5 Estimator KPIs That Drive Roofing Sales Performance

The five estimator KPIs that actually move a roofing company's sales performance are: qualified estimate rate, estimate cycle time, scope accuracy and change-order rate, gross margin variance by estimator, and follow-up quality with clean closed-lost reasons. Together they answer the only questions that matter at a Monday sales meeting: are we chasing the right houses, are we getting estimates out fast enough to win, did the estimate match the roof we actually built, did the job make the money we promised, and do we know why we lost the ones we lost.

Most roofing scorecards fail for one boring reason. They track raw close rate and nothing else. Close rate alone is a vanity number. An estimator can post a 60 percent close rate by discounting every job into the ground, skipping the steep-and-tall work, and writing soft scopes that blow up in production. Another estimator can sit at 28 percent while holding margin, walking away from poor-fit storm chasers, and handing production a clean job every time. The first one looks like a star on a single chart. The second one is the one keeping your doors open.

The goal here is a balanced estimator scorecard you can run weekly, built on definitions everyone agrees on, with sources of data nobody argues about. You will get the exact formulas, the targets to set from your own history (not borrowed averages), the edge cases that wreck the math, and copy-ready templates you can paste into your CRM today. Where insurance work and advertising claims touch these KPIs, you will get the legal boundaries straight, because a fast close that crosses a public-adjusting or false-advertising line is not a win. It is a lawsuit waiting on a calendar.

If you do nothing else, do this: pick consistent definitions, assign each KPI an owner and a data source, and pair every sales-outcome metric with a quality control. That single move turns a scoreboard into a coaching system.

Why most roofing estimator scorecards measure the wrong things

Walk into ten roofing offices and you will see the same dashboard: leads, appointments, sales, revenue. All lagging. All easy to game. None of them tell a sales manager what to do differently next Tuesday.

The problem is that revenue is an outcome, not a behavior. You cannot coach revenue. You can coach whether an estimator followed up inside two business days, whether the inspection photos were complete, whether the scope captured the ridge vent and the pipe boots, whether the bid held margin after production closed it out. Those are the inputs that create revenue, and they are the things a good estimator KPI measures.

There is a second failure mode, and it is sneakier. Inconsistent definitions. If one estimator counts every web form as a "lead" and another counts only inspected roofs, their close rates are not comparable and any ranking you build from them is fiction. If one project manager logs change orders as line items and another buries them in a notes field, your margin-variance report is noise. The single most valuable thing you can do before tracking a single KPI is write down what each term means and make everyone use it. Boring. Decisive.

Third, beware borrowed benchmarks. You will read that a "good" roofing close rate is north of 30 to 40 percent, that referral leads should close above 50 percent, and that paid-ad leads often land below 20 percent. Those ranges, reported across vendor surveys and industry write-ups, are directionally useful and worth knowing. But your number depends on your lead mix, your market, your price position, and how you define a qualified estimate. A retail-replacement shop running referral and past-customer work will and should look nothing like a storm-response crew knocking fresh hail zones. Set targets from your own trailing twelve months, then compare each estimator to your baseline, not to a blog's average.

With that framing, here are the five.

KPI 1: Qualified Estimate Rate

Qualified estimate rate measures how many inbound inquiries turn into real, estimate-ready opportunities. It is the first KPI because it protects every number that comes after it. If you do not separate "someone called" from "someone we can actually bid," your close rate, your cycle time, and your estimator's reputation all get distorted by junk that never belonged in the pipeline.

The formula

Qualified Estimate Rate = Qualified Estimate Opportunities / Total Inquiries

The whole KPI lives or dies on the word qualified. Define it once, in writing, and hold the line. A workable definition for most residential roofers requires all of the following before an inquiry counts as a qualified estimate opportunity:

• A real service address inside your service area
• The property owner or an authorized decision-maker identified
• A stated roof concern or a genuine request for an estimate or inspection
• Contact permission and a working phone or email
• A job type that fits what you actually sell (you do not do flat commercial TPO? a warehouse inquiry is not qualified)
• No disqualifier on file (rental with an absent owner, active dispute, prior non-payment)

Measure it by source, never blended. Website estimate forms, referral calls, door-knock leads, storm-response calls, property-manager requests, and reactivated past customers behave like different species. Blending them into one rate hides where your estimator is genuinely strong and where the channel is just feeding garbage.

