How to Prioritize Which Roofing Estimates to Send First

There is a pile on your estimator's desk, or a queue in your CRM, that quietly decides how much money your company makes this month. It is the backlog of roofs you have measured, walked, or promised a number on but have not yet sent a written estimate for. Most companies work that pile the way a deli works a ticket line: first in, first out, oldest measurement on top. That feels fair. It is also one of the most expensive habits in residential roofing, because the order you send estimates in has little to do with the order they were created and nearly everything to do with which homeowner is ready to say yes.

Think about what a backlog of twenty unsent estimates actually represents. Some of those homeowners called you in a panic after a leak and would sign tonight if you handed them paper. Some are tire-kickers who wanted three numbers for a roof they will not touch for two years. Some are storm-damage homes where the clock on a claim filing deadline is running whether you act or not. If you process all twenty in the order they came in, you let a two-year tire-kicker delay the panicked leak by four days, and you let a slow comparison-shopper push a time-sensitive storm file past a deadline you cannot get back. The work is the same. The sequence is everything.

What follows is a practical system for deciding which estimates leave your shop first. It is built around a scoring model you can run on a sticky note or in a spreadsheet, a set of tie-breakers for when two jobs score the same, and the operational glue, turnaround targets, follow-up cadence, and capacity math, that keeps the system honest once real life starts pulling at it. There are worked examples with real numbers, a few things experienced estimators get wrong, and a short, honest note on where data tools fit. By the time you reach the end you will be able to look at any backlog and say, with a reason you can defend to your crew and your accountant, what goes out next.

Why send-order is a profit lever, not a clerical task

Roofing is a business of perishable attention. A homeowner who calls you is, for a window of a few days, more receptive to you than they will ever be again. Every day that passes, the leak dries, the panic fades, the neighbor recommends a different company, and the other two estimates the homeowner requested show up in their inbox. Industry sales research across home services consistently finds that the contractor who responds first wins a disproportionate share of the work, not because they are better, but because they are present at the moment of highest intent. Your estimate is your response. If it sits in a queue for five days because three lower-intent jobs were ahead of it, you have spent your single biggest advantage on paperwork order.

There is also a cash-flow argument. An estimate that closes turns into a deposit, a material order, and a scheduled crew. An estimate that sits is working capital frozen in a Word document. If your average residential re-roof carries a gross profit in the low-to-mid four figures, then moving one extra close per week from "next month" to "this week" by sequencing better is not a rounding error. It is a measurable change in how fast money cycles through the company. You already paid for the lead, the drive, the measurement, and the estimator's time. Sending the estimate is the cheapest step in the whole chain and the one most likely to be done in the wrong order.

Finally, send-order protects your crews from feast-and-famine whiplash. When you sequence by readiness, you smooth the flow of signed contracts, which smooths the flow of work onto the schedule board. When you sequence by accident, you get clumps, four signs in one day because three sat too long and then all closed at once, followed by a dry week. Sequencing is the throttle on your own pipeline.

A quick reframe before the model

Prioritizing estimates is not the same as prioritizing leads. Lead prioritization happens before you have invested the drive and the measurement; it is about who you go see. Estimate prioritization happens after that cost is sunk; the question is no longer whether the job is worth pursuing but how soon and how hard you pursue the close. That distinction matters because it changes the math. You are not protecting your time anymore, you already spent it. You are protecting the conversion of work you have already done. So a job you might have ranked low as a lead can still deserve a fast estimate if the homeowner's readiness is high.

Capturing the data you need before you can prioritize

You cannot rank a job on a factor you did not capture. The most common reason a prioritization system fails is not a bad model; it is a job sheet that records the address, the measurement, and nothing about the human who has to sign. Garbage in, alphabetical out. So the work starts in the driveway, before anyone scores anything.

Train every estimator to leave each appointment with five things written down, in addition to the measurement: the trigger, the timeline, the decision structure, the competitive picture, and a readiness gut-read. None of these takes more than a sentence, and together they are the raw material the scoring model runs on.

