Roofing Canvassing Funnel Benchmarks: First Knock to Inspection Conversion

Most roofing companies that knock doors cannot tell you their conversion rate. They can tell you how many roofs they sold last month, and they can tell you which rep is "crushing it" based on a gut feeling, but ask them how many doors it takes to book one inspection and you get a shrug or a made-up number. That gap is expensive. A canvassing program is a funnel, and a funnel you do not measure is a funnel you cannot fix.

This breaks down the full path from the first knock to a roof on a ladder: the stages, the ratios that separate a healthy operation from a leaky one, the math to set quotas that are actually hittable, and the specific places where appointments quietly disappear. It is written for the owner or sales manager who already knocks doors and wants to stop guessing. Every number here is presented as a planning range, not a promise. Your market, your storm history, your crew, and your product mix will move these figures, sometimes a lot. The point is to give you a frame to measure against your own data, then beat it.

A note before the numbers: nothing here depends on promising a homeowner a free roof, a waived deductible, or an approved claim. Those promises are how roofers lose licenses and end up in front of a state attorney general. The funnel works because you show up, you document the roof honestly, and you make it easy for a homeowner to decide. Keep that frame and the rest follows.

The canvassing funnel, stage by stage

Before you can benchmark anything you have to agree on what the stages are. Most teams collapse the funnel into "knocks" and "sales," which hides every problem in between. Here is the full chain, and you should be able to count each one independently.

1. Doors approached — a rep physically walks up and knocks or rings. Not houses on the street. Not the territory count. An actual knock.
2. Doors answered — someone opens the door or engages through a video doorbell. This is wildly outside your control and is the single most misunderstood number in canvassing.
3. Conversations — the homeowner stays at the door long enough for the rep to deliver the hook and get a response. A grunt and a slammed door is an answer, not a conversation.
4. Inspections booked — the homeowner agrees to let you look at the roof, either right now or at a scheduled time.
5. Inspections completed — you actually get on or over the roof and document it. Booked is not completed; the no-show rate between these two is where a lot of "good" reps fall apart.
6. Inspections that find actionable damage or end-of-life condition — the roof genuinely warrants a repair, replacement, or an insurance conversation the homeowner can choose to start.
7. Files opened / contingency or inspection agreements signed — the homeowner commits to moving forward with you as the contractor.
8. Jobs closed and built — money changes hands, the crew installs.

The phrase "first knock to inspection conversion" usually means stages 1 through 5: doors approached all the way to a roof you actually got on. That is the right unit to manage, because it is the part of the funnel the canvasser controls and the part that feeds everything downstream. If you only measure approached-to-closed, you cannot tell whether a rep's problem is their knock count, their pitch, their booking, or your install crew's scheduling. Each stage needs its own number.

Why "doors answered" wrecks naive benchmarks

Here is the trap. A rep knocks 80 doors and books 4 inspections. Is that good? You cannot say, because you do not know how many people were home. If 70 of those 80 doors had nobody behind them, then 4 inspections off 10 conversations is excellent. If all 80 answered and the rep still only booked 4, that is a pitch problem, a targeting problem, or both.

Answer rates swing hard by daypart, day of week, neighborhood density, and season. Weekday mornings in a commuter suburb might answer at 15 to 25 percent. The same street on a Saturday late morning, or a weekday between roughly 5 and 7 p.m., can answer at 35 to 50 percent. Retirement-heavy neighborhoods answer far more often and at more hours. Newer subdivisions with two working adults and a video doorbell answer less and answer differently — the conversation now happens through a speaker while the homeowner is at the grocery store.

The practical move: track doors approached and doors answered as two separate columns, always. Then compute your real conversion off conversations, not off raw knocks. A rep with a 5 percent approached-to-booked rate and a 35 percent answer rate is converting roughly 14 percent of the people they actually talk to. A rep with the same 5 percent but a 20 percent answer rate is converting 25 percent of conversations — a much stronger closer who is being failed by their schedule or their territory.

Benchmark ranges you can plan against

These are planning ranges drawn from how door-to-door roofing programs typically behave, not guarantees. Treat them as a starting hypothesis to validate against your own logged data over a few hundred knocks. The single biggest variable is context: a fresh, well-defined storm-restoration territory behaves nothing like cold retail canvassing in a market with no recent weather event.

