across. I've watched companies pick the wrong option for their stage and bleed margin for a year before they figured it out. The goal here is to give you the operational math and the trade-offs so you pick right the first time.

Before we go further, one boundary that matters and that a lot of online advice gets dangerously wrong. A roofing contractor documents the work, measures the roof, and writes an accurate, code-aligned repair estimate for its own scope. The contractor hands that documentation to the homeowner. The homeowner files the claim and the insurer decides what is covered. A contractor does not, for a fee, negotiate the claim, interpret the homeowner's policy, promise an approval or a payout, promise a deductible will disappear, or advertise a "free roof." That last set of activities is unlicensed public adjusting in most states, and it is the fastest way to turn a margin problem into a license-and-lawsuit problem. Everything below assumes you're operating on the document-and-estimate side of that line. We'll come back to it, because it changes how you should evaluate every tool and every hire.

What a supplement actually is, in plain terms

A supplement is a request to revise a repair estimate after the initial scope was written, because the initial scope missed something real. Note the framing: you are revising your estimate. You are not revising the homeowner's claim or the insurer's coverage decision. You document a condition or a code requirement, you put it on an Xactimate-aligned line item with the correct quantity and the carrier's price list, and you give that documentation to the homeowner to submit. The insurer then decides whether it's covered.

The reason supplements exist is that the first estimate is almost never complete. The adjuster spent twenty minutes on the roof, often from the ground or off a drone image, working from a measurement report and a mental checklist. They are not on site when the crew tears off three layers and finds rotted decking. They didn't price the specific drip edge your code jurisdiction requires. They left off the ridge vent because the photo didn't show it clearly. None of that is malice. It's the nature of a fast, remote-ish first pass.

So the supplement is the document that closes the gap between what the first estimate said and what the roof actually needs to be repaired to code and to a workmanlike standard. Done right, it is a factual, photo-backed, line-item description of your scope. Done wrong, it is a sloppy ask that gets denied, slows the job, and trains the carrier to distrust your estimates.

The most common real-world supplement line items, the ones that recur on a huge share of jobs:

• Decking replacement beyond what the original scope allowed, documented by tear-off photos and board count.
• Drip edge / rake edge when the jurisdiction's adopted code requires it and the original estimate omitted it.
• Ice-and-water shield at eaves and valleys per code in cold-climate jurisdictions.
• Ridge vent and ridge cap when the roof has a ridge system the first scope undercounted.
• Step and counter flashing at walls and chimneys, frequently omitted or underpriced.
• Steep / high charges that the original measurement missed because the pitch or stories were wrong.
• Detach and reset of items like satellite dishes, solar attachments, or gutters that the crew has to remove and reinstall.
• Code-required upgrades such as proper underlayment, valley metal, or fastener patterns tied to the locally adopted edition of the building code.

Every one of those is defensible as your scope when you have the photo, the measurement, and the code citation. None of them is defensible when you're just guessing or padding. The whole game of supplementing well is documentation discipline, not argument.

The three operating models, defined honestly

People say "software vs in house" but there are really three models in the market, and the comparison only makes sense once you separate them.

Model A — Software-assisted, written in house. You buy a tool (estimating platform, photo-and-report software, a measurement service, an Xactimate license or equivalent) and one of your own people writes the supplements using it. The software speeds the work; a human on your payroll still owns it.

Model B — Fully in house, minimal tooling. A trained estimator or supplement coordinator on your staff writes supplements largely by hand inside the estimating platform, leaning on experience more than on automation. Common in shops that grew up doing this before the software wave.

Model C — Outsourced supplement service. A third-party firm writes your supplements for a fee, usually a percentage of the supplemented amount or a flat per-file fee. Your crew uploads photos and measurements; their estimators do the documentation and write-up.

When a sales rep says "supplement software vs writing supplements in house," they're usually pitting Model A against Model B, or sometimes Model A against Model C. The honest version is a 3x3 comparison across cost, accuracy, speed, control, and risk. Let's build that out, then give you a decision framework.

