How to Measure Supplement Cycle Time on Roofing Jobs (and Cut It)

If you run a restoration roofing operation, money is leaving the building in a way you probably can't see on your P&L. It isn't theft and it isn't margin erosion. It's time. A supplement that takes 9 days to document, package, and submit is a different business than one that takes 47 days, even if both eventually get the same number back. The slow one ties up cash, buries your supplement manager in follow-up, lets jobs fall out of memory, and quietly trains your crews to stop documenting because "it never gets paid anyway."

Most roofers can tell you their close rate and their average ticket. Very few can tell you their supplement cycle time, and almost none can tell you where inside that cycle the days actually disappear. That blind spot is expensive. Cycle time is the one number that converts a vague feeling of "our paperwork is slow" into a measurable, fixable operation with named bottlenecks.

A few ground rules before the math, because the topic sits next to insurance and the line matters. As a contractor, your lane is documenting the condition of the roof, writing an accurate repair estimate for your own scope of work, and handing a clean, evidence-backed package to the homeowner. The homeowner files. The carrier decides coverage. You are measuring how fast your documentation moves, not how fast you can make an adjuster agree, and certainly not how fast you can "get a claim approved." Keep that frame and everything below is both legal and genuinely useful. Cross it and you're doing unlicensed public adjusting. We'll mark the boundary clearly as we go.

What "supplement cycle time" actually means

Cycle time is elapsed calendar time between two events you define. That's the whole concept. The art is in choosing the right two events and then being ruthlessly consistent about stamping them.

The naive version most shops use is "how long until we get paid." That single number is nearly useless on its own, because it blends together delays you control (writing the estimate, gathering photos, building the packet) with delays you don't (the carrier's review queue, the mortgage company holding depreciation, the homeowner sitting on a check). If you only track the outer bookends, every problem looks like "the insurance company is slow," and you'll never fix the half of the cycle that lives inside your own four walls.

So measure cycle time as a chain of timestamps, not a single span. Here's the spine I recommend, with the event that starts each clock:

Notice the deliberate split at "Packet-to-submit." You assemble the packet; the homeowner submits it. That handoff is where contractors who blur the line get into trouble, and it's also a real measurable delay (homeowners sit on things). Keep the events separate so you can see that wait without ever stepping into the homeowner's role against the insurer.

The two stages you fully own are detection through packet QC. That's your operation. If you can only instrument part of this chain, instrument that part first, because it's the part you can actually change.

Why a single "days to paid" number hides your real problem

Picture two shops. Shop A reports a 40-day average days-to-paid. Shop B reports 38. On the surface, Shop B is winning. But break the chain apart and the story flips. Shop A's 40 days is 4 days of documentation, 3 days of estimate-and-packet, 28 days of carrier review (a slow carrier mix), and 5 days to cash. Shop B's 38 days is 14 days of documentation, 6 days of estimate-and-packet, 12 days of carrier review, and 6 days to cash. Shop B has a faster carrier mix and a far slower operation — 20 controllable days versus Shop A's 12. If both shops only watch the outer number, Shop B congratulates itself while bleeding two extra weeks on every job from pure internal drag. The blended number lied to both of them. Only the per-stage split tells the truth, and the truth is the only thing you can act on.

This is also why benchmarking your total days-to-paid against another company is close to meaningless. Their carrier mix, their storm region, their state's regulatory environment, and their job mix are all different. The only honest benchmark is your own controllable stages, this month versus last month. Compare yourself to yourself, stage by stage, and you'll improve faster than any industry average could ever guide you.

Cycle time vs. lead time vs. aging

Three terms get used interchangeably and they are not the same:

• Cycle time is active, per-stage elapsed time on one job. It answers "how long does this step take when we work it."
• Lead time is the full outer span a customer or your cash flow experiences, detection to money. It includes all the waiting.
• Aging is a snapshot of how long currently-open items have been sitting right now. It's the photograph; cycle time is the film.

You need all three. Cycle time tells you where to fix the process. Lead time tells you what cash conversion really looks like. Aging tells you which specific jobs are on fire today. We'll build reports for each.

The events you have to stamp (and who stamps them)

A metric is only as good as the discipline behind its timestamps. The most common reason a roofing shop "can't measure cycle time" is that nobody owns the click that records each event, so the dates are blank, backdated, or guessed.