A sample qualified-rate table

The table is the point. Two estimators with the same blended close rate look identical until you split by source and discover one of them is feeding on 86-percent-qualified past-customer work while the other is grinding 24-percent aggregator leads. Same effort, very different deck.

Where the right-house question comes in

Qualified estimate rate is also where targeting pays off before anyone climbs a ladder. The single biggest lever on this KPI is the list an estimator works. Knocking a subdivision where every roof was replaced after the last hail event is a guaranteed low qualified rate and a demoralized canvasser. Knowing which homes are actually due for work changes the math.

This is the honest fit for property intelligence. Tools like RoofPredict attach an estimated roof-age range and per-home storm exposure to addresses, so an estimator can skip the brand-new roofs and concentrate on homes a storm likely wore out. It does not inspect the roof, diagnose damage, or certify remaining life, and the age is a planning range, not an exact date. What it does is improve the input to this KPI: a better list raises the qualified rate honestly, instead of forcing the estimator to pad the definition of "qualified" to hit a target.

If you run paid online channels, instrument the inquiry side cleanly. Google's key events in GA4 let you track meaningful actions such as form submits, call clicks, and booking clicks, and the GA4 recommended events reference helps you standardize event names so the data ties back cleanly to source. Just remember GA tells you about inquiries, not qualification. Qualification is a human judgment you log in the CRM.

Common mistakes on KPI 1

• Counting appointments as qualified. An appointment with a renter who cannot authorize work is not a qualified estimate opportunity.
• Letting the definition drift to hit a target. If qualified rate jumps 20 points in a month, suspect the definition before you celebrate.
• Blending sources. Always split. The blended number is almost meaningless.
• Punishing the estimator for a bad list. A low qualified rate on aggregator leads is a marketing problem, not an estimator problem. Read the source column first.

How qualified rate feeds your true close rate

Qualified estimate rate is the denominator most roofers get wrong on close rate. There are two honest close-rate numbers and you should run both. Inquiry-to-close (signed jobs divided by all inquiries) tells you how well the whole funnel works, marketing included. Estimate-to-close (signed jobs divided by qualified estimates delivered) tells you how well the estimator works once a real opportunity is in front of them. An estimator buried in unqualified inquiries can post a brutal inquiry-to-close number and a strong estimate-to-close number at the same time. If you only look at the first, you blame the estimator for a marketing problem. If you only look at the second, you miss that the channel is drowning them in junk. Run both, side by side, by source. The gap between them is one of the most useful diagnostics in the whole scorecard: a wide gap means your qualification or your lead sourcing is leaking, and the fix is upstream of the estimator entirely.

KPI 2: Estimate Cycle Time

Estimate cycle time measures the elapsed time from a qualified opportunity to a delivered estimate, and then from delivered estimate to a decision. Speed wins roofing jobs more often than price does, because the homeowner who gets a clear proposal first, while the storm is still fresh and the ceiling stain is still wet, usually signs first. Industry write-ups consistently put residential decision windows in the days-to-weeks range and commercial in the months range, which means a two-day lag on a retail replacement can be the whole ballgame.

But speed is only half the metric. A fast estimate built on a missing measurement or a guessed scope is not a fast win. It is a future change order with a deadline. So this KPI is always read next to KPI 3.