The cost math, with real numbers

This is where most decisions get made on vibes instead of arithmetic. Let's do the arithmetic. I'll use illustrative figures you can swap for your own; the structure is what matters.

Model B: a dedicated in-house supplement coordinator

Say you hire one supplement coordinator. According to the U.S. Bureau of Labor Statistics, claims adjusters and examiners and related estimating roles sit in a wage band that, depending on region and experience, commonly lands a competent supplement writer somewhere in the range of a mid-five-figure base salary. Add payroll taxes, benefits, software seat, and overhead, and the fully loaded annual cost of one good coordinator realistically runs well into the five figures and sometimes past six figures in high-cost metros.

Let's model it conservatively:

Now the per-file cost depends entirely on volume. A capable coordinator with good tooling can handle a meaningful number of files. If they process 400 supplemented jobs a year, your cost is roughly $219 per job. If they only process 150, it's about $583 per job — and you're paying for idle capacity. If you push them to 600 with the right software, it drops toward $146 per job, but quality starts to suffer past a person's real throughput ceiling.

That sensitivity is the whole point. In-house labor is a fixed cost you must amortize over volume. Below a volume threshold, it's the most expensive option per job. Above it, it's the cheapest. Find your break-even before you hire.

Model C: outsourced supplement service

Outsourced firms typically charge either a percentage of the approved supplement amount (commonly in the range of 8 to 15 percent of what gets added) or a flat fee per file (often somewhere from $50 to $300+ depending on complexity). Percentage pricing aligns their incentive with getting more added — which sounds good and is also exactly why you need to keep them on the document-and-estimate side and never let the model drift into anything that looks like adjusting the claim.

Let's model percentage pricing. Suppose your average supplement adds $2,800 of legitimate, documented scope to a job, and the service charges 12 percent:

• Cost per file: $336
• At 400 jobs/year: $134,400/year
• At 150 jobs/year: $50,400/year
• At 50 jobs/year: $16,800/year

The beautiful property of Model C: it's purely variable. No file, no cost. At low volume it's cheaper than carrying a salaried coordinator. The ugly property: at high volume, you're paying a percentage forever on work a salaried person could do for a flat cost. Cross a volume threshold and you're handing a third party tens of thousands a year that a $87k employee could absorb.

Model A: software-assisted in house

This is the hybrid. You're still paying a person, but the software lets a less specialized person (or fewer people) handle more files, or lets your existing production manager handle supplements as part of their role without a dedicated hire. Software pricing varies widely — per-seat subscriptions, per-report measurement fees, per-claim processing fees, or bundled platforms. Model the cost as: software subscription + per-report fees + a fraction of an existing person's time.

If software lets your production coordinator (already on payroll at, say, $65k loaded) spend 30 percent of their time on supplements and handle 250 files a year with a $6,000/year software-and-reports stack:

• Allocated labor: 30% of $65,000 = $19,500
• Software + reports: $6,000
• Total: $25,500/year over 250 files = ~$102 per file

That's the cheapest per-file number in the whole comparison — if the software genuinely lets a non-specialist produce carrier-credible documentation, and if you actually have an existing person with 30 percent of their week free. Those two ifs are where Model A succeeds or fails.

Putting the cost models side by side

Read the table as a story. At very low volume, Model C wins because it has no fixed cost. As volume climbs, Model A (software leveraging shared labor) is the value sweet spot — until your volume outgrows what one shared person can handle. At that point you either go full Model B (dedicated coordinator, now well-utilized) or scale Model A with more seats. Outsourcing stays flat per file regardless, which is great at the bottom and expensive at the top.

The single biggest mistake I see: companies hire a dedicated coordinator (Model B) at 80 jobs a month of total volume where only 40 actually need supplements, then wonder why the math feels bad. They're paying for a full-time role at a volume where the per-file cost is brutal. Either software-leverage that person across other production duties (move to Model A) or outsource until volume justifies the dedicated seat.

Accuracy and approval quality: where the real money is

Cost per file is the obvious axis and the least important one. A supplement that's $200 cheaper to produce but gets denied, or gets approved at 60 percent of what was justified, costs you far more than the production savings. The question that actually moves margin is: which model produces the most accurate, best-documented, most-approvable scope? And the honest answer is: it depends less on the model and more on the discipline behind it.