Fix that first. Assign every event a single owner and a single trigger. Here's a working assignment that holds up in a 10-to-60-job-a-month operation:

1. Supplement flagged — owner: estimator or production coordinator. Trigger: the moment someone identifies that the carrier's scope is missing items your repair requires (a starter course, drip edge, ice-and-water where code calls for it, steep/high charges, detach-and-reset, etc.). Stamp it the day it's noticed, not the day you get around to it.
2. Evidence complete — owner: field tech / canvasser / production. Trigger: every required photo, measurement, and code citation is captured and uploaded. Not "some photos." Complete against a checklist.
3. Estimate written — owner: estimator. Trigger: the line-item estimate (Xactimate-aligned line items, quantities, and unit logic) is finished and internally reviewed.
4. Packet QC'd — owner: supplement manager. Trigger: packet passes a completeness check (we'll define the checklist).
5. Homeowner submitted — owner: whoever hands the packet to the homeowner. Trigger: confirmed the homeowner sent it to their carrier. You record the date the homeowner tells you they submitted; you are not the filer.
6. Carrier decision — owner: supplement manager. Trigger: any written response — approval, partial, request for more info, or denial. "Request for more info" is a decision for timing purposes; it stops the clock on submit-to-decision and starts a rework clock.
7. Final invoice / depreciation paperwork sent — owner: accounting or supplement manager. Trigger: for recoverable depreciation, the completion documentation and final invoice go out.
8. Cash received — owner: accounting. Trigger: funds land.

The single most valuable discipline here is stamping event #1 honestly. Shops love to start the clock at "estimate written" because it makes them look fast. But the detection-to-documentation gap — the days between noticing a supplement is needed and actually capturing the evidence — is frequently the biggest hidden delay in the whole chain. If your crew left the roof three weeks ago and now you need a photo of the valley, you've got a return trip and a 21-day hole that nobody recorded. Stamp #1 at detection and that hole becomes visible.

A note on backdating and "day zero"

Decide your day-zero convention and never deviate. Two common conventions:

• Calendar days, including weekends and holidays. Simpler, harsher, and it reflects what the homeowner and your bank account actually feel.
• Business days, excluding weekends/holidays. Fairer to your team's controllable stages, but it hides real-world wait the customer experiences.

Use calendar days for lead time and customer-facing numbers, business days for internal stage coaching if you want, but never mix them in the same chart. Label every metric with its convention. A "12-day average" means nothing if half your jobs are counted one way and half the other.

The core calculation, step by step

Let's make this concrete with a worked example. Take a single hail supplement, dates stamped:

• Supplement flagged: March 3
• Evidence complete: March 11
• Estimate written: March 13
• Packet QC'd: March 14
• Homeowner submitted: March 16
• Carrier decision (approved): April 2
• Final invoice sent: April 9
• Cash received: April 24

Stage cycle times (calendar days):

Now the lesson. Your eye goes to the 17-day carrier review, because "the insurance company is slow" is the easy story. But look at what you control: detection-to-documentation was 8 days and decision-to-invoice was 7. That's 15 days, almost as much as the carrier's review, sitting entirely inside your operation. The 8-day evidence gap is a crew that didn't finish documentation before leaving, and the 7-day invoice gap is paperwork sitting on someone's desk. You can't compress the carrier's queue. You can absolutely close those two gaps next week.

That's the entire point of measuring stage-by-stage: it moves your attention from the delay you complain about to the delay you can kill.

Averages lie; use medians and percentiles

Do not report cycle time as a simple mean. One catastrophic 180-day job will drag your average up and make a genuinely good month look terrible, or mask a slow month behind a few fast wins. Roofing supplement timelines are right-skewed — most cluster, a few run very long — so the mean overstates the typical experience.

Report three numbers per stage:

• Median (P50) — the typical job. This is your headline.
• 75th percentile (P75) — your "normal bad" job. Useful for setting follow-up cadence.
• 90th percentile (P90) — your tail. These are the jobs that need an escalation rule.

If your P50 packet-to-submit is 3 days but your P90 is 22 days, you don't have a speed problem, you have a consistency problem on a slice of jobs — usually a specific crew, a specific carrier, or homeowners who go quiet. The percentile spread tells you which.

Weighting by dollars

A 9-day cycle on a 600-dollar gutter supplement and a 40-day cycle on an 11,000-dollar full-scope supplement are not equally important. Track a dollar-weighted cycle time alongside the count-based one: multiply each job's cycle time by its supplement dollar value, sum, and divide by total dollars. When the dollar-weighted number is much worse than the count-weighted one, your biggest supplements are your slowest — which is the worst possible pattern for cash flow, and a sign your team is avoiding the hard, high-value scope work.