Cold retail / non-storm canvassing

Read that bottom row carefully. In cold, non-storm canvassing, getting onto 1 to 3 roofs per 100 doors knocked is a normal, workable outcome. New reps will sit below 1 percent. Seasoned reps with a tight pitch and good territory can push toward and past 3 percent. If someone is quoting you a 10 percent cold knock-to-inspection rate, they are either counting a hot storm territory, counting conversations as knocks, or selling you something.

Storm / post-event canvassing

After a verified hail or high-wind event in a defined damage footprint, every number improves, sometimes dramatically, because the homeowner has a reason to care and a clock running on their claim window.

The spread inside the storm column is enormous and it is mostly about targeting. Knocking the actual hail core where stones were large enough to bruise mats is a different business than knocking three miles out where the report was pea-sized. This is exactly where roof-by-roof storm modeling earns its keep, and we will come back to it.

Rep-day output, beyond ratios

Ratios alone do not pay payroll. You also need absolute throughput per rep per productive field hour. A focused canvasser working a decent territory will physically knock somewhere in the range of 25 to 60 doors per hour depending on lot size and density — rural acreage destroys this number, tight townhome rows inflate it. Over a real 4-to-6-hour evening field block, that is roughly 150 to 300 knocks. Run that against the ranges above and you can sanity-check any rep's claimed output. A rep who says they booked 8 inspections off "a couple hours" of cold knocking is rounding up somewhere; the math does not support it unless they were standing in a storm core.

Seasonality, weather, and the moving baseline

One more thing wrecks naive benchmarks: the same rep, the same script, and the same territory will produce different numbers in March than in October. Canvassing is a seasonal business and your baseline moves with the calendar.

In most U.S. storm markets, the field gets busy when severe weather does — roughly spring through early fall, depending on whether you sit in the southern plains, the upper Midwest, or the Southeast. During an active stretch, homeowners are primed, neighbors are comparing notes over the fence, and answer rates and conversation rates both climb. In the dead of winter, or in a long dry spell with no recent event, you are back to cold-retail economics no matter how good your reps are. If you benchmark your reps against their own July numbers in February, you will conclude the whole team fell apart when really the weather just changed.

The fix is to band your benchmarks by season and by event recency, not to hold one number all year. Keep a separate cold baseline and a separate post-event baseline, and when you compare a rep month over month, compare like to like. A rep whose knock-to-booked dropped from 8 percent to 3 percent did not necessarily get worse — if the storm territory aged out and they are back on cold streets, 3 percent might be a strong cold number. Tag every field block with its context so the comparison stays honest.

Daylight is the other quiet seasonal lever. In December your usable evening knocking window after most people get home might be 45 minutes before it is fully dark and porch lights make a knock feel intrusive. In June you have two-plus hours of comfortable evening light. That alone changes how many quality knocks a rep can physically get in, which changes their absolute output even if every ratio holds. Plan routes and quotas around the actual daylight window, not a flat hours-on-the-clock number.

Setting quotas from the funnel instead of from hope

The reason most canvassing quotas feel arbitrary is that they are set top-down — "everybody books two inspections a day" — without checking whether the funnel math even allows it. Build the quota from the bottom up instead.

Work an example. Say you want one signed file per rep per day in a cold market, and your validated funnel looks like this:

• Approached → booked: 2.5%
• Booked → completed: 75%
• Completed → finds actionable condition: 55%
• Actionable → file signed: 45%

Multiply the chain backward from one signed file:

• 1 signed file ÷ 0.45 = 2.2 inspections that found something
• 2.2 ÷ 0.55 = 4.0 completed inspections
• 4.0 ÷ 0.75 = 5.3 inspections booked
• 5.3 ÷ 0.025 = 213 doors approached

So "one file a day" in this cold funnel means roughly 210 quality knocks per rep per day. If your reps are knocking 120 and you are demanding a file a day, you have not set a goal, you have set a frustration. Either the knock count has to rise, the conversion has to improve, or the quota has to drop. The math is not optional.