Here's what drives approval quality, in roughly the order that matters:

1. Photo documentation completeness. Did the crew capture tear-off photos, decking condition, flashing details, measurements at the right moments? No tool and no service can document a condition that was never photographed. This is a field-process problem, and it's the number-one reason supplements get denied or stalled. Garbage in, garbage out, regardless of who writes it.
2. Correct price list and current code. Using the carrier's correct regional price list for the date of loss, and citing the locally adopted edition of the building code for code-required items, is what makes a line defensible. The International Residential Code is adopted (with amendments) jurisdiction by jurisdiction, so "the IRC says" isn't enough — you need your jurisdiction's adopted edition and local amendments.
3. Line-item accuracy. Right quantity, right unit, right item code. A drip edge measured at 180 LF when the roof has 210 LF of eave-and-rake is a self-inflicted denial.
4. Clean, factual narrative. Each supplemented item paired with the photo and the reason, written as a description of your scope — not as an argument about coverage. "Tear-off revealed 14 sheets of delaminated decking; photos 22–35; replacement required to provide a sound substrate per manufacturer installation requirements" is defensible. "Adjuster missed this, you owe us" is not.
5. Consistency over time. Carriers track patterns. A contractor whose estimates are consistently accurate and well-documented gets faster, smoother handling. A contractor known for padding gets scrutiny on everything.

Notice that none of the top five are "which software you bought" or "whether you outsourced." They're about field discipline and documentation standards. The model you choose affects accuracy mostly through who owns these five things and how consistently they execute.

How each model tends to perform on accuracy

Model B (dedicated in-house): Highest ceiling, because a trained coordinator who knows your crews, your jurisdiction's code amendments, and your typical jobs builds deep pattern recognition. They know that your usual tear-offs in the older part of town hit decking, so they make sure the crew shoots those photos. The ceiling is high; the floor depends entirely on whether you hired and trained well. A weak in-house writer is worse than good software.

Model A (software-assisted): Most consistent floor, because good software encodes the checklist. It prompts for the photos, applies the current price list, flags missing code items, and standardizes the narrative. It raises a mediocre writer to a competent one. Its ceiling is a bit lower than a true expert, because software follows rules and a veteran reads situations. But for most shops, raising the floor matters more than raising the ceiling.

Model C (outsourced): Variable, and you don't control it. A good service has experienced estimators and produces clean work; a cheap one runs a template mill. Critically, the service's accuracy is capped by your field documentation — they can only write what your crew photographed. The best outsourced relationships include a feedback loop where the service tells your crews exactly what photos they're missing. The worst ones just write what they can and let the rest get denied.

Speed, control, and the things nobody puts in the brochure

Turnaround time

Supplement turnaround directly affects cash flow and job close-out. A supplement that sits for three weeks is a job you can't fully invoice and a balance you can't collect.

• Model A is usually fastest once dialed in, because the work happens internally on your schedule with software handling the repetitive parts.
• Model B is fast when the coordinator isn't buried; it stalls when one person becomes a bottleneck and is out sick or swamped.
• Model C introduces a hand-off cycle: you upload, they queue, they write, you review, they revise. A good service turns it fast; a busy one queues you behind everyone else, and you have zero ability to expedite.

Control and institutional knowledge

This is the most underrated axis. With Model B and Model A, the knowledge of how to document and estimate your scope lives inside your company. Every job makes your team better at it. With Model C, that expertise compounds inside the vendor, not you. If you outsource for three years and then the relationship ends or their pricing climbs, you're starting from zero on a core competency. Outsourcing a back-office task is fine; outsourcing a profit center that you never learn to run yourself is a strategic risk.

There's also a softer control issue: a third party writing in your company's name is documenting your scope to your carriers. You are responsible for what goes out under your estimate. If their write-up drifts toward anything that looks like adjusting the claim, interpreting the homeowner's coverage, or promising an outcome, that exposure is yours, not theirs. Which brings us to the part that should govern every one of these choices.

