Roofing Canvassing: How to Stop Wasting Time on Dead Doors

Walk any roofing crew through a full canvassing day and ask them honestly where the hours went, and you will hear the same answer in different words: most of it went to doors that were never going to turn into a job. The roof was three years old. Nobody was home, again. The homeowner just re-roofed with the guy down the street. The renter has no idea who owns the house. The dog. The gate. The six-minute conversation with the retiree who loves to talk but owns a roof that has another decade in it.

The knocks you remember are the ones that landed. The knocks that drained the day are the ones you forget, because nothing happened. That is the trap. Canvassing feels productive because you are moving, sweating, talking. But movement is not the same as progress, and a rep who knocks 80 doors and writes one estimate did not have a good day just because his feet hurt.

The phrase "dead door" gets used loosely. Let's make it precise, because the whole problem hides inside the fuzziness. A dead door is not a no. A no can flip. A dead door is a house where the roof does not need you and will not need you on any timeline that matters to your business, and no amount of charm changes that. You can be the best closer in the county and you cannot sell a replacement to a roof that is eight years old and undamaged. The homeowner would have to be lying to themselves or to you to buy. So the door was dead before you got out of the truck. The only question that matters is whether you knew it was dead before you spent the eleven minutes finding out.

This is a targeting problem dressed up as a sales problem. Reps and owners try to fix it with better scripts, more reps, longer days, pizza-party contests. Those help at the margins. But you cannot script your way out of knocking the wrong houses. If 60 percent of the doors on your route were never buyable, the ceiling on your day was set before anyone said a word. We are going to take that ceiling apart: what a dead door actually is, how to read it from the curb and from data, how to rebuild a route so the buyable roofs are dense instead of scattered, how to measure the right things so you can see the waste, and where roof-age and storm signals change the math entirely. Real workflows, real numbers, the edge cases pros trip on, and an honest accounting of what targeting can and cannot do for you.

The real cost of a dead door (run the math once and you can't unsee it)

Most owners have never put a dollar figure on a wasted knock, so the waste stays invisible. Let's fix that with arithmetic you can do on a napkin.

Take a canvasser earning the equivalent of $22 an hour fully loaded — base, payroll taxes, phone, the truck they ride in, gas, the tablet, the share of the manager who runs them. A productive door-knocker who is actually walking and talking gets through somewhere between 8 and 14 contacted doors an hour in a residential neighborhood, depending on density and how chatty the street is. Call it 10. So each door — each real interaction — costs you roughly $2.20 in pure labor before you count the no-answers, the drive time between streets, and the windshield time getting to the neighborhood.

That $2.20 sounds trivial. It isn't, because of the denominator. Here is a worked example. Say a rep knocks 300 doors in a week to produce the following funnel, which is generous for cold residential canvassing:

Three jobs out of 300 doors. That is a 1 percent door-to-job rate, which is actually a respectable week for true cold knocking. Now look at where the doors went. Of those 300, how many were houses that could never have bought — roofs too new, rentals where the owner is unreachable, homes that re-roofed last season? On most suburban streets that number runs 40 to 65 percent. Let's say 55 percent, so 165 dead doors. At $2.20 each that is $363 of labor poured into houses that had a zero percent chance, plus the drive time woven through them, plus the opportunity cost — every minute on a dead door is a minute not spent on a buyable one.

The insight is not "door knocking is bad." Door knocking is one of the highest-ROI customer-acquisition methods in roofing when the doors are good. The insight is that your effective cost per signed job is dominated by the dead doors, not the live ones. If you could cut dead doors from 55 percent to 25 percent of the route, you would not improve your day by 30 percent. You would roughly double the buyable doors a rep touches per hour, and the funnel scales off the buyable count, not the total count. That is the entire game.

Why "just knock more doors" is the expensive answer

When production sags, the reflex is volume: knock more, work later, add a rep. Volume works, but it is the most expensive lever you have because it scales the dead doors right along with the live ones. If your route is 55 percent dead and you double the doors, you doubled your dead-door spend too. You bought more of the same mix.

The cheap lever is mix. Improve the percentage of buyable doors and every downstream number — conversations, inspections, signed jobs — moves without anyone working harder. A rep who knocks 200 doors at 30 percent dead beats a rep who knocks 300 doors at 55 percent dead, on less mileage and less burnout. Burnout matters more than owners admit: canvasser turnover is brutal, and the fastest way to lose a green rep in week three is to march them down streets where every other house is a polite brush-off. People quit dead doors. They stick around when knocking feels like it works.