A worked example with a denial and a re-submission

The clean linear example above is the happy path. Real supplements loop. Here's a second job that hits a documentation request, because that's where most of the lost time actually lives:

• Supplement flagged: May 2
• Evidence complete: May 9
• Estimate written: May 10
• Packet QC'd: May 11
• Homeowner submitted: May 13
• Carrier responds — request for additional documentation: May 28
• Additional evidence captured (return trip for valley + flashing photos): June 6
• Revised packet submitted by homeowner: June 9
• Carrier decision (partial approval): June 24
• Final invoice sent: June 27
• Cash received: July 14

Now time it with the loop made visible:

The rework loop added 12 days plus a return trip, and it pushed the job from a clean ~50-day cycle into a 73-day one. Where did the rework come from? The carrier asked for valley and flashing photos that should have been in the packet at QC. That's not a carrier-speed problem — it's a packet-completeness problem you caused, and it cost you 12 days and a truck roll. This is the single most important pattern to instrument: tag every "request for additional documentation" and measure your rework rate (share of supplements that trigger at least one documentation request) and your rework days (time lost to the loop). When you cut rework, you cut both days and field cost at once, and you don't need the carrier to do anything differently.

A quick formula sheet

For anyone building this in a spreadsheet, here are the calculations spelled out:

• Stage cycle time = (stage-end date) − (stage-start date), in your chosen day convention.
• Total lead time = (cash received) − (supplement flagged).
• Controllable cycle time = (packet QC'd) − (supplement flagged). This is the number you own.
• Median (P50) = sort the stage's values, take the middle one. Even N: average the two middle values.
• P90 = sort, take the value at the 90th-percentile position (for N values, position ≈ 0.9 × N, rounded up).
• Dollar-weighted cycle time = Σ(job cycle time × job supplement value) ÷ Σ(job supplement values).
• Rework rate = (supplements with ≥1 documentation request) ÷ (all supplements submitted).
• Abandonment rate = (supplements that never reach a decision) ÷ (all supplements entered).
• Days-in-stage (for aging, as of today) = today − (date current stage started).

Keep these in one tab with raw timestamps in another, and your three reporting views all build off the same source data — no double entry.

Building the supplement aging report

Cycle time is historical. Aging is live, and it's what you actually manage day to day. An aging report buckets every open supplement by how long it's been sitting in its current stage as of today.

Build it like accounts-receivable aging, but per stage. Standard buckets:

Each open supplement appears once, in the bucket matching its current-stage age. The report's job is to make "what do I work today" obvious: everything in Stale and Critical gets touched today, everything in Aging gets a scheduled follow-up, Watch gets an eye.

The trick that separates a useful aging report from a guilt-inducing one is the next-action column. An aging line that just says "Job 4471, 38 days, Stale" produces anxiety, not action. An aging line that says "Job 4471, 38 days in Submit-to-decision, next action: homeowner to call carrier for status, owed since 3/12" produces a phone call. Every aging row needs: stage, days, owner, and the single next action with who does it.

Stage-specific aging thresholds

A flat 30-day threshold across all stages is lazy. Estimate-writing should never sit 30 days; carrier review legitimately can. Set per-stage thresholds based on what's reasonable for that step:

• Detection-to-documentation: flag at 5 days. There's no excuse for evidence sitting a week.
• Documentation-to-estimate: flag at 3 days.
• Estimate-to-packet: flag at 2 days.
• Packet-to-submit: flag at 5 days (homeowner-dependent; gentle nudges).
• Submit-to-decision: flag at 21 days, escalate at 30. This is the one carrier-controlled stage, so your "action" is a documentation-status follow-up, not pressure.
• Decision-to-invoice: flag at 3 days. Money this close to the door should not sit.
• Invoice-to-cash: flag at 30 days, then a structured collections cadence.

When you set thresholds per stage, your aging report stops being a wall of red and starts pointing at the two or three places that are actually abnormal.