The compliance line that governs all three models

Whichever model you pick, the activity has to stay on the document-and-estimate side. This is not legal advice and your state's rules are the authority — check your state Department of Insurance — but the pattern is consistent nationwide and it is worth burning into every coordinator's brain and every vendor contract.

What you (or your software, or your service) MAY do:

• Inspect the roof and document its condition with photos and measurements.
• Write an accurate, code-aligned repair estimate for your own scope of work, using the correct price list.
• State facts about your scope to the carrier ("we found 14 sheets of bad decking; here are the photos").
• Hand that complete documentation to the homeowner so the homeowner can submit it.

What you (or your software, or your service) MAY NOT do, for a fee, on the homeowner's behalf:

• Negotiate, adjust, or "handle" the claim with the insurer.
• Interpret the homeowner's policy or tell them what their coverage means.
• Promise a specific payout, a specific approval, or that the claim "will be covered."
• Promise that the deductible will be waived, absorbed, eaten, or made to disappear.
• Advertise or imply a "free roof."
• Represent the homeowner against the insurer.

That last block is, in most states, unlicensed public adjusting, and it carries real penalties. The Federal Trade Commission's guidance on truthful advertising also bites here: a "free roof" or "we'll waive your deductible" pitch is both a licensing problem and a deceptive-advertising problem.

Why does this belong in a piece about software vs in house? Because the model you choose changes who's tempted to cross the line. A percentage-based outsourced service has a built-in incentive to maximize the supplemented amount; make sure their incentive never turns into adjusting the claim or promising outcomes in your name. An in-house coordinator under pressure to "get it approved" might start calling adjusters to "work the claim" — keep them documenting scope, not negotiating coverage. Good software helps here by structuring output as factual scope documentation rather than coverage arguments. When you evaluate any tool or vendor, ask explicitly: does this keep us on the document-and-estimate side, or does it nudge us toward acting like an adjuster? If it nudges, walk.

A compliance-safe phrasing cheat sheet

Make this cheat sheet part of onboarding for any coordinator and part of the contract with any service. It protects the company regardless of which model writes the supplement.

A real worked example: the same job, three ways

Let's run one realistic job through all three models so the trade-offs stop being abstract.

The job. Two-story, 28-square architectural shingle replacement after a hailstorm. Initial carrier estimate came back at $14,200. The crew tears off and finds: 12 sheets of delaminated decking, the original scope omitted code-required drip edge on 195 LF of eave and rake, no ice-and-water shield was included despite the cold-climate jurisdiction requiring it at eaves, the steep charge was missing because the original measurement called it one story, and the ridge vent (62 LF) wasn't counted. All of it is real, photographed, and code-backed.

Model B — in-house coordinator. Your coordinator knows this neighborhood hits decking, so the crew already shot 30 tear-off photos. She pulls the current regional price list, cites the jurisdiction's adopted code edition for the drip edge and ice-and-water, recounts the measurement for the steep charge, and writes five clean line items with photo references. Documented supplement: roughly $3,900 of additional scope. Time: about 45 minutes of her day. Loaded cost of that 45 minutes at her salary: maybe $35. She hands it to the homeowner to submit.

Model C — outsourced service. You upload the photos and measurement. Their estimator writes a clean supplement — but they weren't on site, so they only document what the photos clearly show. The ridge vent photo was ambiguous, so they leave it off rather than risk a denial. Documented supplement: about $3,400. Their fee at 12 percent: roughly $408. You netted $3,400 of scope minus $408 = ~$2,992 added, and you lost the $500 ridge-vent item to a documentation gap nobody flagged in the field.

Model A — software-assisted, written by your production coordinator. The software's checklist prompted the crew to shoot decking, flashing, and ridge photos, and flagged ice-and-water as a code item for the jurisdiction. Your production coordinator, who isn't a supplement specialist, follows the software's structured workflow, applies the current price list automatically, and produces all five line items plus the ridge vent. Documented supplement: about $3,900. Cost: roughly $20 of his time plus the per-file software cost of maybe $25.