• The trigger. What made them pick up the phone now? "Ceiling stain after Tuesday's rain" is a different animal from "saw your truck in the neighborhood." The trigger is the single best predictor of intent, and it is gone the moment you leave if you do not write it down.
• The timeline. When do they say they want it done? Get a real answer, not a polite one. "Before my daughter's wedding in August" is a deadline. "Eventually" is not. If they will not commit to a window, that itself is data: it scores low.
• The decision structure. Who has to agree? Is a spouse not home today? Is there financing in play, and have they been approved? A clean one-person, cash-or-approved decision closes far faster than a two-spouse, financing-pending one, even at identical intent.
• The competitive picture. Are you the first bid, the third, or the only one? Did they mention a brother-in-law in the trades or a number a neighbor quoted? This feeds win probability directly and keeps you from over-investing in a courtesy bid.
• The readiness gut-read. Your estimator stood in their kitchen. Did this feel like someone about to buy a roof or someone collecting numbers for a file? That instinct, captured as a one-to-five hunch, is surprisingly predictive and costs nothing to record.

Build these five fields into your job sheet or CRM intake so they are not optional. An estimate that enters the queue without them gets scored on guesswork, and guesswork is how a leak ends up behind a tire-kicker. If your estimators resist, frame it the way it actually pays them: the notes they write are what let the office send their hot jobs same-day, which is what turns their measurements into commissions instead of stale PDFs.

The five factors that decide send-order

Every prioritization system, however fancy, is really a way of weighing a handful of factors against each other. For unsent roofing estimates, five factors carry almost all of the predictive weight. Learn to read these five and you can rank any pile.

1. Intent: how ready is this homeowner to buy

Intent is the strongest single predictor of whether an estimate closes, and it is the one estimators most often ignore because it does not show up on the measurement. Intent is about the homeowner's situation and language, not the roof. A homeowner with an active leak, a buyer's inspection deadline, an insurance file already opened, or a flat "I just need someone to do it" is high intent. A homeowner who says "we're getting a few numbers," "maybe next spring," or "my brother-in-law might do it" is lower intent, no matter how bad the roof is.

Read intent from three signals captured at the appointment: the trigger (what made them call now), the timeline (when they say they want it done), and the decision structure (who has to agree, and whether financing is in play). High intent is a clear trigger, a near-term timeline, and a simple decision. Train your estimators to write one intent line on every job sheet before they leave the driveway. It is the most valuable sentence in your CRM.

2. Time-sensitivity: is there an external clock

Some jobs have a deadline that exists whether you act or not, and those jump the line regardless of intent or value. The classic examples in roofing:

• A real-estate transaction with a financing or inspection contingency tied to a closing date.
• A storm-damage home where the homeowner intends to file with their carrier; many policies require notice within a defined window after the date of loss, and proof-of-loss documentation has its own deadlines.
• An active, worsening leak threatening interior finishes, where delay raises the homeowner's damage and their urgency to grab whoever responds.
• A homeowner leaving town, a seasonal weather window closing, or a permit or HOA approval that itself takes weeks.

Time-sensitivity is binary in practice: either there is an external clock or there is not. When there is, that job moves up, even past a higher-value job, because a missed deadline can mean the work disappears entirely.

3. Value and margin: what is the job worth to you

Not all jobs are equal dollars, and not all dollars are equal margin. A straightforward architectural-shingle re-roof at a healthy margin is worth more to your business than a low-slope tear-off you priced thin to win, even if the second has a bigger sticker. When you sort, weigh expected gross profit, not revenue. A $9,000 job at 35 percent margin contributes more than a $16,000 job at 12 percent, and it usually closes faster and causes fewer headaches.