Now run the same target in a storm territory at 8% approached-to-booked, 85% show, 70% find-rate, 55% sign:

• 1 ÷ 0.55 = 1.8 inspections that found something
• 1.8 ÷ 0.70 = 2.6 completed
• 2.6 ÷ 0.85 = 3.1 booked
• 3.1 ÷ 0.08 = 38 doors approached

Same one-file goal, 38 knocks instead of 213. This is why dropping a rep into a real storm footprint feels like a different job, and why measuring "doors per file" without noting storm vs. cold context produces garbage comparisons between reps.

The quota table every manager should keep

Keep a living version of this for your market, updated as your real data comes in. Even rough, it changes how you coach.

When a rep's knocks-per-booked number drifts far outside the band for their context, you have found a coaching target before it shows up as a bad month.

A worked example: comparing two reps the right way

Numbers in the abstract are easy to nod at and hard to use. Walk through a real comparison the way a sharp manager would.

Two reps turn in their week. Both worked five evening blocks. Here is the raw board:

The naive read is that Rep A is the workhorse — 640 knocks to Rep B's 410 — and Rep B is the lucky one who caught a storm. Now compute the ratios.

Rep A: answer rate 30%, conversation rate 15%, knock-to-booked 3.4%, show rate 55%, doors per file 160.

Rep B: answer rate 41%, conversation rate 33%, knock-to-booked 6.8%, show rate 86%, doors per file 46.

The storm context explains a lot of Rep B's edge, exactly as expected — but two numbers tell a coaching story the context does not. Rep A's show rate is 55 percent against Rep B's 86. Rep A is booking appointments and then losing nearly half before anyone gets on the roof. That is not a storm-versus-cold difference; that is a confirmation-and-same-day-inspection difference, and it is fixable this week. If Rep A simply lifted their show rate from 55 to 80 percent by inspecting on the spot and confirming the rest, their completed inspections jump from 12 to roughly 17 with zero extra knocking.

The second story: Rep A's conversation rate is half of Rep B's even after you account for the storm lift. Some of that is context, but enough of it points at the opener that it is worth a ride-along to hear Rep A's first seven seconds. The point of the exercise is that the raw board said "Rep A is the grinder, Rep B got lucky," and the ratios said "Rep A has a show-rate leak and maybe an opener leak, and both are coachable." That is the entire value of stage-level measurement.

Where the funnel actually leaks

Aggregate numbers hide the leak. You have to look stage by stage. In practice the appointments disappear at five specific places, and each has a different fix.

Leak 1: The first seven seconds at the door

Most lost conversations are lost before the pitch even starts. The homeowner opens the door already deciding whether you are a threat or a salesperson to get rid of. Reps who lead with their company name, a clipboard held like a weapon, and "do you have a few minutes" get the door closed on them at a brutal rate.

The fix is structural, not motivational. Standardize the opener around three things: a reason you are specifically on this street, a low-commitment ask, and an immediate exit ramp that lowers the homeowner's guard. Something in the shape of: "Hey, I'm not going to keep you — we're checking roofs on [street] because of the [month] storm that came through, and I noticed yours has [specific observable thing]. Mind if I point it out real quick?" It names the street, gives a concrete reason, and the "I'm not going to keep you" defuses the get-rid-of-them reflex. Reps who nail the first seven seconds routinely double their conversation rate versus reps who launch into a feature pitch.

Measure this leak by comparing each rep's answer rate to their conversation rate. If a rep gets doors open at the team average but conversations crater, the opener is the problem, not the territory.

Leak 2: Booking for "later" instead of now

The second giant leak is the gap between booked and completed. A homeowner who says "sure, come back Thursday" is a far weaker yes than one who walks you to the side of the house right now. Thursday bookings no-show, cancel, or "forget" at rates that can run 30 to 50 percent. Same-day or on-the-spot inspections complete at 85 to 95 percent because there is no gap for the yes to cool off.

The operational answer is to bias hard toward inspecting on the spot. If a rep is equipped to get on a ladder or fly a drone the moment the homeowner agrees, the funnel tightens by an entire stage. When a same-day look is impossible, the booking has to be locked: confirmed time, calendar invite or text confirmation sent before the rep leaves the porch, and a reminder the morning of. A booking with no confirmation sent is barely a booking.

Track this leak as a clean show-rate column. If your booked-to-completed is under 65 percent, the problem is almost never the reps' charisma — it is that you are booking soft future appointments and not confirming them.