The scoreboard:

One job won't decide your model. But multiply the gap by 300 jobs a year and the differences are enormous — and notice the real story isn't "software vs human." It's that the field documentation drove the outcome, and Models A and B captured the full scope because the checklist (software) or the experienced coordinator (human) made sure the photos existed. Model C lost money to a documentation gap, not to bad writing. Fix the field process and any model improves.

The decision framework: which model for which company

Use volume, complexity, and your team as the three inputs.

Step 1 — Count your real supplement volume

Not your total job count. Count jobs that actually need a supplement. A useful rule of thumb: in storm-and-insurance work, a large share of jobs need at least one supplemented line. Track it for 60 days. If you're under ~10 supplemented files a month, you're low-volume. 10–25 is mid. 25+ is high.

Step 2 — Match volume to model

• Under 10/month: Outsource (Model C) or software-assist an existing person (Model A). Do not hire a dedicated coordinator yet — the per-file math is brutal at this volume.
• 10–25/month: Software-assisted shared labor (Model A) is usually the sweet spot. You likely already have a production person with capacity, and software makes them productive on supplements without a specialist hire.
• 25–50/month: You're at the threshold for a dedicated coordinator (Model B), ideally still software-equipped. The per-file cost of a salaried person finally beats the outsourced percentage.
• 50+/month: Dedicated team, software-equipped, possibly two people. Outsourcing this volume means paying a percentage on hundreds of thousands of dollars of scope you could process for a flat payroll cost.

Step 3 — Adjust for complexity and team

• If your jobs are complex (lots of steep, lots of code-upgrade jurisdictions, commercial), weight toward in-house expertise (B or software-heavy A), because templated outsourcing struggles with non-standard scope.
• If you lack anyone who understands estimating, start outsourced (C) while you build the knowledge, then bring it in house once you've learned the patterns. Don't start fully in-house with nobody who's done it.
• If you have a sharp production person with spare capacity, Model A is almost always your best dollar.

Step 4 — Decide what you refuse to outsource

Even if you outsource the writing, never outsource the field documentation standard or the compliance line. Those stay in house, owned by you, always. The photos, the measurement discipline, and the rule that you document scope rather than handle claims — those are non-negotiable company standards no matter who types the line items.

The lever everyone ignores: only supplement the roofs worth your time

Here's the strategic point that sits underneath the whole software-vs-in-house debate, and almost nobody frames it this way. The cheapest, most accurate supplement is the one on a job you should have been chasing in the first place. A huge amount of supplement pain comes from doing insurance work on roofs that were marginal candidates — roofs where the damage is ambiguous, the age is uncertain, and every line item is a fight. The further upstream you fix targeting, the less your back office has to grind.

This is where knowing which roofs are actually due changes the economics. If your crews are knocking and inspecting roofs that are genuinely aging out and genuinely sat under a real storm, the documentation is cleaner, the conditions are clearer, and the supplements are more straightforward to write and defend — because there's actually decking to replace and there are actual code gaps to close. When you instead chase every door, you generate a pile of weak, contested files that bury whatever supplement model you chose.

This is the part of the problem RoofPredict is built for. It tells roofing contractors which roofs are due house by house: a roof-age range per address read from aerial imagery, plus storm physics modeled per roof, so you can rank doors, routes, and mailing lists toward the roofs a storm actually wore out and the roofs aging out on their own. It also enriches your own CRM or list with roof-age and storm signals, so the targeting lives in the tools you already use. Be clear about the honest limits: roof age comes back as a range, not a build date, and the storm model gives you odds, not proof that a specific roof was hit. It points you at the right doors; your crew and your camera still do the documenting. But pointing the truck at better roofs upstream is the single highest-leverage thing you can do to make whatever supplement model you choose work harder — better candidates in, cleaner files out.

It is not a supplement tool and won't write a line item for you. What it does is shrink the share of your jobs that are marginal, contested, and supplement-heavy for the wrong reasons, so your coordinator or software or service spends its time on roofs that genuinely deserve the scope.

Implementation: a 30-day rollout for each model

Whatever you pick, don't drop it in cold. Here's a tight rollout.