Value also includes strategic worth: a visible house on a busy street where your yard sign earns impressions, a referral from your best past customer, or a property in a neighborhood you are trying to break into. Those carry weight beyond the immediate margin. Keep that weight modest, though, future referral value is real but speculative, and real cash this month beats a maybe.

4. Win probability: how likely is this one to close

Win probability blends intent with fit. Even a high-intent homeowner is a long shot if you are the fourth bid, if they anchored on a number far below your range, or if a relative is a roofer. A medium-intent homeowner who called you specifically, has only your card, and liked your estimator is a strong bet. Estimate win probability honestly, in broad bands rather than false precision: high (you are clearly the front-runner), medium (competitive but uncertain), or low (you are a courtesy bid). Sending a polished estimate fast to a low-probability courtesy bid is effort better spent elsewhere.

5. Effort to produce: how long until you can hit send

The four factors above describe the prize. This one describes the cost of claiming it. A simple gable re-roof you already measured can be turned into a clean estimate in fifteen minutes. A cut-up roof with three slopes, a chimney cricket, skylights, and a partial deck replacement might need two hours, supplier pricing on a special-order tile, and a structural look. When two jobs are otherwise close, the one you can send in fifteen minutes should usually go first, because finishing it frees the homeowner to decide while you work the harder one. Do not let a quick win rot behind a complex job that is going to take all afternoon anyway.

A simple scoring model you can run today

You do not need software to do this well. You need a consistent way to turn those five factors into a number so the pile sorts itself. Here is a 100-point model that has held up in the field. Score each unsent estimate, sort high to low, and work the top of the list.

Score each factor with simple anchors so two estimators rate the same job the same way:

• Intent (0 to 30): 30 = active leak, signed-buyer deadline, or "do it now"; 20 = wants it this season, clear trigger; 10 = vaguely shopping; 0 = "just curious."
• Time-sensitivity (0 or 25): 25 = a real external clock (transaction, claim-notice window, worsening leak, leaving town); 0 = none.
• Win probability (0 to 20): 20 = you are the only or clearly leading bid; 12 = competitive; 5 = courtesy bid.
• Value and margin (0 to 15): 15 = strong gross-profit dollars; 8 = average; 3 = thin.
• Effort to send (0 to 10): 10 = ready in under 20 minutes; 6 = an hour; 2 = needs supplier pricing, a second visit, or engineering.

Add them. Anything 70 and up is a today job. 50 to 69 is this-week. Below 50 is when-you-have-air. The exact thresholds matter less than the discipline of always working from the top.

Why these weights

Intent and time-sensitivity together hold 55 of the 100 points on purpose. Those two factors describe the homeowner's readiness and the external clock, the two things you cannot manufacture and cannot get back. Value sits lower than most owners expect because a big job you will not win, or will win in two months, is worth less today than a medium job that closes this week. And effort sits lowest of all because it is a tie-breaker, not a driver; never let "it is easy" pull a low-intent job ahead of a leak. The point of writing the weights down is consistency: when your sales manager and your estimator both score the pile and land within a few points, the system is working.

Worked example: ranking a real backlog

Here is a backlog of six unsent estimates scored with the model. The notes column is what the estimator wrote in the driveway.

Send-order falls out cleanly: Alvarez (91) tonight, Donnelly (82) and Bryce (71) tomorrow morning, Fitch (56) by end of week, and Castellano and Eames when there is room. Notice that Castellano is the biggest sticker price in the pile and lands near the bottom, because it is low intent, low win probability, and high effort. A first-in-first-out shop would have sent Castellano days before Alvarez and possibly lost the leak. That single re-sequencing is the whole point.

Working the special cases

The model gets you ninety percent of the way. The last ten percent is judgment about specific situations that recur in roofing.

Storm and hail work: the deadline is real, the lane is narrow

Storm-damage homes almost always score high because the external clock is real. When a homeowner intends to file with their insurer, the policy typically requires prompt notice of loss, and there are documentation deadlines that follow. That makes a fast, thorough estimate genuinely time-sensitive. So storm files move up the list, with one hard discipline attached: stay rigorously on your side of the line.