There is a compounding effect here that owners miss. A rep who hits a good street stays sharp — the energy on door fifty matches the energy on door five, because the doors keep rewarding the effort. A rep grinding a dead route gets flat by mid-afternoon: the pitch goes robotic, the smile goes, the body language reads "please say no so I can move on," and even the genuinely buyable doors on that route get a worse pitch because the rep has been beaten down by the dead ones leading up to them. So dead doors don't just waste their own slot; they degrade the quality of every live door that follows. Mix isn't only a math problem. It's a morale problem, and morale converts.

What actually makes a door dead

Dead doors fall into a handful of categories. Learn to name them and you start seeing them from the truck.

1. The roof is too young

This is the big one and the most fixable. An asphalt-shingle roof installed within the last 8 to 10 years is, barring storm damage, not a replacement candidate. You can spot a young roof from the curb if you know the tells — crisp, uniform granule color, sharp shingle edges, no cupping or curling, clean and bright flashing, ridge cap that still looks factory. But you can't always see the roof from the street, especially on a steep front-facing pitch or behind mature trees, and you certainly can't see it from the route list before you drive over.

The trap inside this category is that year-built is not roof-age. A house built in 1994 tells you almost nothing about the roof, because that roof has very likely been replaced once or twice since. The county record, the Zillow listing, the tax assessor — they carry year-built, not last-reroof date. Re-roofs are mostly invisible to those records unless a permit was pulled and digitized, and a huge share of residential re-roofs skip the permit or never make it into a searchable database. So a rep who tries to target "older homes" off public records is half-blind: they will knock old houses with new roofs and skip newer houses whose original builder-grade roof is failing right on schedule at 18 years. Age of the structure and age of the roof drift apart the moment the first roof comes off.

2. Unreachable owner / wrong decision-maker

Rentals, recently-sold homes, estates, homes where the person at the door is a tenant, an adult child, a house-sitter. The roof might be perfect for you and the conversation still goes nowhere because the person who can say yes is not standing there. Renter-occupied homes are a large slice of many neighborhoods — in some areas a third or more of single-family stock — and you usually can't tell from the curb. A messy yard or multiple cars is a weak tell at best.

3. Recently serviced

The homeowner re-roofed last year, or had a competitor out last month, or is mid-claim with another contractor. The roof may genuinely be old and worn, but the window is closed. These doors feel especially bad because the conversation often starts warm before it dead-ends.

4. No damage where damage is the whole pitch

If you knock storm-restoration, a house that did not actually take hail or wind impact is a dead door for your motion, even if the roof is fifteen years old, because your offer is built around documenting storm damage for the homeowner's own insurance process. A hail map that shows the storm passed over the ZIP does not mean every roof in the ZIP took damaging impact. Hail falls in streaks and swaths; one side of a street can get hammered while the other side a block over sees nothing. Knocking the whole storm ZIP as if every roof is a candidate is one of the biggest sources of wasted storm-canvassing time there is.

5. The chronic non-buyer

The house that has been knocked by every roofer in town and never bites — sometimes because of a known difficult owner, sometimes because the roof simply keeps not failing. Mature canvassing operations keep a "do not re-knock until" date so they stop re-burning these.

Notice that only one of these — category 1, young roof — is about the roof itself, and it is the one you can attack hardest with data before you ever walk up. The others you mostly read on-site, though good list hygiene (sold dates, rental flags, prior-contact history in your CRM) chips away at 2, 3, and 5.

Reading a dead door from the curb: the 15-second triage

Before data, there is craft. A sharp canvasser triages every house in the seconds between the truck and the door. Train your reps to run a quick read so they invest their energy proportionally — full pitch on the live ones, polite and fast on the obvious dead ones.