If you're going software-assisted (Model A)

1. Week 1: Pick the tool. Demo at least two. Score them on: does it apply the correct current price list automatically, does it prompt the crew for the right photos, does it keep output as factual scope documentation (not coverage arguments), and does it integrate with your CRM and measurement source.
2. Week 2: Build your photo checklist into the field app. This is the real work. Decking, flashing, ridge, eaves, valleys, measurement shots — make them required before a job is marked ready for supplement.
3. Week 3: Run 10 real jobs through it in parallel with your old process. Compare documented scope and turnaround.
4. Week 4: Cut over. Set a standard: every supplemented file gets a photo-completeness check before it goes out.

If you're hiring in house (Model B)

1. Before hiring: Confirm your volume justifies the seat (25+/month). Write the role around documentation and estimating, explicitly not claim handling.
2. Week 1–2: Train on your jurisdiction's adopted code edition and amendments, your carriers' price lists, and your compliance cheat sheet. The compliance training is not optional.
3. Week 3: Shadow real tear-offs so the coordinator understands what the crew sees and what photos they need.
4. Week 4: Own the queue, with you reviewing every file for the first month.

If you're outsourcing (Model C)

1. Selection: Interview the service like a hire. Ask how they handle missing photos (do they flag your crew or just write what they can?), how they keep output on the document-and-estimate side, and what their pricing does at your volume.
2. Contract: Put the compliance line in writing — they document scope, they do not negotiate, interpret coverage, or promise outcomes in your name.
3. Field loop: Demand a feedback mechanism so they tell your crews what photos are missing. Without it, you'll bleed scope to documentation gaps forever.
4. Review: Audit their first 20 files for accuracy and compliance before you trust the volume.

Building the documentation standard that makes any model work

Because field documentation is the variable that decides accuracy in every model, it's worth spelling out the standard itself. This is the asset you own no matter who writes the line items, and it's the highest-return thing you can build. Treat it as a written company standard, trained into every crew, enforced before any file is marked ready.

A workable field documentation standard has five parts:

1. A required photo set per job, by phase. Pre-tear-off (overall roof, each slope, existing flashing, penetrations, ridge), during tear-off (decking condition, any rot, layer count, fastener patterns), and post-tear-off (replaced decking, installed flashing, ice-and-water placement). Each phase has a checklist the crew can't skip in the field app.
2. A measurement source of record. One authoritative measurement report per job, captured the same way every time, so the steep charges, squares, eave-and-rake lengths, and ridge counts are never guesses. Most denied steep and quantity items trace back to a sloppy or missing measurement.
3. A code reference per jurisdiction. Maintain a short internal sheet of the adopted code edition and key local amendments for each jurisdiction you work in, so the coordinator or software cites the edition your county actually enforces, not a generic national reference. This is the single most common defensibility gap on code-upgrade items.
4. A price-list date discipline. Always tie the estimate to the correct regional price list for the date of loss. A line priced on the wrong list is a self-inflicted denial that no model can save.
5. A factual narrative template. Each supplemented item gets a one-line description that states the condition, the photo references, and the reason it's part of your scope. The template keeps the language on the document-and-estimate side and out of coverage-argument territory.

Build those five and your supplements get cleaner regardless of model, because every model is just turning your documentation into line items. Skip them and even the best coordinator or the priciest software is writing from incomplete information. The standard is the foundation; the model is the machine that runs on top of it.

A useful internal metric to track once the standard is live: photo-completeness rate at hand-off. Measure the percentage of jobs that arrive at the supplement step with the full required photo set. If that number is below 90 percent, fix the field process before you spend a dollar comparing software to in-house writers, because the documentation gap is costing you far more than the model choice ever will.

The hidden costs nobody quotes you

The pitch decks compare subscription price to salary and stop there. The costs that actually decide whether a model works show up later, off the price sheet. Build them into your comparison before you sign anything.