Leak 3: The inspection that finds nothing and ends cold

Not every roof you get on warrants work, and that is fine — a roof in good shape with years of life left should be documented as exactly that. The leak here is different: it is the rep who finds a genuinely aging or storm-worn roof and still fails to translate what they saw into a decision the homeowner can make. They climb down, mumble "yeah it's got some wear," hand over nothing, and leave.

The fix is documentation discipline. Every completed inspection should produce photos of the actual conditions — granule loss, mat exposure, bruising, lifted or creased shingles, flashing failures, soft metals with impact marks — tied to specific locations on the roof, plus a written summary the homeowner can hold. When there is enough damage or end-of-life condition that an insurance conversation makes sense, the rep's job is to document thoroughly and prepare an accurate, Xactimate-aligned repair estimate for the homeowner's file. The homeowner is the one who decides to file a claim, and the insurer is the one who decides coverage. Your rep does not negotiate the claim, does not interpret the policy, does not promise an approval or a payout, and does not tell anyone their deductible will vanish. Cross that line and you are no longer doing roofing sales, you are doing unlicensed public adjusting, which is illegal in most states and a fast way to lose your contractor's license.

Leak 4: No system, so the data is fiction

You cannot fix a funnel you record on the back of a business card. The fourth leak is invisible because it is a measurement failure: reps self-report round numbers, managers compare apples to storms, and nobody knows the real ratios. "I knocked a ton today" is not data.

The fix is a canvassing app or CRM where every door gets a disposition tapped in the field — no answer, not interested, callback, booked, inspected — with the territory and timestamp attached. The instant you have honest dispositions, the leaks become visible per rep and per street, and coaching stops being a debate about whose memory is better.

Leak 5: Knocking the wrong doors at scale

The most expensive leak is also the quietest, because it never shows up as a bad conversation — it shows up as reps burning a full evening on streets where almost no roof is a real candidate. Two houses on a block might be 22-year-old roofs sitting in a verified hail swath. The other thirty are six years old and were tarped two streets over from the actual core. A rep with a perfect pitch still converts poorly there, because most of those doors had nothing to find.

This is a targeting problem, and it is upstream of every script and every quota. Fixing the pitch on a bad street is polishing the deck chairs. The leverage is in choosing better streets before anyone knocks.

Targeting: the upstream multiplier

Everything above assumes reps are walking decent territory. If they are not, no amount of script work saves the funnel. Two inputs decide whether a door is worth a knock: how old the roof likely is, and whether that specific roof took a real beating from a recent storm. Get both right and your approached-to-booked rate can jump several multiples — not because the reps got better, but because they are finally knocking roofs that genuinely need attention.

Roof age as a range, not a guess

A roof that is 18 to 24 years old is a fundamentally different sales conversation than one that is 4 to 8. The older roof is at or past the end of typical asphalt-shingle service life, the homeowner has probably already noticed a granule pile or a ceiling stain, and the inspection tends to find something real. The problem is you cannot see roof age from the street, and county permit records are patchy, missing for re-roofs done without a pull, and often years out of date.

Aerial and satellite imagery analysis can estimate a roof-age range per address across a whole territory — not an exact install date, a range, because that is what the imagery honestly supports. "This roof reads as roughly 17 to 22 years old" is enough to prioritize a street. Sort your territory by estimated age and your reps spend their evening on the blocks where the inspection actually finds wear, instead of on six-year-old roofs that waste the knock.

Storm exposure modeled per roof, not per county

The other half is storm exposure, and this is where most targeting goes wrong. The typical approach is to grab a hail report polygon — "1.5 inch hail reported in this county" — and knock everything inside it. But a county-level or even city-level report is far too coarse. Hail size, wind speed, and the angle and direction of the storm vary street to street and even roof to roof. A north-facing slope might be hammered while the south face on the same house is untouched. The actual report points from the field are sparse and biased toward where people happened to be looking.

Modeling storm exposure per roof — combining the storm's track, estimated hail size and wind field, and each roof's geometry and orientation — gets you much closer to which specific addresses likely took real damage. It is still odds, not proof: a model tells you a roof had a high probability of damaging impacts, and the inspection confirms or rules it out. But knocking the modeled high-exposure roofs first, instead of the whole county polygon, is the difference between a 4 percent storm street and a 12 percent one.