Reading the aging report in a Monday huddle

The aging report only earns its keep if it drives a short, repeatable meeting. Here's a 15-minute format that works for a supplement desk:

1. Critical first (61+ days). Read each line aloud with its next action and owner. These don't move to the next agenda item until the owner commits to a specific step today. If a job has been Critical for two cycles in a row, decide whether it's genuinely recoverable or whether it belongs in the abandoned column — keeping dead jobs on the aging report inflates your open-dollar number and demoralizes the desk.
2. Stale (31–60 days), grouped by stage. If five of seven Stale jobs are stuck in the same stage, that's a process flag, not five coincidences. Spend the time on the stage, not the jobs.
3. Aging (15–30 days) — scan, don't dwell. Confirm each has a scheduled next-touch date. Move on.
4. One systemic question. Pick a single recurring pattern from the report ("three of our Stale jobs are documentation requests from the same carrier") and assign one person to investigate the root cause before next week.

The discipline is finishing in 15 minutes. A two-hour aging meeting means you're using the meeting to do the follow-up work instead of to direct it, and that's a sign the work isn't owned the rest of the week.

The abandoned-supplement column nobody wants to look at

Every aging report needs a column most shops refuse to maintain: supplements that have effectively died — the homeowner went dark, the job got too old, or it's been Critical so long it isn't coming back. Hiding these makes your cycle-time numbers look better than reality, because dead jobs have effectively infinite cycle time and dropping them quietly is the worst kind of survivorship bias.

Keep a running abandoned list with the dollar value and the reason each died. Over a quarter, the reasons cluster, and the clusters are gold: "homeowner went dark after submission" points at a follow-up-cadence gap; "aged out waiting on a return-trip photo" points at field documentation discipline; "never got past detection" points at a flagging process that creates supplements nobody works. Each cluster is a fix. The abandoned column is where your most expensive process problems confess themselves.

Recoverable depreciation: the stage everyone forgets to time

Here's where a lot of restoration shops quietly leak the most money, because depreciation recovery isn't even on their cycle-time radar.

Quick refresher on the mechanics, on the documentation side where you belong. When a roof is covered, the carrier often pays the homeowner the actual cash value (ACV) first — the replacement cost minus depreciation — and holds back the recoverable depreciation until the work is complete and documented. The homeowner gets that held-back amount released by their carrier after they submit proof of completion and the final invoice. That second check is real money, frequently thousands of dollars, and it's the last thing in the chain — which means it's the easiest to forget once the roof is built and the crew has moved on.

Your role is purely documentary: produce clean completion evidence and an accurate final invoice so the homeowner can request release of their held-back funds. You don't "recover depreciation" for the homeowner against the insurer, you don't promise it'll be released, and you don't tell them coverage is guaranteed. You make the paperwork airtight and timely. The carrier decides.

The depreciation cycle has its own clock that most shops never start:

The killer stat: many shops have completed roofs sitting for 30, 60, 90+ days with depreciation paperwork never sent, simply because nobody owns the trigger "roof finished → start completion docs." That's not slow cycle time, that's a missing process. The first time you build a completion-to-evidence aging report, you'll usually find a pile of finished jobs with un-started depreciation packets. Each one is money owed to your customer (and your final draw) sitting in a drawer.

Depreciation completeness checklist

A depreciation release request typically needs, on the documentation side:

• Signed completion certificate or equivalent proof the work is done
• Dated completion photos showing the finished roof and key details (ridges, valleys, flashings, accessories)
• Final invoice matching the approved scope, line for line
• Any code-required upgrade documentation that was part of the approved estimate
• Permit close-out / inspection sign-off where the jurisdiction requires it

Missing any one of these is the usual reason a release request stalls. Score the packet for completeness before the homeowner submits it, and your submit-to-release time drops because there's no back-and-forth for missing items.

The deductible, tracked but never touched

While you're timing depreciation, track the deductible the homeowner owes on the same job record — but understand exactly what "track" means here. You document the deductible amount, you invoice it cleanly as part of the homeowner's obligation, and you reconcile it against the final numbers. That's legitimate accounting on your own contract.

What you never do — and what gets contractors fined, sued, or worse in many states — is advertise or arrange to waive, absorb, rebate, or "eat" the deductible, or tell a homeowner the deductible is "gone" or that the roof is "free." That's insurance fraud territory in a lot of jurisdictions and it's on every state regulator's radar. Your cycle-time system should make the deductible visible and invoiced, never disappeared. If your templates or your sales scripts hint at making a deductible vanish, that's a liability you want out of your operation before you scale any of this.