Rework and revision cycles. A supplement that comes back for revision costs you twice: the original write-up plus the redo, plus the days the job's balance sat uncollected. Software with a strong photo-completeness check and a current price list cuts the revision rate; a cheap outsourced template mill quietly inflates it. When you price a model, ask what its first-pass acceptance rate looks like, alongside its per-file cost. A model that's $50 cheaper per file but doubles your revision rate is more expensive in practice.

Carrier relationship erosion. This one never shows up on an invoice and costs the most. Carriers track which contractors submit accurate, well-documented estimates and which submit padded or sloppy ones. A contractor with a clean track record gets faster, smoother handling on every future file. A contractor flagged for padding gets every line scrutinized, which slows every job and raises your effective cost per supplement across the whole book. Whichever model you choose, it has to protect that reputation. A percentage-based service with a padding habit can quietly torch a relationship you spent years building.

Cash-flow drag. Every day a supplement sits unwritten or unapproved is a day a job balance can't be invoiced and collected. At any real volume, turnaround speed translates directly into working capital. A model that's slightly more expensive per file but turns supplements in days instead of weeks can be cheaper once you account for the cash you're not floating.

Training and ramp time. A new in-house coordinator isn't fully productive on day one. Budget a ramp period where their output is lower and you're reviewing every file. Software-assisted models shorten this because the tool carries part of the expertise; a fully manual in-house model has the longest ramp.

Switching costs. If you start outsourced and later bring it in house, you're rebuilding a skill from scratch with no internal expertise to lean on. That migration has a real cost. It's a reason to keep at least some documentation knowledge in house even while you outsource the writing.

A simple scorecard for choosing

When you can't decide, score the three models on the axes that matter and weight them for your situation. Here's a starting scorecard; adjust the weights to your business.

Score each cell 1 to 5, multiply by the weight, total it. The point isn't the exact number; it's forcing yourself to weigh control and accuracy instead of fixating on the price tag. Most companies that do this honestly find the middle-volume answer is software-assisted shared labor, because it scores well on cost, keeps knowledge in house, and raises the accuracy floor without a specialist hire.

One more practical note: these aren't permanent. The smart practice is to re-run the math every six months as your volume changes. The shop doing 8 supplements a month that correctly outsourced last year may be doing 30 this year and bleeding percentage fees it should now convert to a salaried, software-equipped seat. The decision is a function of your current volume, not a one-time identity.

What pros get wrong

A short list of the expensive mistakes, drawn from watching companies make them:

• Hiring a coordinator at the wrong volume. Paying a salary at 8 supplements a month is the most common money-loser. Software-assist an existing person until volume justifies the seat.
• Blaming the writer for a field problem. Most denied supplements die from missing photos, not bad writing. Fix the field checklist before you switch tools or fire the coordinator.
• Letting an outsourced service operate as a black box. If you never see what's going out in your name, you've lost control of both quality and compliance.
• Drifting into adjusting. The moment anyone — staff or vendor — starts "working the claim," interpreting coverage, or promising approvals or a waived deductible, you've traded a margin problem for a license problem. Document scope; let the homeowner file and the insurer decide.
• Supplementing your way out of a targeting problem. If half your jobs are marginal candidates, no supplement model saves you. Fix which roofs you chase upstream.
• Using stale price lists or the wrong code edition. A line item priced on last year's list or citing a code edition your jurisdiction hasn't adopted is a self-inflicted denial.
• Outsourcing the knowledge forever. Fine to outsource while you learn. Risky to never learn a core profit-center skill.

So, which one?

If I had to compress the whole decision into a few sentences: at low volume, outsource or software-assist an existing person, because a dedicated salary can't be justified. In the broad middle, software-assisted shared labor is the best dollar most companies will spend, because it raises the floor on quality and keeps the knowledge and the cost inside your shop. At high volume, bring it fully in house with a software-equipped, well-trained coordinator, because the per-file math finally favors the salaried seat and the institutional knowledge compounds.

But the model is the second decision. The first one is field documentation discipline, which determines accuracy in every model, and the targeting that decides whether your back office is grinding on strong files or weak ones. Get the right roofs in the front of the funnel and document them completely in the field, and the supplement-software-vs-in-house question gets a lot smaller — because clean candidates and complete photos make almost any model work, and no model can save a pile of marginal, under-documented jobs.