Your job as a roofer is to inspect the roof, document the damage thoroughly with dated photos and measurements, and prepare an accurate repair estimate, ideally aligned to the line-item format carriers expect, that reflects your scope to restore the roof. You hand that documentation and estimate to the homeowner. The homeowner files the claim. The insurer decides coverage. That is the safe and legal lane, and it is also the genuinely valuable one, because thorough documentation is what a homeowner actually needs and cannot produce themselves.

What you must not do, for a fee, is negotiate or "handle" the claim, interpret what the policy covers, promise a specific approval or payout, tell a homeowner their deductible will be waived or absorbed or that the roof will be "free," or otherwise represent the homeowner against their carrier. In many states that crosses into unlicensed public adjusting, and a 2024 Texas matter made clear that even labeling yourself an insurance or claims "specialist" in your advertising can run afoul of those rules. Keep a simple do-not-say list posted where your sales team can see it:

• Do not say you will "get the claim approved" or "handle/manage/negotiate" it.
• Do not say "we'll get your deductible covered/waived" or imply a free roof.
• Do not promise a settlement amount or guarantee the carrier will pay.
• Do not interpret the homeowner's coverage or tell them what their policy means.
• Do say: "We document the damage and write an accurate estimate; you file, and your insurer decides."

Prioritize storm estimates for speed and thoroughness of documentation. Never prioritize them by promising something on the claims side you are not licensed to deliver.

The big job that will not close fast

A large, complex, high-margin re-roof is seductive. It is also often a low-priority send, because complex jobs take longer to produce, longer to decide on, and frequently involve more comparison shopping. Do not ignore it, but do not let it block the board. The smart play is to split it: send a one-page preliminary scope and a price range fast, then schedule the time to produce the detailed estimate. The homeowner gets a quick response that holds their attention; you protect your afternoon for the easy wins; and you have a documented reason to follow up with the full proposal.

The referral and the relationship job

A referral from your best customer scores modestly on the cold factors but carries weight you should honor, because the cost of being slow is reputational as much as transactional. If your top referrer sent you to their neighbor and that neighbor waits five days for a number, your referrer hears about it. Give relationship jobs a one-tier bump and, more importantly, a personal touch on delivery: a call rather than just an emailed PDF.

The price-shopper who wants a number now

Low-intent comparison shoppers are where send-order discipline saves you the most. They feel urgent because they ask a lot of questions, but they convert poorly and consume estimating hours. Do not refuse them, just sequence them last and produce them efficiently. A clean, fast, standardized estimate is the correct response to a low-intent shopper, not a custom-crafted proposal that took an hour you owed to a leak.

The estimate itself: speed without sloppiness

Prioritizing which estimate goes first only pays off if the estimate that goes first is good. A fast, vague, error-riddled proposal loses the same job a slow one would, so the send-order discipline has to sit on top of a production process that can move quickly without cutting the things that actually close roofs. Two ideas make that possible: templates and tiers of detail.

Build standardized templates for your common roof types so the estimator is filling in numbers, not writing prose from scratch. A typical residential template carries the scope language, the warranty terms, the payment schedule, and the standard exclusions already written; the estimator drops in measurements, material selection, and price. That is how a hot job goes out in fifteen minutes without looking thrown together. The homeowner should never be able to tell a fast estimate from a slow one by its quality, only by how quickly it arrived.

Match the depth of detail to the tier. A hot job needs a clean, complete, professional estimate, but it does not need a forty-line breakdown that takes an hour; it needs the scope, the price, the timeline, and a clear next step, fast. A cool comparison-shopper needs an efficient, standardized number, not a custom masterpiece. Reserve your deep, itemized, photo-rich proposals for the jobs where that detail is the thing that wins, complex roofs, premium materials, and storm-documentation files where thoroughness is the actual product. Spending estimating craft where it does not move the close is the same mistake as sending in the wrong order, just one layer down.