Break cycle time into stages

Do not track one blob of time. Track the stages, because each one points at a different fix:

Stage 1: Qualified lead  ->  inspection scheduled
Stage 2: Inspection scheduled  ->  inspection completed
Stage 3: Inspection completed  ->  estimate delivered
Stage 4: Estimate delivered  ->  first follow-up
Stage 5: Estimate delivered  ->  signed or closed-lost

When the total looks slow, the stage view tells you where. Stage 1 dragging means a scheduling or capacity problem. Stage 3 dragging usually means missing inputs: no clean measurements, no supplier quote back yet, no decking note from the inspection. Stage 4 dragging is a follow-up discipline problem and it is the most common and most fixable of all.

Set different clocks for different work

One cycle-time target for every job type is a mistake. An emergency repair, a retail replacement, an insurance-related claim job, a multi-building commercial bid, and a specialty-material job (slate, standing-seam metal, cedar) all carry different honest timelines. Hold each to its own clock.

The follow-up clock is the cheapest win in roofing

Most lost residential roofing jobs are not lost on price. They are lost on silence. The estimate goes out, the estimator gets busy, three days pass, the homeowner signs with the contractor who called back. Tracking Stage 4 as its own number, with a hard target like first follow-up within two business days on every delivered estimate, recovers more revenue than almost any pricing change. It costs nothing.

Speed never excuses a bad claim

When estimators lean on urgency, financing, warranty, or energy-savings language to compress the decision, every one of those claims still has to be true and supportable. The FTC's advertising and marketing basics lay out the standard plainly: claims must be truthful, not misleading, and backed by evidence. "This shingle will cut your cooling bill in half" or "the manufacturer's strongest warranty" said to rush a signature is a fast cycle time that bought you a complaint. Coach speed on the process (scheduling, takeoff, follow-up), not on the pitch.

Common mistakes on KPI 2

• One clock for all job types. Commercial and emergency work get judged against the same target and somebody always looks bad unfairly.
• Measuring delivery but not follow-up. Stage 4 is where the money leaks.
• Rewarding speed that skips inputs. A 24-hour estimate with no decking note is a change order in disguise.
• Ignoring the inspection-to-estimate gap. If Stage 3 is your bottleneck, the fix is usually a measurement or supplier-quote workflow, not pressure on the estimator.

KPI 3: Scope Accuracy and Change Order Rate

Scope accuracy tracks whether the estimate described the work production actually performed. It is the bridge between the sales department's promises and the field's reality, and it is where margin quietly dies. An estimator can win on volume and speed and still bleed the company dry if every other job needs a change order for a skylight nobody priced, a layer of decking nobody flagged, or an access problem nobody saw from the driveway.

Change orders are not automatically bad

This is the nuance that separates a fair scorecard from a witch hunt. Some change orders are legitimate and unavoidable: rotten decking you cannot see until tear-off, a code upgrade the permit triggers, a customer who chooses to upgrade to a designer shingle mid-job, a hidden chimney flashing failure. Punishing an estimator for discovered conditions teaches them to pad every bid and lose work. The KPI's job is to separate the honest discovery from the avoidable miss.

So categorize every change order. This is the most important table in your whole scorecard:

When you read this table monthly, patterns scream at you. If one estimator generates the same "missed visible component" miss on steep cut-up roofs over and over, that is a coachable, specific behavior, not a vague "improve scope accuracy" goal.

Tie it to real measurement discipline

Material-quantity misses usually trace back to takeoff quality. The NRCA roofing guidelines and the NRCA Roofing Manual remain the established reference for how roof assemblies are specified and measured, and aerial and satellite measurement is now treated as accepted industry practice. A disciplined estimator applies a documented waste factor, typically in the single-digit to low-double-digit percentage range depending on roof complexity and pattern, rather than a gut number that changes by mood. Variance against a documented waste factor is measurable. Variance against a vibe is not.

Flag code and permit review early

Scope misses also come from late code awareness. The model codes give you the frame of reference: residential roof assemblies are covered in IRC Chapter 9 and commercial roof assemblies in IBC Chapter 15. Do not treat a model-code page as proof of what a specific job requires. Local adoption, amendments, the permit's actual scope, and manufacturer installation instructions all govern. The KPI question is narrower and fairer: did the estimator flag code and permit review early enough for the job to be priced right? An estimator who notes "jurisdiction requires ice-and-water to code, drip edge, and a re-nail inspection; price accordingly" is doing the job. One who discovers it at the permit desk is creating a change order.