If the upstream targeting is the part you've never had real data on, that's exactly where a per-address roof-age range plus storm modeling earns its keep: point the trucks at the roofs that are genuinely due, document them properly, write accurate scope, and let the homeowner file. Everything downstream gets easier.

FAQ

Is supplement software or writing supplements in house cheaper?

It depends entirely on volume. Below roughly 10 supplemented jobs a month, software-assisting an existing employee or outsourcing is cheaper, because a dedicated salary can't be amortized. Above about 25 a month, a software-equipped in-house coordinator usually wins on a per-file basis. Run your own numbers: fully loaded labor cost divided by files handled versus the software-plus-shared-labor cost versus an outsourced percentage.

What is the break-even volume for hiring a dedicated supplement coordinator?

Most shops find the math turns favorable somewhere around 25 to 50 supplemented files a month, when a fully loaded salary divided by files handled drops below what an outsourced service charges per file or what a percentage model costs. Track your true supplement volume for 60 days before hiring, and count only jobs that actually need a supplement, not total jobs.

Does supplement software produce more accurate supplements than a person?

Software raises the floor by encoding the checklist, applying the current price list, and prompting for required photos, which turns a non-specialist into a competent writer. A veteran in-house estimator has a higher ceiling because they read non-standard situations software can't. But accuracy is driven mostly by field photo completeness and using the correct price list and adopted code edition, not by who types the line items.

Why do supplements get denied, and can software fix it?

The most common reasons are missing or incomplete field photos, wrong quantities, stale price lists, and citing a building code edition the jurisdiction hasn't adopted. Software helps with price lists, quantities, and photo prompts, but it cannot document a condition that was never photographed. Fixing your field documentation checklist usually does more for approval rates than switching tools.

What is the risk of outsourcing roofing supplements?

Three risks: the expertise compounds inside the vendor instead of your company, so you never learn a core skill; their output quality is capped by your field documentation and you have less control over turnaround; and a percentage-based service has an incentive that must be kept strictly on the document-and-estimate side. Put the compliance line in the contract and audit their first files.

Can a roofing contractor negotiate the supplement with the insurance company?

A contractor can state facts about its own scope and provide documentation, but negotiating, adjusting, or handling the claim on the homeowner's behalf for a fee is unlicensed public adjusting in most states. The safe workflow is: document the damage, write an accurate code-aligned estimate of your scope, hand it to the homeowner, and let the homeowner file while the insurer decides coverage. Check your state Department of Insurance for specifics.

Can I tell a homeowner their deductible will be covered or that the roof is free?

No. Promising a waived or absorbed deductible, advertising a 'free roof,' or promising a specific approval or payout is both a licensing problem (it looks like adjusting) and a deceptive-advertising problem under FTC guidance. Say instead that the deductible is the homeowner's responsibility per their policy and that coverage is the insurer's decision based on the documentation you provide.

What photos does a crew need so any supplement model works?

At minimum: tear-off and decking condition, all flashing details (step, counter, headwall, valley), ridge and ridge vent, eaves and ice-and-water areas, penetrations, and measurement reference shots. Make these required before a job is marked ready for supplement. No software or service can document a condition that was never photographed, so the field checklist is the foundation of every model.

How does better roof targeting reduce supplement workload?

When crews inspect roofs that are genuinely aging out and genuinely sat under a real storm, conditions are clearer and code gaps are real, so supplements are cleaner and easier to document and defend. Chasing marginal candidates generates contested, under-documented files that bury any supplement model. Tools like RoofPredict point trucks at roofs that are likely due using a per-address roof-age range plus storm modeling, which improves the candidates your back office works on.

Can RoofPredict write my supplements?

No. RoofPredict is a targeting and list-enrichment tool that tells contractors which roofs are likely due using a roof-age range per address and storm physics modeled per roof. It does not write line items or document a specific roof's condition. Its honest limits: roof age is a range not a build date, and storm modeling gives odds not proof. It improves the roofs in your funnel; your crew and camera still document, and your coordinator, software, or service still writes the scope.