Here is a curb-read checklist your reps can internalize:

• Granule color and uniformity. Faded, blotchy, uneven color = aged or sun-beaten. Bright, even, almost shiny = young, skip the age angle.
• Shingle edges and lay. Cupping, curling, lifting tabs, a wavy or "tired" surface = candidate. Flat, crisp, tight = young.
• Granule loss. Bald spots, shiny asphalt showing through, granules in the gutters or at downspout splashes = aging. Clean = young.
• Flashing and penetrations. Rusted, lifting, or tar-patched flashing around chimneys and vents = older roof, possible leaks. Clean metal = newer.
• Ridge and hip caps. Worn, cracked, or missing caps = old. Factory-sharp = young.
• Sag or deck irregularity. Any waviness in the deck line viewed across the roof plane = potential structural or long-term moisture issue = worth a conversation.
• Storm tells (if you knock storm). Spatter marks on metal, dinged gutters and downspouts, dented soft-metal (AC fins, mailbox, gutter caps), missing or bruised shingles, exposed mat = real impact worth documenting. A clean roof under a storm ZIP is likely a streak-miss.
• Occupancy and owner tells. Personalized landscaping, a flag, holiday decor, a well-kept yard lean owner-occupied. Generic, neglected, multiple unrelated vehicles, a lockbox lean rental or transitional.

This triage is a skill, and it has a ceiling. From the street you often can't see the back slopes, the side facing away, or anything on a low-slope or heavily-treed lot. You're reading the front pitch and inferring the rest. A green rep gets maybe half of these right; a five-year vet reads a street like a book. The vet's edge is exactly the edge you're trying to give every rep without waiting five years — which is where pre-route data earns its keep.

Rebuild the route, not the rep: density beats hustle

The single highest-leverage change most roofing canvassing operations can make is not a better pitch — it is a better route. Two things drive a route's value: the percentage of buyable doors and the density of those buyable doors (how close together they are). Both matter, and they trade off.

The density problem

Suppose you identify the best 30 percent of doors in a 1,000-home area. If those 300 doors are evenly scattered, your rep walks past two dead houses for every good one and burns half the day on sidewalk. If those 300 doors cluster into a few streets and pockets, your rep walks a tight loop hitting good house after good house. Same doors, completely different day. The reason cluster matters is that walking and driving between doors is dead time too — it just doesn't feel like it because you're not getting rejected, you're just moving.

So route planning has two jobs: pick the buyable doors, then sequence them so the rep spends minutes-per-buyable-door, not minutes-per-door. A practical workflow:

1. Score the area, don't just pick a ZIP. Pull a roof-age and (where relevant) storm read on the whole target area so each address carries a "due / not due" signal rather than only an address.
2. Rank and threshold. Sort by likelihood the roof is due. Set a cut line — say, work everything above the 60th percentile of roof-age likelihood, plus any address with real storm-impact signal regardless of age.
3. Cluster the survivors. Group the above-the-line addresses geographically. Identify the dense pockets where good doors are tightly packed and start there. A street that is 70 percent above-the-line is gold; a street that is 15 percent above-the-line is a drive-by, not a walk.
4. Sequence for a tight walking loop. Order the doors so the rep walks a continuous loop with minimal backtracking, both sides of the street where density allows, and minimal truck repositioning.
5. Mark the skips explicitly. Give the rep the dead doors too, flagged as skip-or-soft-touch, so they don't waste a full pitch but can still drop a door-hanger or note a re-knock date.

That last point matters more than it sounds. A skip is not the same as ignorance. A rep who knows why a house is a skip can leave a branded door-hanger in ten seconds and move on, planting a seed for the day that roof does age into your window, without burning a conversation today.

Timing: the hours that quietly kill your contact rate

Route targeting is wasted if you walk it at the wrong time, because an unanswered door is its own kind of dead door — you paid for the walk and got no conversation. In commuter neighborhoods, weekday mid-afternoon is a graveyard: the people who can say yes are at work. Contact rates climb sharply in the late-afternoon-into-early-evening window on weekdays (roughly the two hours before dusk, when people are home but it isn't yet dark or dinner-locked) and across weekend daytime. A practical rule: stack your highest-value scored streets into those high-contact windows and save the lower-priority cleanup, door-hanger drops, and re-knocks for the dead hours. Knocking your best route at 1 p.m. on a Tuesday is throwing good targeting after bad timing.

Season and weather matter too. The day after a regional storm, contact and receptiveness spike because the event is on everyone's mind — that's the moment a per-roof storm read earns its keep, because you can walk the actually-hit streets while the homeowner is still looking up at their roof. Brutal heat, rain, and the dinner-to-bedtime window kill both contact and patience. None of this replaces good targeting; it multiplies it. A great list walked at the wrong hour still produces a mediocre day.