A few elements earn their place on almost every estimate and quietly lift close rates: a dated photo or two from the inspection that show the homeowner what you saw, a plain-language scope so they understand what they are buying, a clear material and workmanship warranty, and one unambiguous next step ("reply or call to get on the schedule"). None of these slow you down once they live in a template, and all of them reduce the back-and-forth that lets a competitor's bid catch up.

Turnaround targets: how fast each tier should move

A priority order is only useful if it is attached to a clock. Otherwise everything is "important" and nothing moves. Set explicit turnaround targets by score tier and treat them as commitments, the way you would treat a crew's start time.

The single most important number on that chart is "same day" for hot jobs. The drop-off in close rate as response time stretches from hours to days is steep and well documented across home services. If you do nothing else from these pages, commit to a same-day estimate for any job scoring 70 or above, and protect a daily block of estimator time to make that promise real.

Building the daily block

Reserve a fixed window, say the first ninety minutes of the office day, for producing and sending hot estimates only. Nothing else touches that window: not measuring, not material orders, not callbacks. The estimator opens the queue, sorts by score, and clears the 70-plus jobs first. This one habit does more for close rate than any script, because it converts good intentions into a scheduled, defended action.

A full week, sequenced: what the system looks like in motion

Abstract rules are easy to nod at and hard to run. So walk through a single week the way the queue actually flows, using the six-job backlog from earlier plus the new jobs that arrive midweek, because a real backlog never holds still.

Monday morning. The estimator opens the queue and scores the six standing jobs. Alvarez (91, active leak) and Donnelly (82, fresh hail, wants documentation) are both hot. In the protected 90-minute block, the estimator produces both, Alvarez first because the leak is live, Donnelly second because the documentation needs care, and both go out before lunch. Bryce (71, home under contract, closes in 18 days) is hot too but the contingency window gives a little room; it goes out Monday afternoon. Each send triggers a same-day confirmation call.

Tuesday. Two new measurements come in overnight: a referral from a past customer with an active attic leak (scores 88) and a price-shopper gathering a third bid (scores 28). The leak-referral jumps straight to the top of Tuesday's block and goes out same-day with a personal call, ahead of Fitch (56) and Castellano (32), which were already waiting. This is the system working as intended: a job created Tuesday correctly beats jobs created last week because its readiness is higher. The price-shopper gets a fast, standardized number by Thursday and no custom effort.

Wednesday. Follow-up day. Alvarez has questions about the deck repair line; the estimator walks the homeowner through it on a call and Alvarez signs. Donnelly's homeowner is filing with their carrier and asks whether the deductible will be "taken care of", the estimator holds the line cleanly: "We document the damage and write the estimate; what your policy covers and how the deductible works is between you and your insurer." Donnelly's file stays clean and trustworthy precisely because nobody promised something they could not deliver.

Thursday and Friday. Warm-tier jobs (Fitch, the referral cohort) get produced and sent, the cool tier gets efficient standardized numbers, and the week's outcomes get logged: Alvarez won (leak), referral-leak won (referral), Bryce pending (in contingency), Castellano lost (price), Eames stalled (timeline). That log is Friday's most valuable artifact, because over a quarter it tells you whether your intent scores are predicting wins.

The shape of the week is the lesson: the queue is re-sorted every morning, new high-intent jobs leapfrog older low-intent ones without guilt, follow-up is scheduled not hoped for, and the claims line is held without flinching. Nothing here requires talent the average shop lacks. It requires running the same loop every day.

Who runs this: roles and handoffs

A prioritization system dies in the gap between people. The estimator captures the intent data, the office scores and sends, the salesperson follows up, and if any handoff is fuzzy, jobs fall into the crack and age. Decide explicitly who owns each step.