Access and safety belong in the scope

An estimate that ignores how the crew safely gets on and works the roof is underpriced, full stop. Steep slope, three-story height, tight lot access, fragile low-slope edges, and rooftop equipment all change the labor and the staging. OSHA's fall protection requirements are not optional and not free; the general-industry and construction standards drive real cost into staging, anchors, and crew time on certain roofs. A scope that prices a 12/12 cut-up like a walkable ranch is not a competitive bid. It is a loss the field will eat. Reward estimators for documented access and safety assumptions, and treat "didn't price access" as a scope miss, not a clever low number.

Common mistakes on KPI 3

• Treating all change orders as failures. You will teach padding and lose jobs.
• Not categorizing change orders. Without categories you cannot tell discovery from a miss.
• No documented waste factor. Quantity variance becomes unmeasurable.
• Pricing access and safety as an afterthought. It is one of the most common silent margin killers.

KPI 4: Gross Margin Variance by Estimator

Gross margin variance compares the profit you estimated on a job to the profit the job actually produced, attributed to the estimator who bid it. This is the KPI that ties sales behavior to the bank account, and it is also the one most likely to be wrong, because it is only as good as your job-costing discipline. Reported roofing gross margins commonly sit in a wide band that varies sharply by job type, with repairs running far richer than full replacements, so always read margin by job type, never blended.

The formula and the inputs

Gross Margin %         = (Revenue - Direct Job Costs) / Revenue
Gross Margin Variance  = Actual Gross Margin % - Estimated Gross Margin %

Direct job costs are materials, direct labor, subcontractors, dump and disposal, permits, equipment rental, and warranty/callback cost tied to the job. To attribute variance fairly, you have to capture the moving parts:

That last row is the silent killer. A change order that production performed but sales never got the customer to approve and pay is pure lost margin, and it shows up as the estimator's variance unless you label it.

Your scorecard is only as good as your records

Margin variance turns into opinion the second the cost record is weak. The IRS recordkeeping guidance for small businesses is a useful north star here: records should support every transaction. In roofing terms, that means purchase orders, supplier invoices, supplier credits, subcontractor bills, signed change orders, crew timesheets, and the original signed scope all need to tie to the job number. If your actual costs live in a foreman's text messages and a stack of receipts, your estimator scorecard is a fight every month. Keeping the lead source, signed scope, photos, costs, change orders, and handoff notes attached to one property record, the kind of recordkeeping that tools like RoofPredict help organize alongside a roofer's CRM, is what makes the variance number defensible instead of debatable.

Separate avoidable variance from the market

Do not punish an estimator for price moves they could not have known. Material input prices swing, and you can watch them: the BLS Producer Price Index program publishes the asphalt-shingle and roofing-materials series that drive a big share of a roof's cost. Local demand context matters too; the Census Bureau's Building Permits Survey is a clean read on how hot your local construction market is, which moves both labor availability and pricing. When a supplier raises shingle pricing mid-quarter, that is a market variance, not an estimating error. Tag it as market, separate it from the estimator's controllable variance, and your coaching stays honest.

The variance buckets you must keep separate:

• Estimator-controllable: quantity errors, missed components, unpriced access, soft labor assumptions, failure to capture an approved change order.
• Market-driven: material price changes, labor-rate spikes, fuel and disposal cost moves.
• Production-driven: rework, damaged material, blown labor on execution, callbacks.
• Customer-driven: mid-job upgrades, delays, scope changes the customer requested.

Only the first bucket belongs on the estimator's scorecard.

Common mistakes on KPI 4

• Blending margin across job types. Repairs and replacements have different honest margins; the blend hides everything.
• No clean cost record. Garbage in, argument out.
• Blaming the estimator for market or production variance. Tag the buckets or lose your estimators' trust.
• Ignoring uncollected change orders. This is often the single largest hidden margin leak.