A worked route comparison

The scored-and-clustered route plans fewer total doors, finishes in less time, and puts the rep in front of nearly twice as many buyable roofs per hour as knocking the whole ZIP. Same rep, same pitch, same close rate per buyable conversation — the output roughly doubles because the input mix and density changed. This is the difference between working harder and working the right street.

Where RoofPredict fits: knowing which roofs are due before anyone climbs a ladder

Everything above hinges on one hard input: knowing, before you drive over, which roofs on a street are actually old enough or storm-worn enough to be worth a knock. Curb-reading gets you part way but only at the door, only on what you can see, and only as good as the rep's eye. Public records give you year-built, which is the wrong number. That gap is the specific thing RoofPredict is built to close.

Here is what it does, plainly. RoofPredict takes aerial and satellite imagery of your target area and returns, address by address, a roof-age estimate as a range (for example, 18 to 22 years — not an exact install date, because no one can read an exact date off a photo and we won't pretend otherwise), plus a storm read modeled per roof: beyond whether a storm passed over the ZIP, a per-roof model of the hail and wind that roof actually took, so you can separate the houses a storm genuinely wore out from the ones a block over that it missed. The difference between "it hailed in this ZIP" and "this specific roof took damaging impact" is the difference between knocking a streak and knocking a whole grid hoping. You can also feed in your own mailing list or CRM and have each address enriched with those signals, so the list you already own gets sharper instead of starting from scratch.

What that gives a canvassing operation is a route where the dead doors are flagged before the rep leaves the truck. The young roofs drop off the walk list. The old roofs and the genuinely storm-hit roofs rise to the top and cluster into the dense pockets worth working. A green rep handed that list knocks like a vet on day one, because the list already did the curb-read that takes years to learn — which house, before they climb anything.

Now the honest limits, because a tight trade compares notes and overselling is how you lose roofers. Roof age comes back as a range, not a date — it narrows the field, it does not certify a specific roof's exact age. The storm model is odds and modeled impact, not proof — it tells you which roofs most likely took damage and deserve a look, not which roofs are guaranteed to have a claimable hit; the roof itself, inspected in person, is the proof. It does not measure the roof (that's a measurement tool's job, a different category) and it does not identify the shingle product or read condition from the inside. It will occasionally flag a young roof as older or miss a recent re-roof that imagery hasn't caught up to. What it does reliably is move your route from "every house" to "the houses most likely to need you," which is exactly the lever that kills dead doors. It sharpens the outbound you already do; it is not a lead service and it does not hand you a homeowner who raised their hand. You still knock. You just knock the right doors.

The storm-canvassing trap (and how to stay on the right side of the line)

Storm restoration deserves its own section because it is where dead doors multiply fastest and where the legal and ethical lines are sharpest.

The canvassing trap is simple: a storm gets reported, a hail map lights up a ZIP, and crews knock the whole ZIP. But hail and damaging wind do not fall evenly. Hail comes down in streaks and swaths driven by the storm's track and the local wind field; one street takes golf-ball impact and the parallel street two blocks over takes pea-sized hail that does nothing to a shingle. Wind damage concentrates on exposed elevations and edges. So "the storm hit this ZIP" produces a route that is mostly dead doors — roofs that saw the storm pass but took no damaging impact. The rep wears out the day knocking houses with nothing to document.

The fix is the same lever as everywhere else: target the roofs that actually took impact, not the geography the storm crossed. Per-roof storm modeling (paired with a real roof-age read, since an old roof that took marginal impact is a different conversation than a two-year roof that took the same impact) collapses the storm ZIP down to the streaks that are worth your crew's hours. Then your rep walks up already knowing this specific roof likely took a hit — and confirms it with their own eyes and camera on the ladder.

The compliance line you must teach every rep

Storm canvassing is where roofers get into legal trouble, and a rep with a bad script can put your license and your business at risk. Draw the line clearly and teach it as part of canvassing training, not as an afterthought.

What a roofer may do: inspect the roof, thoroughly document damage with photos and measurements, prepare an accurate repair estimate aligned to standard estimating practice (Xactimate-style line items for the work), and hand that documentation to the homeowner. You can state facts about your own scope of work. That documentation-and-estimate role is entirely legitimate and is genuinely valuable to a homeowner.