In a small shop, one person may wear all the hats, which is simple but fragile; the moment that person is on a roof, the queue stalls. The fix is to make the queue a shared, visible artifact, a board, a pipeline view, a shared sheet, so anyone can see the ranked list and the next action on each job. Visibility is what lets a second person cover the same-day promise when the primary owner is tied up.

In a larger shop, separate the roles cleanly. Estimators capture and produce; a sales coordinator or office manager owns the scoring, the daily block, and the follow-up cadence; the sales manager reviews the win-loss log and tunes the weights. The single most important handoff rule: every estimate has exactly one named owner of its next action at all times. "The team is handling it" means nobody is. When you can point at any job in the queue and name the person responsible for its next touch and the date it is due, the system holds. When you cannot, it is already leaking.

Guard against the two predictable people-failures. The first is the estimator who measures fast and documents nothing, starving the model of inputs; fix it by making the five intake fields mandatory and reviewing them. The second is the salesperson who works only the jobs they personally like, usually the big, exciting ones, instead of the ranked list; fix it by managing to the queue order, not to gut feel. The model exists precisely so that the next send is decided by a defensible score, not by whoever's favorite job is loudest.

Following up: the part everyone skips

Sending the estimate is the start of the close, not the end. A large share of estimates that eventually close do so after multiple contacts, yet a large share of estimators stop after one. Your follow-up cadence should mirror your send priority: hot jobs get fast, persistent follow-up; cool jobs get a lighter touch. A workable default cadence after sending:

1. Same day, within an hour of sending: a quick call or text confirming the homeowner received it and asking if the scope looks right. This is a service touch, not a pitch.
2. Day 2: a call to answer questions and offer to walk through any line item. Most real objections surface here.
3. Day 4 to 5: a value or proof touch, a relevant photo from the inspection, a note on the material warranty, or a reminder of the timing window.
4. Day 8 to 10: a direct ask. "Are we a fit to get you on the schedule, or has something changed?" A clear yes or no is more valuable than a maybe that sits.
5. Day 14+: move to a long-cycle nurture, a monthly check-in, rather than dropping it. Roofs do not get younger.

The discipline is to schedule the next touch the moment you finish the current one, so nothing falls through. An estimate with no scheduled next action is an estimate you have already half-lost.

A note on logging outcomes

Every estimate should end in a recorded outcome: won, lost, or stalled, with a one-word reason. Over a quarter, that log tells you which of your five scoring factors actually predicted wins in your market, so you can tune the weights. If high-intent jobs you scored 30 keep closing and big-value jobs you scored 15 keep dragging, the data is telling you to lean harder on intent. The model is a starting point; your own win-loss log makes it yours.

Capacity: matching what you send to what you can build

There is a failure mode at the other end of this. If you sequence purely for close probability and send aggressively, you can sell more roofs than your crews can install in a reasonable window. Then your fast estimate turns into a six-week schedule, the homeowner who signed on intent cools off waiting, and your reputation takes the hit instead of your competitor's.

So close the loop between the send queue and the schedule board. A simple weekly check: how many crew-days are open in the next three weeks, and how many crew-days of signed work do you already have. If you are oversold, do not stop sending hot estimates, those are your highest-intent homeowners and you must not lose them, but be honest about install timing in the estimate and the follow-up, and slow your push on the cool tier. If you are undersold, push the warm and cool tiers harder to fill the board. The send queue is the intake valve; the schedule board is the tank. Watch both gauges.

A worked capacity check

Suppose you run two crews that each complete roughly one standard residential re-roof per day, so ten crew-days a week, call it eight to ten roofs depending on complexity. Your backlog scoring shows nine jobs at 70-plus and eleven at 50 to 69. You send all nine hot jobs same-day, as you should. If even six of those close this week, you have filled most of next week's capacity from the hot tier alone. That is your signal to ease off the cool tier and set realistic install dates in your warm-tier estimates, rather than overselling and creating a backlog of irritated, waiting customers. Capacity math is what keeps prioritization from turning into over-promising.