KPI 5: Follow-Up Quality and Closed-Lost Reasons

The last KPI measures whether opportunities move through a clean decision process and whether you actually learn from the ones you lose. A roofing company that writes a hundred estimates a month and cannot tell you why it lost forty of them is flying blind. Follow-up quality plus disciplined closed-lost reasons turn losses into a feedback loop that fixes the other four KPIs.

What to track

• First follow-up completed on time (the Stage-4 clock from KPI 2)
• Number of touches before a decision
• Decision-maker confirmed (not only "talked to someone")
• Estimate revised and re-sent when the customer asked
• Closed-won reason
• Closed-lost reason, from a fixed picklist
• No-decision / stalled reason
• Review or referral request after a completed job

Make closed-lost a picklist, not a free-text shrug

If closed-lost is a free-text field, you will get "price" on ninety percent of records and learn nothing. Force a structured picklist so the data is analyzable:

CLOSED-LOST REASON (pick one):
[ ] Price - lost to a lower bid
[ ] Price - customer not ready to spend
[ ] Went with another contractor (not price)
[ ] Timing - deferred work
[ ] Scope - we don't do this work / out of area
[ ] Insurance - claim not approved by carrier
[ ] No response - went dark after estimate
[ ] DIY / handyman
[ ] Lost on speed - competitor proposed first

Now patterns become visible. If "lost on speed" is climbing, your cycle time (KPI 2) is the problem. If "price - lost to a lower bid" clusters on one estimator, look at their scope and discounting. If "no response" dominates, your follow-up discipline is broken. The picklist is what makes the loss column teach you something.

The insurance line you do not cross

Closed-lost on insurance work is where roofers get themselves into legal trouble, so read the boundary carefully. An estimator can document observed roof conditions, photograph them, measure the roof, and provide a scope and price. The estimator can hand the homeowner a clean, factual report that supports the homeowner's own claim. What the estimator, your company, and any "we" in your marketing cannot do is negotiate, manage, adjust, settle, or "maximize" the claim, or promise approval, coverage, or payment. The insurer decides coverage. Period.

This is not a soft suggestion. In 2024 the Texas Supreme Court ruled in the Stonewater Roofing case that a roofing contractor advertising itself as able to handle and negotiate insurance claims was engaged in unauthorized public adjusting, and that the licensing law restricting that conduct did not violate the contractor's free-speech rights, as reported by Insurance Journal. Public-adjuster licensing is enforced state by state; the Texas Department of Insurance's agent and adjuster licensing pages are one example of the regulator that governs who may act on an insured's behalf. The NAIC's consumer guide on filing a homeowners claim is a good plain-language reminder of where the homeowner's and carrier's roles sit.

Here is the safe "say this, not that" boundary every estimator should know cold:

That deductible line matters as much as the adjusting line. Offering to waive, rebate, or absorb a homeowner's insurance deductible is insurance fraud in many states. It is not a sales tactic. It is a felony exposure. A closed-lost reason of "insurance not approved" is a legitimate, useful data point. A closed-won reason that involved any of the left-column phrases is a problem you want to find in a KPI review before a regulator does.

Reviews and email, inside the lines

Follow-up often includes review requests and email outreach, and both have rules. The FTC's endorsement and review guidance governs how you can solicit and use reviews and testimonials: do not gate them on a positive experience, do not write fake ones, disclose material connections. The FTC's CAN-SPAM compliance guide governs commercial follow-up email: honest subject lines, a real physical address, a working unsubscribe. Track review-request completion as a KPI, but pair it with a compliance check so you are not rewarding pressure tactics or fabricated testimonials.

Common mistakes on KPI 5

• Free-text closed-lost. You will learn nothing. Use a picklist.
• Crossing the public-adjusting line to win. A fast close that crosses it is a liability, not a sale.
• Any talk of waiving deductibles. Stop. It is fraud in many states.
• Soliciting reviews without FTC-compliant language. A win on the board, a problem in the file.