What a roofer may not do — the do-not-say list every canvasser should have memorized:

• Do not offer to "handle," "manage," "negotiate," or "adjust" the insurance claim for the homeowner. For a fee, that is unlicensed public adjusting in most states and it is a serious violation.
• Do not interpret the homeowner's policy or tell them what is or isn't covered. That is the carrier's and the homeowner's territory.
• Do not promise a specific payout, an approval, or that the claim "will go through."
• Do not promise the deductible will be waived, absorbed, covered, eaten, or made to disappear. Offering to cover or rebate a deductible is illegal in many states and is insurance fraud.
• Do not advertise or imply a "free roof." The homeowner pays their deductible; that is not free, and the language invites fraud scrutiny.
• Do not represent the homeowner against their insurer or position yourself as their advocate in the claim.

The safe frame, said at the door: "If you'd like, I'll get up there, document anything the storm did, and write you a detailed estimate to repair it. That's yours to keep. If you decide to file, you file it and your insurance company decides what's covered — I just make sure the damage is documented right and the repair is scoped right." That captures every bit of the homeowner's actual interest while keeping you cleanly on the documentation-and-estimate side of the line. RoofPredict's role in this is the same as everywhere: it tells you which roofs likely qualify by age and modeled storm impact and helps you prioritize the documentation workload — it never touches the claim, never interprets coverage, and never promises an outcome.

Measure the right things or you're flying blind

You cannot manage dead doors if you only track doors and jobs. The middle of the funnel is where the waste hides, and most roofing operations don't instrument it. Here is a tight metric set worth tracking per rep, per route, and per area.

The diagnostic move is to read these together. If a rep's buyable rate is high but their conversation rate is low, the route is good and the rep needs coaching. If the buyable rate is low and you're knocking the whole ZIP, fix the route before you touch the rep. If contact rate is low, your reps are knocking the wrong hours (mid-afternoon weekday is a graveyard in commuter neighborhoods; aim for early evening and weekend daytime). Separating route quality from rep quality is the whole point — without the buyable rate you can't tell whether your reps are bad or your streets are.

A simple weekly review your managers can actually run

1. Pull each rep's funnel for the week (the table above).
2. Flag any rep whose buyable rate is below your target threshold — that's a route conversation, not a rep conversation.
3. Flag any rep whose buyable rate is fine but whose conversation or booking rate lags — that's coaching.
4. Look at which scored areas produced the best doors-per-job and feed that back into next week's route selection.
5. Update the do-not-re-knock and re-knock-on dates so chronic dead doors and not-yet-due roofs stop eating future routes.

That loop — score, work, measure, refeed — is what turns canvassing from a grind into a system that gets sharper every week.

Edge cases and what pros get wrong

A few situations break the simple rules. Knowing them separates a thoughtful operation from a brute-force one.

The new roof that's still a candidate. A roof can be young and still need you if it was installed badly — wrong nailing, no ice-and-water where code requires it, improper valley work, missing drip edge. These show as premature failure: granule loss, lifting, leaks on a roof that should have a decade left. Curb-reading and a real inspection catch these; pure age data will (correctly) rank them low, so don't treat the age score as gospel — it's a filter, not a verdict. The rep's eye still matters.

The old roof that will never buy. Some homeowners will run a 30-year-old roof until it leaks into the living room and then call the cheapest guy on the internet. Age says due; reality says chronic non-buyer. This is why you track outcomes and set re-knock dates rather than re-knocking the same tired roof every quarter.

Re-roof invisible to imagery. A roof re-done last month may not show up in the latest available aerial imagery yet, so it can score as old when it's brand new. The curb-read catches it instantly. Treat the data as the plan and the rep's eyes as the final check — they cover each other's blind spots.

HOA and uniform-development streets. In some developments the whole subdivision was built in the same year and re-roofed in the same wave, so they age together. When one roof on the street is due, the neighbors often are too — these streets are unusually dense in buyable doors when they hit their window, and unusually dead between windows. Score the street, not the house, and you'll catch the wave.

Steep-slope and tree-occluded fronts. When you can't see the roof from the street or clearly from above, lean harder on the data and on a polite "mind if I take a quick look from your driveway?" rather than guessing. Don't write off a house just because the front pitch hides the story.