A clean deductible-tracking line on each job also tightens your invoice-to-cash stage: when the deductible is documented up front, the final invoice math is unambiguous and the homeowner isn't surprised, which is one less reason for the last check to sit.

What pros get wrong about measuring this

A decade of watching shops try to instrument this produces a short list of repeat mistakes.

Starting the clock too late. Covered above, but it's the number one error. If day zero is "estimate written," you've amputated the detection gap, which is often your worst stage. Start at detection.

Counting only the won supplements. If you only measure cycle time on supplements that eventually got approved, you've hidden every supplement that died in process — the ones that aged out, got abandoned, or where the homeowner went dark. The abandoned ones often have the worst (infinite) cycle time and they're the ones costing you the most. Measure every supplement that enters the funnel, and track an abandonment rate alongside cycle time. A 14-day median looks great until you notice 30 percent of supplements never reach a decision at all.

Treating "request for more info" as the carrier being slow. When a carrier asks for additional documentation, that's frequently a signal your packet was incomplete. Tag those events. If a meaningful share of your supplements trigger a documentation request, your real bottleneck isn't carrier speed — it's packet quality, and you fix it upstream at QC, not downstream with follow-up calls.

Averaging across wildly different supplement types. A drip-edge add and a full code-upgrade supplement with steep-slope and detach-reset have nothing in common timing-wise. Segment before you average (next section). A blended average is a number you can't act on.

Confusing activity with progress. "We made 40 follow-up calls this week" is activity. If those calls were on jobs already past the carrier's normal review window, that's just noise on stages you don't control. Spend follow-up energy where the aging is abnormal for that stage, not where it's simply long-but-normal.

Letting the metric become a weapon. The fastest way to kill data quality is to use cycle time to publicly punish a crew. They'll respond by gaming the timestamps — stamping "evidence complete" before it is, backdating submissions. Measure to find process bottlenecks, coach privately, and protect the honesty of the stamps above all.

Segmenting cycle time so the number means something

A single company-wide cycle time is a starting point. The insight lives in the segments. Cut your cycle-time and aging data at least these ways:

By supplement type/size. Small adds, mid-scope, and full code-upgrade supplements behave differently. Track each band separately. You'll often find your team is fast on the easy adds and slow on the high-dollar complex ones — exactly backwards from where you want speed.

By carrier. Submit-to-decision varies a lot by carrier and region. You can't change a carrier's process, but you can set realistic per-carrier follow-up cadences and stop wasting energy chasing one that simply reviews on a longer normal cycle. Just keep your follow-up strictly to documentation status — you're confirming they have everything, not arguing coverage.

By estimator / supplement writer. Your controllable stages (documentation, estimate, packet) vary by person. If one writer's estimate-to-packet P50 is 1 day and another's is 6, that's a training and workload signal, not a character flaw. The data lets you coach with specifics.

By crew / production source. Detection-to-documentation depends heavily on whether the crew captured complete photos before leaving. A crew with a chronic evidence gap is creating return trips and cycle-time holes. That's a documentation-discipline fix at the field level.

By lead source / storm event. Jobs from a single hail event documented while the damage is fresh and crews are mobilized move faster than scattered jobs documented months later. This connects directly to which roofs you target and when — getting to recently storm-exposed roofs early shortens the entire downstream cycle because the evidence is fresh and access is easy.

A worked segmentation that changes a decision

Segmentation is only worth the effort if it changes what you do. Here's a realistic cut and the decision it forces. Say your blended controllable cycle (detection to packet QC) is a respectable 6 days. Feels fine. Now segment by supplement size:

The blended 6-day number was hiding the real picture: your team is quick on the cheap, easy supplements and slow on the expensive, complex ones. The dollar-weighted cycle time is far worse than 6 days because your biggest dollars sit the longest. That's exactly backwards. The decision this forces: the constraint is complex-scope estimating capacity, not general speed. The fix isn't "work faster" — it's either training a second writer on full-scope supplements, building reusable code-citation templates for the recurring upgrade items, or pulling the opportunity-detection step earlier so the missing scope surfaces with its evidence instead of being hunted line by line. Without the segment, you'd have spent your energy shaving a day off the small adds that don't matter. With it, you put the effort where the money sleeps.

Where RoofPredict fits: timing the work you control

Everything to this point is process and discipline, and you can do a lot of it in a spreadsheet. The reason most shops still can't see their cycle time is that the timestamps live in five different places — photos in one app, the estimate in another, the packet in email, the invoice in QuickBooks — and nobody stitches them into a per-stage clock. The measurement dies in the seams.