Put the five together: the estimator scorecard

Five KPIs only work as a system. Run them on one page, reviewed weekly, with each metric owned by someone and sourced from somewhere nobody disputes.

The pairing column is the whole philosophy. Never reward one number alone. Reward close rate with closed-lost reason quality. Reward speed with scope accuracy. Reward margin with captured-and-approved change orders. A single-metric incentive always breeds a single-metric distortion.

A copy-ready weekly scorecard

ESTIMATOR WEEKLY SCORECARD - [Name] - Week of [date]

PIPELINE
- New qualified opportunities .......... ___  (by source: ___)
- Qualified estimate rate (rolling 30d)  ___%
- Estimates due this week .............. ___
- Estimates delivered .................. ___
- Median estimate cycle time (by type)   ___ days

FOLLOW-UP
- First follow-up within 2 biz days .... ___ / ___ delivered
- Open decisions awaiting follow-up .... ___
- Closed-won this week ................. ___
- Closed-lost this week (with reasons) . ___

QUALITY / FLAGS
- Missing inspection photos ............ ___ jobs
- Change orders opened (by category) ... ___
- Any insurance-language flag? ......... Y / N
- Any review-request compliance flag? .. Y / N

HANDOFF
- Jobs sent to production this week .... ___
- Handoff completeness (complete/minor/critical): ___

Production handoff: the KPI most scorecards forget

Most estimator scorecards stop at the signed contract. That is a mistake, because a huge share of production failures and margin misses are born at the sales-to-production handoff. An estimate can be fast, accurate, and profitable on paper and still detonate in the field because the estimator never recorded the shingle color, the decking note, or the access plan.

Score the handoff. Before a job moves from sales to production, the estimator should have delivered:

PRODUCTION HANDOFF CHECKLIST
[ ] Signed scope of work
[ ] Approved price and payment terms
[ ] Inspection photos (overview + each slope + penetrations + problem areas)
[ ] Measurements / takeoff record with waste factor
[ ] Material selections (product line, color, accessories)
[ ] Underlayment, ventilation, and flashing spec
[ ] Warranty expectations set with customer
[ ] Access, staging, and safety notes
[ ] Permit / code-review flags
[ ] Customer communication notes (preferences, constraints, gate code, pets)
[ ] Approved change requests (if any)
[ ] Open questions for production

Score each handoff as complete, missing minor items, or missing critical items. A critical miss is the unpriced skylight, the wrong color, the missing decking note, the unapproved upgrade, or an unsupported insurance statement in the file. This single KPI lets sales and production argue about facts instead of personalities. "Three of your last ten jobs went to the crew with no color selection" is a conversation. "Production hates working with you" is a fight. Requiring the handoff fields to be filled before the job advances, the kind of gate a connected record system like RoofPredict can enforce, keeps the field crew from inheriting holes the sales department dug.

Setting targets without creating bad incentives

Targets shape behavior, and the wrong target reliably produces the wrong behavior. Set them deliberately.

• Reward close rate alone and estimators discount to win, write soft scopes, and chase poor-fit jobs.
• Reward gross margin alone and estimators pad bids and walk away from winnable work.
• Reward estimate speed alone and scope accuracy collapses into a change-order machine.
• Reward job count alone and the handoff falls apart under volume.

The fix is the balanced pairing you already saw, enforced in the compensation plan. A commission plan that pays purely on signed contract value ignores rework, missing scope, uncollected change orders, and low-margin jobs, and it quietly rewards the estimator who creates the most production losses. A better structure ties part of the payout to confirmed margin after production, or holds back a portion pending a quality and handoff review. The exact compensation mechanics need payroll and legal review in your state, but the operating principle is simple and non-negotiable: do not pay people to create losses downstream.

And set the targets from your trailing twelve months. Pull your own qualified rate by source, your own cycle time by job type, your own margin by job type, and set each estimator's goal as a realistic step from your baseline. Borrowed benchmarks make for good blog reading and bad targets.