Where RoofPredict fits

This is the gap RoofPredict is built for. It scores a contractor's territory house by house: a roof-age range per address from aerial imagery, plus storm physics modeled per roof so you can see which specific addresses a given event most likely wore out. The output is a ranked door list and route — the roofs that are aging out and the roofs the storm hit hardest, sorted so your canvassers start where the inspection is most likely to find something real. It also enriches a list you already own: feed in your existing CRM or mailing list and it appends the roof-age and storm-exposure signals to addresses you were already going to work, so a mailer or a knock route gets pointed at the right doors.

The honest limits matter, and stating them is the point. Roof age comes back as a range, not a precise install date, because imagery cannot honestly give you a date. Storm exposure is a probability per roof, not a guarantee of damage — the model ranks likelihood, and the physical inspection is still what confirms condition. RoofPredict tells you which doors are worth the knock and the order to take them in. It does not knock the door, deliver the pitch, get on the ladder, or close the homeowner. The funnel discipline in the rest of this piece is still the work. What good targeting does is make every other stage convert better by feeding your reps roofs that deserve their time, and it is not a substitute for honest documentation or for staying on the right side of the claims line.

The inspection itself: turning a roof into a decision

The completed inspection is the hinge of the whole funnel. Everything before it is about earning the chance to look; everything after it depends on what you do with what you saw. Reps who treat the inspection as a formality — climb up, glance around, climb down — waste the hardest-won stage in the chain. Treat it as the moment you build the homeowner's file.

Work the roof systematically so nothing gets missed and the documentation holds up. A repeatable inspection pass looks like this:

1. Ground and elevation first. Before the ladder, walk the perimeter and photograph each elevation. Note soft metals — gutters, downspouts, fascia, AC condenser fins, mailbox, garage door. Impact marks on soft metals are some of the clearest physical evidence that hail of a damaging size actually fell at that address, and they are at eye level.
2. Field of each slope. On the roof, document each slope separately and label it by orientation (north, south, east, west). Storm exposure is directional, so a slope-by-slope record matters. Photograph granule loss, mat exposure, bruising you can feel as a soft spot, and any fractured shingles.
3. Penetrations and transitions. Flashing at chimneys, walls, valleys, and pipe boots is where roofs leak first and where age shows. Photograph cracked boots, rusted flashing, and failed sealant.
4. Wear versus impact. Be honest in the file about what is age-related wear and what looks like storm impact. Conflating the two is how documentation falls apart later. Note both, label both.
5. Overall condition and remaining life. A plain-language read for the homeowner: what you saw, roughly where the roof is in its service life, and what the realistic options are.

The output the homeowner keeps is a photo report tied to roof locations plus a written summary, and when conditions warrant repair or replacement, an accurate estimate. If the roof has storm damage the homeowner may want to involve their insurer, the right document is an itemized, Xactimate-aligned repair estimate that reflects your actual scope of work. You hand that to the homeowner. They decide whether to file. The insurer decides coverage. You never tell the homeowner the claim is approved, never promise a number, and never describe their deductible as something that disappears — your estimate is a statement of what the repair costs, not a prediction of what the carrier will pay.

The estimate hand-off that converts

The gap between a good inspection and a signed file is almost always presentation. A rep who shows the homeowner the actual photos on a tablet, points to the specific slope and the specific condition, and walks through the written summary converts far better than one who verbally summarizes "it's pretty worn." People decide on what they can see. Build the hand-off as a fixed sequence: show the elevations, show the slope photos, explain wear versus impact in plain words, present the estimate, and explain the homeowner's next step in their own decision. Standardizing that sequence is the single biggest lever on your find-rate-to-sign-rate conversion, and it is entirely within your control.

A week-one tracking system you can stand up immediately

You do not need to buy anything to start measuring. You need disciplined dispositions and a simple sheet. Here is a setup that produces real benchmarks inside one week.