RoofPredict's RoofClaim module is built to put those stamps in one place and start the clocks automatically, on the documentation side of the line. Concretely, here's what your supplement manager actually does with it:

• Claim intake linked to the home. A supplement is created against the specific property and its roof profile, so the clock starts at detection — event #1 — instead of whenever someone remembers to open a spreadsheet row. That alone fixes the most common measurement error.
• Upload, auto-classify, and OCR the documents. Carrier estimates, contractor estimates, photos, denial letters, and invoices get read and sorted automatically. The OCR pulls line items off the carrier's estimate so you're not re-keying them to compare scope. Less re-keying means the documentation-to-estimate stage gets shorter, and the timestamp is captured the moment the doc lands.
• Opportunity detection with evidence anchors and pricing. This maps the carrier's estimate line items against a roofing knowledge base and flags missing scope, code-required items, and missed supplements — each tied to the photo or measurement that proves it, with pricing attached. Instead of an estimator hunting line by line for what's absent, the gaps surface with their evidence. That compresses the slowest controllable stage (writing the estimate) and, just as importantly, raises packet quality so you trigger fewer "send more documentation" requests downstream. It does not negotiate with anyone or promise an outcome; it documents what's missing and why, for the homeowner to submit.
• Packet-completeness scoring. Before the homeowner submits, the packet gets scored for completeness against the checklist for that supplement type. A low score means a missing photo or code citation gets caught at QC — the cheapest possible place to fix it — instead of coming back as a carrier query two weeks later. This is the single biggest lever on submit-to-decision time, because incomplete packets are what generate the back-and-forth.
• Supplement aging plus a follow-up cadence. The system tracks each open supplement's age in its current stage and runs the per-stage follow-up cadence, so the Stale and Critical buckets surface themselves and the next-action list builds itself. You stop maintaining the aging report by hand.
• Recoverable-depreciation autopilot. When a roof is marked complete, it starts the completion-evidence and final-invoice checklist automatically — closing the "nobody owns the trigger" gap that leaves finished jobs with un-started depreciation paperwork. The completion-to-evidence clock starts on its own.
• Deductible tracking sits alongside, so the deductible the homeowner owes is documented and invoiced cleanly. You're tracking and invoicing what's owed; you are not waiving, absorbing, or erasing a deductible, and the templates won't let you advertise that you do.

Honest limits, because the do-not-say list is the whole game in this category: the templates are locked and UPPA-gated on purpose. They produce contractor-documentation-only outputs — supplement packets, depreciation-release letters, deductible invoices, missing-docs letters, and audit reports. They do not negotiate claims, interpret policy or coverage, or promise a payout or approval. The opportunity detection flags what your repair scope requires and what's missing from the carrier's line items, with evidence — it does not adjust the claim. That boundary is deliberate, and it's what keeps you on the right side of the public-adjusting line while still cutting days out of every controllable stage.

The upstream lever: fewer slow jobs by targeting better

The fastest supplement is the one documented on a fresh, accessible roof right after a storm, by a crew that's already mobilized in that neighborhood. The slowest is a scattered one-off documented months later, with a return trip for missing photos. So part of cycle-time management happens before the supplement even exists — in which roofs you go after, and when.

This is where RoofPredict's targeting side feeds the claim side. The platform scores every home in a service area by roof-age band (recent, mid-life, due, overdue) and per-roof storm exposure, then builds a ranked target audience — house by house — of roofs that are likely due and recently storm-exposed, each with a "why this home" evidence chain. Two honest caveats that matter: roof age is estimated as a range, not an exact install date, and a storm-exposure score is odds of exposure, not proof of damage. It tells you where to look first; the roof and the documentation tell you what's actually there.

Why this shortens cycle time: when you concentrate documentation on recently storm-exposed roofs while the evidence is fresh and crews are in the area, your detection-to-documentation stage shrinks, return trips drop, and packet quality rises — all of which pull days out of the downstream cycle. You can turn that ranked list into tracked direct mail with personalized reports and QR codes, build canvassing routes for the field app, and capture leads into the pipeline, so the jobs entering your supplement funnel are concentrated, fresh, and well-documented from day one rather than scattered and stale.

Tying cycle time to cash, capacity, and morale

It's worth being explicit about why this number deserves a seat next to close rate and average ticket, because the case is stronger than "faster is nicer."