Mixing motions on one street. A street can hold storm candidates and age candidates and dead doors all at once. Don't force one pitch down the whole street. A rep working a scored list should switch frames house to house: storm documentation here, aging-roof conversation there, soft door-hanger on the young one. Flexibility per door beats one script per street.

The over-knocked neighborhood. After a notable storm, every roofer in a hundred-mile radius descends on the same hot ZIPs, and out-of-town crews swarm in. By the time you knock, the homeowner has heard the same pitch six times this week and the door is dead from fatigue, not from the roof. Two defenses: get to genuinely-hit streets fast with a per-roof read instead of waiting for the hail map to make the rounds, and lean on your steady age-based work and your own customer book, which doesn't depend on a storm and isn't swarmed by transient crews. The roofs aging out of warranty on your home streets are buyable every month of the year, storm or no storm, and almost nobody is knocking them — that's the least-contested territory you have.

Confusing a slow door with a dead door. Not every non-yes is dead. A homeowner who is interested but distracted, mid-dinner, or not the only decision-maker is a deferred door, not a dead one. The skill is to disposition it correctly in your CRM with a real callback or re-knock date instead of mentally filing it as a loss. Dead means the roof can't buy; slow means the moment was wrong. Mixing those two up either burns you re-knocking dead roofs or loses you live ones you never circled back to.

Your own list is the most under-worked street you have

Before you knock a single cold door, look at the doors you already own. Every roofing company is sitting on a CRM full of old estimates that never closed and past customers whose roofs are now years older than when you last touched them. Those records carry something a cold street doesn't: prior contact, a name, sometimes a measurement, a reason they once cared. An estimate you wrote four years ago on a roof that was "a few years out" is, today, a roof that may be right in your window — and you already have the homeowner's information and a relationship, however thin.

The workflow: export your aged estimates and past-customer addresses, enrich them with a current roof-age and storm read so you can see which ones have aged into the buyable window or taken a storm since you last spoke, and prioritize a phone-and-door pass on those before you spend a dime on cold canvassing. It's the warmest "cold" outreach you have, the cheapest doors on your list, and most operations never work it. Cold canvassing is essential, but starting there while your own book sits idle is leaving the easy money on the table.

A 30-day plan to cut your dead-door rate

If you want to turn all of this into action, here's a concrete month.

Week 1 — Measure the baseline. Instrument the funnel from the metrics table for every rep. Specifically capture buyable rate — have reps tag each contacted door as candidate or dead and why (young roof / unreachable / recently serviced / no damage / chronic). You can't fix a number you've never measured, and most owners are shocked when they see their real dead-door percentage.

Week 2 — Fix the route, not the rep. Stop knocking whole ZIPs. Score your next target area by roof age (and storm impact if you run storm), threshold to the top tier, cluster the survivors into dense walking pockets, and hand reps a list that flags the skips. Compare the week's buyable rate to your Week 1 baseline.

Week 3 — Work your own book. Pull aged estimates and past customers, enrich them with current roof-age and storm signals, and run a warm phone-and-door pass on the ones that have aged into window. Track close rate against cold doors — it should be markedly higher.

Week 4 — Tighten the loop. Review the funnel by rep and by area. Separate route problems (low buyable rate) from rep problems (low conversation or booking rate) and coach accordingly. Set re-knock and do-not-knock dates so chronic dead doors stop reappearing. Feed your best-performing scored areas into next month's plan.

Do this for a month and the change shows up in the only number that matters: cost per signed job. Not because anyone worked harder, but because you stopped paying your crew to knock roofs that were never going to need you.

The bottom line

A dead door is a roof that doesn't need you and won't on any timeline that pays your bills. The hours your crew loses to those doors are the biggest, most invisible cost in canvassing, and you cannot script, hustle, or volume your way out of it — because all three scale the dead doors right along with the live ones. The lever that actually works is mix: knowing which roofs are due before anyone leaves the truck, clustering those roofs into tight routes, working your own book first, and measuring the buyable rate so you can see the waste and kill it.

Curb-reading craft gets your reps part of the way. A real roof-age read and a per-roof storm model — the kind RoofPredict produces from aerial imagery, as a range and as odds, with honest limits — gets the rest of the way by doing the hard targeting before the walk starts, so the dead doors are flagged and the due roofs rise to the top. You still knock every door yourself. You just stop knocking the ones that were never going to open.