Capture, per rep per field block:

• Date, daypart (e.g., 5–8 p.m.), territory/street name
• Doors approached
• Doors answered
• Conversations (real engagement)
• Inspections booked (and flag on-the-spot vs. scheduled-later)
• Inspections completed
• Inspections that found actionable condition
• Files signed
• Context flag: cold vs. storm territory

Compute, per rep weekly:

• Answer rate = answered ÷ approached
• Conversation rate = conversations ÷ approached
• Knock-to-booked = booked ÷ approached
• Show rate = completed ÷ booked
• Find rate = found-actionable ÷ completed
• Sign rate = signed ÷ found-actionable
• Doors per file = approached ÷ signed

Once you have two weeks of this across a few reps, your own benchmark bands replace every number in this piece. That is the goal — your data, measured honestly, beats anyone's generic averages.

Reading the numbers like a manager

The ratios tell you exactly where to coach, and they let you stop coaching the wrong thing:

• Low answer rate, normal everything else: schedule or territory problem. The rep is knocking dead hours or empty streets. Move their block to evenings or weekends; do not lecture them about pitch.
• Good answer rate, low conversation rate: the opener is failing. Drill the first seven seconds. Ride along and listen to the actual words.
• Good conversation rate, low knock-to-booked: the rep talks well but cannot ask for the inspection. Drill the booking ask and the on-the-spot offer.
• Good booking, low show rate: they are booking soft "come back later" appointments and not confirming them. Push same-day inspections and lock confirmations.
• Good show rate, low find rate: almost always a targeting problem. The rep is getting on roofs that have nothing wrong because the territory is full of newer or unaffected roofs. Fix the door list upstream.
• Good find rate, low sign rate: documentation and presentation. The rep finds real condition but cannot turn it into a decision. Standardize the photo report and the estimate hand-off.

Notice how often the diagnosis is upstream of the rep. A manager without stage-level data blames the canvasser for a territory problem and burns out good people. A manager with the funnel sees that the find rate is fine and the territory is the issue.

What pros consistently get wrong

A handful of mistakes show up over and over in canvassing programs that plateau.

Comparing reps across different contexts. Rep A in a storm core books 10 percent; Rep B cold-knocking a quiet suburb books 3 percent. Rewarding A and benching B is a mistake — adjusted for context, B might be the stronger closer. Always compare within context, or normalize to conversation rate, which strips out a lot of the territory noise.

Chasing knock count as the only metric. Yes, throughput matters, and a rep knocking 60 doors a night will usually beat one knocking 20. But pure knock count rewards spray-and-pray. A rep who knocks 200 garbage doors can look more productive on the board than one who knocks 90 well-targeted doors and books more. Pair knock count with knock-to-booked so you reward effective volume rather than raw volume alone.

Treating booked as done. Managers celebrate a big booking day and then wonder why installs lag. Booked is a promise; completed is a fact. The show rate is where optimism goes to die, and it is the most ignored number in the whole chain.

Letting reps freelance the claims conversation. A rep under pressure to close will reach for the easy line — "we'll get your whole roof covered," "don't worry about the deductible," "this is basically a free roof." Every one of those is a compliance landmine. They invite chargebacks, complaints, and regulatory action, and several states treat deductible-rebating and claim-negotiation by a contractor as outright illegal. Train the do-not-say list as hard as you train the pitch: do not promise approval or payout, do not say the deductible is waived or absorbed or gone, do not advertise a free roof, do not interpret the homeowner's policy, and do not negotiate with the carrier on the homeowner's behalf. The rep documents the roof, prepares an honest estimate, and hands it to the homeowner — the homeowner files, the insurer decides.

No feedback loop from inspection back to targeting. Every inspection is a data point about whether your targeting was right. If reps keep getting onto roofs that turn out to be six years old and undamaged, that is telling you the door list is wrong, and you should be feeding that back into how you choose streets. Programs that close this loop get sharper every storm season; programs that ignore it knock the same bad blocks every year.

Rep ramp and retention: the funnel behind the funnel

There is a second funnel that quietly determines whether your first one works: the one that turns a new hire into a productive canvasser. Door-to-door has brutal early attrition. A lot of new reps quit in the first two to three weeks, before they ever reach the conversion numbers that make them worth the training. If you do not manage the ramp deliberately, you pay to recruit and onboard people who leave right before they would have become profitable.