Cash. Every day a supplement or depreciation packet sits is a day your money is financing the carrier's review queue and the homeowner's procrastination instead of your payroll. If you carry, say, 60 open supplements averaging a few thousand dollars each, shaving a week off the controllable portion of the cycle pulls a meaningful five-figure sum forward in time, every month, permanently. That's working capital you create out of process discipline, with zero change to your pricing or close rate.

Capacity. A supplement desk's real capacity isn't measured in jobs — it's measured in jobs times how long each one stays open and demanding attention. A job that closes in 18 days occupies a slot for 18 days. A job that drags 60 days occupies that slot for 60 days, and a desk full of slow jobs is a desk that feels permanently underwater even when volume is flat. Cutting cycle time is the cheapest capacity increase available to you. You don't hire; you just stop letting jobs marinate.

Morale and documentation discipline. This one is underrated. When supplements routinely die in process or take two months to pay, your field crews and estimators learn — correctly — that careful documentation doesn't get rewarded, so they stop doing it. That's a doom loop: weak documentation causes rework and abandonment, which causes slow cycles, which teaches the team that documentation is pointless, which weakens documentation further. A fast, visible cycle-time program breaks the loop in the other direction. When crews see that a complete photo set on the roof leads to a clean packet that pays in three weeks, the documentation discipline becomes self-reinforcing. The metric isn't just measuring the behavior you want — it's manufacturing it.

What a healthy supplement operation looks like

If you want a target picture to aim your monthly scorecard at, a well-run restoration supplement desk tends to show this profile (your carrier and storm mix will shift the carrier-controlled stages):

• Detection-to-documentation: P50 of 2–3 days, P90 under a week. Crews document before leaving.
• Documentation-to-packet (estimate + QC): P50 of 2–3 days. Reusable templates and surfaced scope gaps keep this tight.
• Rework rate: low and trending down, because packet-completeness scoring catches gaps before submission.
• Submit-to-decision: variable by carrier, but predictable per carrier, with follow-up cadence matched to each carrier's normal review window.
• Decision-to-invoice: P50 of 1–2 days. Money this close to the door does not sit.
• Completion-to-depreciation-sent: P50 of a few days, with no finished jobs aging past two weeks with un-started paperwork.
• Abandonment rate: low and explained — every abandoned job has a logged reason that feeds a process fix.

The shape that matters: the controllable stages are short and consistent, the carrier-controlled stage is predictable and worked with documentation-status follow-up only, and nothing valuable is sitting un-started in a drawer.

A 30-day plan to start measuring

You don't need a six-month rollout. Here's a sequence that gets you a real cycle-time number in a month.

Week 1 — define and assign. Lock your stage list and event definitions. Pick calendar vs. business days and write it down. Assign one owner per event. Build the completeness checklist for each supplement type. Don't automate anything yet; get the definitions right.

Week 2 — start stamping. Every new supplement from day one of this week gets all events stamped, honestly, by the assigned owner. Stamp detection at detection. Resist backfilling old jobs with guessed dates — start clean with new flow.

Week 3 — build the three views. From the stamps you're collecting, build: (1) a per-stage cycle-time table with P50/P75/P90, (2) a live aging report bucketed by stage with a next-action column, (3) a completion-to-evidence aging report to surface un-started depreciation packets. Even with two weeks of data the aging reports are immediately useful.

Week 4 — find the two worst stages and fix one. Look at your P50 and P90 by stage. The two stages with the biggest gap between "reasonable" and "actual" are your targets. Pick the one you fully control — almost always detection-to-documentation or decision-to-invoice — and put a hard rule on it (e.g., "no crew leaves a roof until the documentation checklist is green," or "final invoices go out within 48 hours of approval"). Re-measure next month.

Then repeat. Cycle-time improvement is not a project, it's a monthly rhythm: measure, find the worst controllable stage, fix one rule, re-measure.

A simple scorecard to review monthly

Keep the monthly review to one page:

The last two rows are the ones that catch the money everyone else misses: abandoned supplements and aged-out depreciation. If your scorecard only had those two lines plus controllable cycle time, you'd already be ahead of nearly every shop in your market.