Where the backlog comes from: better inputs, easier sorting

Everything above assumes the pile already exists and your job is to sort it. But the quality of the pile, how many high-intent, on-time, winnable jobs are in it, is set earlier, by which roofs you chose to go measure in the first place. If your backlog is full of low-intent comparison shoppers and roofs with years of life left, no scoring model will save the week; you will just be efficiently sorting weak jobs.

This is where knowing which roofs are actually due changes the inputs. A roof's age is the single best predictor of whether a homeowner is near a replacement decision, and storm exposure stacks on top of it. The problem is that the two public sources roofers reach for, year built on the county record or a listing site, do not tell you the roof's age at all; a re-roof is invisible to them, so a house built in 1998 might have a roof from last year or a roof from 2006. And a hail map tells you where it hailed, not which specific roofs the storm actually wore out.

This is the gap RoofPredict is built to close. It reads aerial imagery to estimate a roof's age as a range, house by house, and models storm physics, hail and wind, on each roof rather than just flagging the ZIP a storm crossed. The output is a ranked list of the addresses most likely to be due, so the roofs you choose to go measure are weighted toward homeowners who are genuinely closer to a decision. It also enriches the list you already own, your past customers and old estimates, with the same age and storm signals, which is often the highest-intent pile of all. Honest limits apply, and they matter: roof age comes back as a range, not an install date, and storm modeling gives you odds of meaningful wear, not proof of damage on a specific roof; you still inspect, you still verify, the homeowner still files any claim and the insurer still decides. What it does is tilt the top of your funnel toward the right doors, so by the time a job lands in your unsent-estimate pile, it is more likely to be one worth sending first. Better inputs make the sorting easier; they do not replace it.

Common mistakes experienced estimators make

Even seasoned shops fall into predictable traps. Watch for these.

• Sorting by sticker price. The biggest job feels like the priority. But a $20,000 roof at low margin you probably will not win is worth less today than a $9,000 leak repair that closes tonight. Sort by expected gross profit times probability, not revenue.
• Letting the complex job hold the queue. A two-hour estimate sits half-finished while five fifteen-minute jobs wait behind it. Split the complex one: quick range now, detail later.
• Treating loud as urgent. The homeowner who calls four times is not necessarily the one most likely to buy; sometimes they are the most anxious price-shopper. Score on the five factors, not on volume.
• One-and-done sending. Firing off a PDF and never following up wastes most of the value of a fast send. The cadence is the close.
• No outcome logging. Without a win-loss log, you can never tune your weights, and you keep making the same misjudgments. One word per closed estimate is enough.
• Over-promising install dates to win the send race. Speed on the estimate is good; lying about when the crew shows up is not. Tie the queue to capacity.
• Over-investing in courtesy bids. A custom hour-long proposal for a fourth-place bid is a gift to your competitor's close rate. Standardize and speed up low-probability sends.
• Crossing the claims line on storm jobs. Promising approval, payouts, or deductible relief to move a storm estimate up the list is both a sales mistake and a legal one. Sell thorough documentation and an accurate estimate; let the homeowner file and the carrier decide.

A one-page system you can put on the wall

Pull it together into something an estimator can run without thinking hard:

1. Capture intent in the driveway. One line per job: trigger, timeline, decision-maker. No estimate enters the queue without it.
2. Score the five factors (intent 30, time 25, win 20, value 15, effort 10). Add to 100.
3. Sort high to low. 70+ is today, 50 to 69 is this week, below 50 is when there is air.
4. Hit the turnaround targets. Same-day for hot, 24 to 48 hours for warm, 3 to 5 days for cool. Protect a daily block for hot sends.
5. Schedule the next touch the moment you send. Run the five-step follow-up cadence; never leave an estimate with no next action.
6. Check capacity weekly. Open crew-days versus signed work. Adjust the push, never oversell.
7. Log every outcome in one word and review the weights each quarter.