The ramp has a predictable shape. A brand-new rep in their first week is below 1 percent knock-to-booked in a cold market because their opener is stiff, they freeze on objections, and they take rejection personally. By week three or four, with reps drilling and ride-alongs, a keeper climbs toward the seasoned band. The reps who quit almost always quit during the worst stretch — the gap between "I've taken 200 rejections" and "I just booked my first inspection." Two management moves shorten that gap.

First, set early-week goals on activity, not outcomes. A first-week rep should be measured on quality knocks and clean dispositions logged, not on files signed. Asking a green rep to sign files in week one sets them up to feel like a failure during exactly the window when they are most likely to walk. Once their conversation rate stabilizes, shift the goal downstream.

Second, get them an early win fast, which usually means putting new reps on the best-targeted streets you have. A rep who books their first inspection on day two because they were knocking genuinely-aging or storm-worn roofs is far more likely to still be here in month two. Sending your greenest rep to the worst territory "to toughen them up" is how you lose people who would have been good. Good targeting is not only a conversion lever — it is a retention lever, because it lets new reps experience success before they decide to quit.

Safety is part of the funnel, not separate from it

The completed-inspection stage involves ladders and roofs, and an injury does more than hurt a person — it takes a producer off the board and exposes the company. Fall protection is governed by federal OSHA rules for construction work at height, and roof inspections are not exempt from basic ladder and fall-hazard sense. Build the safety expectations into the inspection workflow: stable ladder setup, awareness of slope and surface condition, no climbing wet or frosted roofs, and a clear policy that a rep who cannot safely get on a roof documents from a ladder, a pole camera, or a drone instead of forcing it. A drone or pole-camera pass also happens to lift your completion rate, because it lets a rep document a roof that is too steep, too high, or too fragile to walk. Safety and throughput point the same direction here.

A realistic 90-day plan to lift the funnel

If you are starting from "we knock but we don't measure," here is a sequence that compounds.

Days 1–14: Instrument. Get honest dispositions on every door, even on paper. Define the eight stages. Stop accepting self-reported round numbers. By day 14 you have a baseline knock-to-booked, show rate, and doors-per-file for each rep, split by cold vs. storm.

Days 15–30: Fix the cheapest leak first. Almost always that is the opener or the confirmation step. Standardize the first seven seconds and require a confirmation text on every scheduled booking. These two cost nothing and typically move conversation rate and show rate within two weeks.

Days 31–60: Fix targeting. Stop knocking by intuition. Rank territory by roof-age range and storm exposure so reps start on the doors most likely to convert. This is the stage where roof-by-roof data — age range plus per-roof storm modeling — does the heaviest lifting, because it lifts every downstream ratio at once. Re-measure; the find rate and knock-to-booked should both climb.

Days 61–90: Tighten the bottom of the funnel. Standardize the photo report and the estimate hand-off so a good inspection reliably becomes a signed file. Build the inspection-to-targeting feedback loop so bad streets stop getting re-knocked. By day 90 you have a measured funnel, fixed openers and confirmations, targeted routes, and a closing process — and you can set quotas from real math instead of hope.

Bringing it together

A canvassing program lives or dies on two things: knocking the right doors, and measuring what happens at each one. The first-knock-to-inspection conversion rate is not one number you chase — it is a chain of stages, each with its own leak and its own fix. Cold markets get you onto 1 to 3 roofs per 100 honest knocks; a well-targeted storm street can do several times that, and the difference is mostly upstream targeting, not rep heroics. Set quotas from the funnel math, compare reps within context, treat booked as a promise and completed as the fact, and keep every claims conversation strictly on the document-and-estimate side of the line.

The single highest-leverage move for most teams is to stop knocking by intuition. When your reps walk a route ranked by roof age and per-roof storm exposure — the roofs aging out and the roofs the storm actually wore down — every ratio below it improves at once. That is the part RoofPredict is built to hand you: which roofs are due, house by house, and the order to knock them. The funnel discipline is still yours to run, but you will be running it on the right doors. Pull one month of your own knock dispositions, compute the seven ratios, find your worst leak, and fix that one first. Then point your best reps at the roofs the data says are ready.

FAQ

What is a good first-knock-to-inspection conversion rate for roofing canvassing?