Account for territory and job mix

Apply the same definitions to everyone, but do not force the same targets on different work. The estimator running commercial low-slope bids with six-month decision cycles cannot be held to the same cycle time as the one closing referral retail replacements in three days. The storm-response estimator documenting roofs for insurance-related work carries documentation time the retail estimator does not. Same yardstick, different expectations, and that is fair.

Data hygiene: the unglamorous thing that makes all of this real

Every KPI above is downstream of clean data, and roofing CRMs rot fast. Set explicit rules and enforce them, or your trend lines become fiction inside a quarter.

Write down, for your whole team, the answer to each of these:

• When is a lead created?
• When does a lead become qualified?
• When is an estimate considered delivered (sent? viewed? presented in person?)?
• When is a job closed won?
• When is a job closed lost?
• Which costs count as direct in gross margin?
• How are customer-requested upgrades categorized vs. discovered conditions?
• How are canceled jobs counted?
• How are duplicate leads merged?

Then run a monthly cleanup. Hunt for missing lead sources, blank closed-lost reasons, estimates with no delivered date, jobs with no final cost record, and change orders with no category. Those gaps are not only ugly reports. Each one blinds a sales manager to whether the estimator needs coaching, the channel is weak, or production is inheriting holes.

A practical trick: make the most-skipped fields required at the stage where they matter, not all at once. Requiring a closed-lost reason the moment a job is marked lost, and a lead source the moment a lead is created, catches the two gaps that wreck reporting most, without burdening the estimator with a wall of fields they will learn to fake. Required-at-the-right-moment beats required-everywhere every time, because a field an estimator resents is a field an estimator games.

Good hygiene protects the estimator as much as the company. If a customer changed scope after signing, that should be tagged so it never lands on the estimator's variance. If a supplier raised pricing after the bid, that is market variance, labeled and separated. If production damaged material or skipped a documented step, the record should show the estimating was clean. Clean data is what lets you be fair, and fairness is what keeps good estimators from quitting.

A coaching cadence that actually changes behavior

KPIs only improve performance if you review them while the work is fresh and turn each review into one or two concrete changes, not a list of vague aspirations.

Weekly (active pipeline): new qualified opportunities, estimates due and delivered, overdue follow-up, open decisions, missing inspection records. Fast, fifteen minutes, forward-looking.

Monthly (trends): gross margin variance by job type, scope-miss categories, closed-lost reason patterns, production handoff scores, lead-source quality. This is where you spot the repeating miss.

Quarterly (structure): estimator capacity, territory performance, commission structure, training needs, template and checklist changes, channel budget shifts.

Keep the coaching brutally specific. "Improve your close rate" is useless. "Add ventilation and pipe-boot lines to every estimate before delivery" and "call back every retail estimate within two business days" are behaviors an estimator can actually change by Friday. When a KPI flags a problem, pull the underlying estimate, the photos, the notes, the lead source, and the handoff record before you assign blame. Half the time the data exonerates the estimator and indicts the process, and that is exactly the kind of finding that makes the whole system pay for itself.

The payoff of running all five together, with clean data and tight coaching, is that you stop guessing. You know which inquiries to chase, how fast you have to move to win them, whether your bids hold up against the build, whether the work made money, and why you lost what you lost. That is what separates a roofing sales team that is merely busy from one that is genuinely performing.

Sources checked: June 18, 2026.

FAQ

What KPIs should a roofing estimator track for sales performance?

Track five core metrics: qualified estimate rate (how many inquiries become real, biddable opportunities, measured by source), estimate cycle time (days from qualified lead to delivered estimate to decision, by job type), scope accuracy and change-order rate (did the bid match what production built, categorized by cause), gross margin variance by estimator (estimated profit versus actual, separated into estimator, market, and production buckets), and follow-up quality with structured closed-lost reasons. Pair every sales-outcome metric with a quality control so no single number can be gamed.

What is a good close rate for a roofing company?