That is the entire system. It fits on an index card, and it will change which estimates leave your shop first, which is the same as saying it will change how much you close.

The bottom line

The order you send roofing estimates is a decision your company makes dozens of times a week, usually by accident. Made on purpose, with intent and external deadlines weighted heaviest, value and effort weighted lightest, and a clock attached to each tier, it pulls your highest-intent homeowners to the front of the line and gets them paper while they are still ready to sign. The pile on the desk stops being a queue and becomes a ranked list of your next jobs. And if you also fix the inputs, choosing to measure the roofs that are genuinely due rather than whichever door happened to call, the whole funnel gets easier to sort, because more of what lands in the pile deserves to go out first.

FAQ

Should I always send the oldest estimate request first?

No. First-in-first-out feels fair but ignores readiness. An estimate created today for a homeowner with an active leak should usually go out before one created last week for a comparison shopper with three years of roof life left. Sort by intent and any external deadline, not by the date the request came in.

What single factor matters most when ranking estimates?

Homeowner intent, how ready they are to buy, read from the trigger that made them call, their timeline, and how simple the decision is. It outpredicts roof condition and job size. Capture one intent line at every appointment; it is the most valuable note in your CRM.

How fast should I send an estimate to a high-priority job?

Same day, ideally within a few hours, for any job that scores hot (high intent or a live deadline). Close rates drop steeply as response time stretches from hours into days. Protect a fixed daily block of estimator time so the same-day promise is real and not aspirational.

How do I prioritize storm and hail estimates without crossing legal lines?

Move them up because the deadline is real, many policies require prompt notice of loss and have documentation deadlines, but stay strictly on the documentation side. Inspect, photograph, and write an accurate repair estimate, then hand it to the homeowner. The homeowner files and the insurer decides. Do not promise approval, a payout, or deductible relief, and do not negotiate or interpret the claim; in many states that is unlicensed public adjusting.

Should the biggest-dollar job always be the priority?

No. Sort by expected gross profit times the probability you win, not by sticker price. A large, thin-margin job you probably will not win, or will not win for two months, is worth less today than a smaller job at a healthy margin that closes this week.

What about a referral from a past customer?

Give it a one-tier bump and, more importantly, a personal delivery, a call rather than just an emailed PDF. The cost of being slow on a referral is reputational; your referrer hears about it. Honor the relationship even when the cold factors score it only moderately.

How many times should I follow up after sending an estimate?

Plan on several touches over about two weeks: a same-day confirmation, a day-2 questions call, a day-4-to-5 value touch, a day-8-to-10 direct ask, then a long-cycle nurture rather than a hard stop. Most estimates that close need more than one contact, and most estimators quit too early. Schedule the next touch the moment you finish the current one.

Can I sell more roofs than my crews can install by sending too aggressively?

Yes, and it backfires. If you oversell, a fast estimate turns into a six-week schedule and the homeowner who signed on urgency cools off waiting. Check open crew-days against signed work weekly. Keep sending hot estimates, but be honest about install timing and ease off the low-intent tier when you are oversold.

Do I need software to prioritize my estimate backlog?

No. A five-factor, 100-point score on a sticky note or a spreadsheet works. Software helps once volume grows, mainly by automating follow-up reminders and logging outcomes so you can tune your scoring weights. The discipline of always working from the top of a ranked list matters more than the tool.

How do I get a backlog full of winnable jobs in the first place?

Prioritization sorts the pile, but the pile's quality is set by which roofs you choose to measure. Roof age is the best single signal of who is near a replacement, and storm exposure stacks on it, yet county year-built and listing sites miss re-roofs and hail maps only show where it hailed. Tools like RoofPredict estimate roof age as a range house by house and model storm wear per roof to rank the doors most likely to be due, so more of what enters your backlog deserves to go out first. Age comes back as a range and storm modeling gives odds, not proof, so you still inspect and verify.