The boundary, one more time, because it pays to be careful

Everything in a cycle-time program is about speed and completeness of your documentation. Keep your follow-ups to documentation status — "do you have everything you need from us" — never to coverage or payout. Write accurate, Xactimate-aligned repair estimates for your own scope and hand them to the homeowner to file. Never tell a homeowner their claim will be approved, never promise a number, never advertise a free roof, never tell them the deductible is waived or gone, and never represent them against their insurer for a fee. Document thoroughly, estimate accurately, package completely, hand it over. The homeowner files; the insurer decides.

Stay in that lane and a cycle-time program is pure upside: faster cash, less rework, fewer abandoned supplements, and a depreciation pipeline that doesn't leak. The shops that win the restoration game aren't the ones with the most aggressive claim tactics — those are the ones who get in trouble. They're the ones whose documentation is so fast and so complete that there's nothing to argue about, and whose finished roofs never sit a single extra day waiting on a packet nobody started.

Measure the chain. Stamp every event. Fix the worst controllable stage each month. The number will move, and so will the cash.

FAQ

What is supplement cycle time in roofing?

It's the elapsed calendar time between two events you define on a supplement — most usefully measured as a chain of per-stage times (detection, documentation, estimate, packet QC, homeowner submission, carrier decision, invoice, cash) rather than a single span. Measuring it stage by stage shows you exactly where the days disappear, separating delays you control from delays you don't.

When should I start the clock on a supplement?

At detection — the moment someone notices the carrier's scope is missing items your repair requires — not when the estimate is written. Starting at estimate-written hides the detection-to-documentation gap, which is frequently the biggest controllable delay in the whole chain (think return trips for missing photos).

Should I use average or median cycle time?

Median. Supplement timelines are right-skewed — most cluster and a few run very long — so a simple average gets dragged up by outliers and overstates the typical job. Report P50 (median) as your headline, plus P75 and P90 to see your normal-bad and tail jobs that need escalation rules.

How do I build a supplement aging report?

Like AR aging, but per stage. Bucket every open supplement by how long it's been in its current stage as of today (Current 0–7, Watch 8–14, Aging 15–30, Stale 31–60, Critical 61+). Set per-stage thresholds instead of one flat number, and give every row a next-action column with stage, days, owner, and the single next step.

Why is recoverable depreciation the stage most shops forget to time?

Because it's the last thing in the chain and happens after the crew has moved on. Nobody owns the trigger 'roof finished → start completion docs,' so finished jobs sit for 30, 60, or 90+ days with depreciation paperwork never sent. Build a completion-to-evidence aging report and you'll usually surface a pile of finished jobs with un-started depreciation packets — money owed to the customer and your final draw, sitting idle.

Can a roofing contractor recover depreciation for the homeowner?

No — that crosses into claim handling. Your role is documentary: produce clean completion evidence and an accurate final invoice so the homeowner can request release of their own held-back funds from their carrier. You don't promise the funds will be released, you don't interpret coverage, and the homeowner submits the request. The carrier decides.

How do I separate carrier delays from my own delays?

Split the chain at the submission handoff. Stages from detection through packet QC are fully yours; submit-to-decision is the carrier's. When you measure each stage separately, you'll often find a big chunk of total time (detection-to-documentation plus decision-to-invoice) sits inside your operation, where 'the insurance company is slow' was getting the blame.

Does a 'request for more information' mean the carrier is slow?

Usually it means your packet was incomplete. Tag those events. If a meaningful share of your supplements trigger a documentation request, the real bottleneck is packet quality, which you fix upstream at QC with a completeness checklist — not downstream with more follow-up calls.

How does RoofPredict's RoofClaim help with cycle time?

It puts every timestamp in one place and starts the clocks automatically on the documentation side: claim intake linked to the home (clock starts at detection), OCR and auto-classification of carrier/contractor docs, opportunity detection that flags missing scope and code items with evidence anchors and pricing, packet-completeness scoring before submission, supplement aging with a follow-up cadence, and a recoverable-depreciation autopilot that starts the completion checklist when a roof is marked done. It produces documentation-only outputs on UPPA-locked templates — it doesn't negotiate claims, interpret coverage, or promise a payout.

What's the single highest-leverage fix to shorten cycle time?

Raise packet completeness before submission. Incomplete packets generate the back-and-forth 'send more documentation' requests that blow up submit-to-decision time. A completeness checklist scored at QC catches the missing photo or code citation at the cheapest possible place — before it comes back as a carrier query weeks later. After that, the two best controllable fixes are closing the detection-to-documentation gap (no crew leaves a roof until documentation is complete) and the decision-to-invoice gap (invoices out within 48 hours of approval).