It depends heavily on your lead mix, so set targets from your own history rather than a published average. As rough industry context, vendor surveys and trade write-ups often put a healthy blended residential close rate somewhere around the low-to-mid thirties percent, with referral and past-customer leads closing much higher and paid aggregator leads much lower. Always measure close rate by lead source, never blended, because a 30 percent rate on stale shared leads and a 30 percent rate on referrals reflect completely different performance.

How do you measure estimate cycle time in roofing?

Break it into stages rather than one number: qualified lead to inspection scheduled, inspection scheduled to completed, inspection completed to estimate delivered, estimate delivered to first follow-up, and estimate delivered to signed or closed-lost. Track the median per stage and set different targets per job type, since an emergency repair, a retail replacement, an insurance job, a specialty-material job, and a commercial bid all carry different honest timelines. The follow-up stage is usually the cheapest place to recover lost revenue.

Are roofing change orders always the estimator's fault?

No, and that distinction is the whole point of tracking scope accuracy. Discovered conditions like rotten decking found at tear-off, a permit-triggered code upgrade, or a customer-requested mid-job upgrade are legitimate and unavoidable. Avoidable misses are different: a skipped visible skylight, unpriced roof access, a wrong material quantity, or an unpriced safety requirement. Categorize every change order so you can separate honest discovery from a coachable estimating error, then coach only the avoidable category instead of teaching estimators to pad every bid.

How is gross margin variance attributed to a specific estimator fairly?

Capture estimated versus actual cost for materials, labor, subcontractors, disposal, rework, and uncollected change orders on a clean per-job record, then sort each variance into four buckets: estimator-controllable (quantity errors, missed components, soft labor assumptions, failed change-order capture), market-driven (supplier price moves you can track via the BLS Producer Price Index), production-driven (rework, blown labor, callbacks), and customer-driven (mid-job changes). Only the estimator-controllable bucket belongs on the estimator's scorecard, and fair attribution requires disciplined job-costing records tied to the job number.

Can a roofing estimator handle a homeowner's insurance claim?

No. A roofer can document and photograph roof conditions, measure the roof, and provide a scope and estimate that support the homeowner's own claim, but cannot negotiate, manage, adjust, settle, or maximize the claim, or promise approval or payment. The insurer decides coverage. The 2024 Texas Stonewater Roofing decision held that a contractor advertising itself as able to handle and negotiate claims was engaged in unauthorized public adjusting. Never offer to waive or absorb a deductible either; that is insurance fraud in many states.

Why should closed-lost reasons be a picklist instead of a text field?

Because a free-text field collapses into the word price on nearly every record and teaches you nothing. A fixed picklist (lost to a lower bid, customer not ready, went with another contractor, deferred timing, out of scope, insurance not approved, went dark, lost on speed) makes losses analyzable. Then patterns surface: rising lost-on-speed points at your cycle time, clustered lost-to-lower-bid on one estimator points at scope or discounting, and dominant went-dark points at broken follow-up discipline. Structured loss data feeds back into fixing the other KPIs.

Should production handoff be part of an estimator's KPIs?

Yes, because many production failures and margin misses begin at the sales-to-production handoff, not in the field. Score each handoff as complete, missing minor items, or missing critical items, where a critical miss is an unpriced skylight, wrong color selection, missing decking note, unapproved upgrade, or an unsupported insurance statement. Require signed scope, photos, measurements, material selections, access notes, permit flags, and approved change orders before a job advances. This lets sales and production discuss facts instead of personalities.

How can RoofPredict help with roofing estimator KPIs?

RoofPredict improves the inputs to these metrics rather than replacing the scorecard. By attaching an estimated roof-age range and per-home storm-exposure scoring to addresses, it helps estimators work a better list, skip brand-new roofs, and raise their qualified estimate rate honestly instead of padding the definition. Keeping lead source, photos, scope, costs, change orders, and handoff notes attached to one property record also makes margin variance defensible. It does not inspect roofs, diagnose damage, certify remaining life, or decide insurance coverage.