In cold, non-storm markets, getting onto 1 to 3 completed inspections per 100 honest knocks is a realistic, healthy range; new reps fall below 1 percent and seasoned reps with tight targeting push past 3 percent. After a verified storm in a well-targeted footprint, that completed-inspection rate commonly runs 4 to 12 percent. These are planning ranges to validate against your own logged dispositions, not guarantees, and the biggest swing factor is whether you are knocking the right doors.

How do I separate the knock-to-inspection funnel into measurable stages?

Track eight distinct stages: doors approached, doors answered, real conversations, inspections booked, inspections completed, inspections that found actionable condition, files signed, and jobs built. First-knock-to-inspection conversion specifically covers approached through completed. Measuring only approached-to-closed hides every problem in between and makes coaching impossible, because you cannot tell whether a rep's issue is their knock count, opener, booking, show rate, or your install scheduling.

Why is 'doors answered' so important to track separately from doors knocked?

Answer rate is largely outside the rep's control and swings from roughly 15 percent on weekday mornings to 50 percent on weekend late mornings or weekday evenings. If you only divide bookings by raw knocks, you cannot tell a closing problem from a scheduling problem. Computing conversion off actual conversations instead of raw knocks reveals whether a rep is a weak closer or simply being sent out during dead hours.

How many doors does a rep need to knock to book one inspection?

It depends entirely on context. In a cold market at a 1 to 3 percent knock-to-booked rate, expect roughly 33 to 100 knocks per booked inspection. In a strong storm territory at 6 to 12 percent, it can drop to 8 to 17 knocks per booked inspection. Build quotas backward from this math rather than imposing a flat number, or you will set goals the funnel cannot support.

Where do canvassing funnels leak the most appointments?

Five places: the first seven seconds at the door, where a weak opener kills the conversation; booking for 'later' instead of inspecting on the spot, which drops the show rate; inspections that find real condition but produce no documentation or decision; no tracking system, so the data is fiction; and knocking the wrong doors at scale. The last one is the most expensive and the quietest, because it shows up as low conversion on streets where most roofs were never real candidates.

How do I set fair canvassing quotas across different reps and territories?

Compare reps within the same context, or normalize to conversation rate to strip out territory noise. A rep booking 3 percent cold may be a stronger closer than one booking 10 percent in a storm core. Build the quota from your validated funnel ratios working backward from your target signed-file count, and flag any rep whose doors-per-file drifts far outside the band for their cold-versus-storm context.

How does roof-age and storm data improve canvassing conversion?

Targeting is upstream of every script and quota, so it lifts every downstream ratio at once. Knocking roofs that are aging out, or roofs a storm most likely wore down, means inspections find real condition far more often. RoofPredict scores a territory house by house with a roof-age range from aerial imagery plus storm physics modeled per roof, then ranks the doors and route. Roof age comes back as a range, not an exact date, and storm exposure is a probability per roof confirmed by the inspection, not a guarantee.

Why do scheduled 'come back later' inspections convert worse than on-the-spot ones?

Future appointments give the homeowner's yes time to cool off, so they no-show, cancel, or 'forget' at rates that can run 30 to 50 percent. On-the-spot inspections complete at 85 to 95 percent because there is no gap. Bias hard toward inspecting the moment the homeowner agrees, and when that is impossible, lock the booking with a confirmed time and a text confirmation sent before the rep leaves the porch.

What can a canvasser legally say about insurance and deductibles at the door?

A rep may inspect, document damage with photos, and prepare an accurate repair estimate for the homeowner to keep. They may state facts about their own scope. They may not promise an approval or payout, say the deductible is waived or absorbed or gone, advertise a free roof, interpret the homeowner's policy, or negotiate with the carrier on the homeowner's behalf. Those acts can constitute unlicensed public adjusting and deductible-rebating, which are illegal in many states. The homeowner files the claim and the insurer decides coverage.

What is the fastest way to start measuring my canvassing funnel?

You can start on paper. Capture per rep per field block: date, daypart, territory, doors approached, doors answered, conversations, inspections booked (flagging on-the-spot versus scheduled), inspections completed, inspections that found actionable condition, and files signed, with a cold-versus-storm flag. Within two weeks you can compute answer rate, conversation rate, knock-to-booked, show rate, find rate, sign rate, and doors-per-file, which replaces every generic benchmark with your own honest numbers.