Notice what this does to the kind of business you run. A crew that knocks the right doors closes more per hour, so the green reps make money early and stay instead of quitting in week three. A route built on roof age works every month, not only the weeks after a storm, so your pipeline stops swinging between feast and famine. And the streets you mine are yours — your scored area, your own customer book — not a lead resold to five competitors and not a storm you have to wait on. That's the real prize hiding inside a boring-sounding problem like dead doors: fix your targeting and you don't just save labor, you change what your company depends on to grow.

If you want to see what your own streets look like scored by roof age and storm impact — which houses are actually due, and which ones your crew has been burning hours on — that's exactly the picture RoofPredict is built to hand you. Book a demo and bring a street you already know well; you decide whether we read it right.

FAQ

What exactly counts as a "dead door" in roofing canvassing?

A dead door is a house where the roof does not need you and won't on any timeline that matters to your business, so no pitch can convert it. That's different from a "no," which can flip. The common types are roofs too young to replace, unreachable owners (rentals, recent sales, tenants at the door), recently serviced roofs, storm-ZIP houses that took no actual damaging impact, and chronic non-buyers. Only the young-roof category can be filtered hard with data before you knock; the rest you catch through list hygiene and on-site triage.

Why can't I just target old houses using Zillow or county records?

Because those records carry year-built, not roof age, and the two drift apart the moment the first roof comes off. A 1994 house has likely been re-roofed once or twice, and most re-roofs never make it into a searchable record because permits get skipped or never digitized. Targeting old structures means you'll knock old houses with new roofs and skip newer houses whose original builder-grade roof is failing at 18 years. You need a read on the roof itself, not the building's age.

How much money do dead doors actually cost?

Run the math. A fully-loaded canvasser around $22/hour who contacts about 10 doors an hour costs roughly $2.20 per contacted door. On a typical suburban route, 40 to 65 percent of doors were never buyable. At 55 percent dead across 300 doors, that's about $363 of labor a week poured into zero-chance houses, plus drive time woven through them and the opportunity cost of not being at a buyable door. Your effective cost per signed job is dominated by the dead doors, not the live ones.

Isn't the answer just to knock more doors?

Volume is the most expensive lever you have, because it scales your dead doors right along with your live ones. If a route is 55 percent dead and you double the doors, you doubled the dead-door spend too. The cheap lever is mix: improve the percentage of buyable doors and every downstream number moves without anyone working harder. A rep knocking 200 well-targeted doors beats a rep knocking 300 random ones, on less mileage and less burnout.

How do I read a dead door from the curb?

Train a 15-second triage: granule color and uniformity (faded/blotchy = aged, bright/even = young), shingle lay (cupping or curling = candidate, flat and crisp = young), granule loss in gutters, rusted or patched flashing, worn ridge caps, deck sag, and storm tells like spatter on metal and dinged gutters. Owner tells (personalized yard vs. lockbox/neglect) hint at occupancy. The limit is that you only see the front pitch from the street, and a green rep gets maybe half of these right, which is why pre-route data helps level the field.

What does RoofPredict actually give a canvassing crew, and what are its limits?

It reads aerial imagery for your target area and returns, address by address, a roof-age estimate as a range (like 18 to 22 years, not an exact install date) plus a storm read modeled per roof, so you can separate roofs a storm actually wore out from ones a block over it missed. You can also enrich your own list or CRM with those signals. The honest limits: age is a range not a date, the storm model is odds and modeled impact not proof, it doesn't measure the roof or identify the shingle, and it can occasionally miss a fresh re-roof imagery hasn't caught. It sharpens the outbound you already do; it's not a lead service and you still knock every door.

Why is knocking the whole storm ZIP a waste of time?

Hail and damaging wind don't fall evenly. Hail comes in streaks and swaths, so one street takes golf-ball impact while a parallel street two blocks over takes harmless pea-sized hail. A hail map showing the storm crossed the ZIP doesn't mean every roof took damage. Knocking the whole ZIP produces a route that's mostly dead doors. Targeting the roofs that actually took modeled impact collapses the ZIP down to the streaks worth working, then your rep confirms it in person on the ladder.

What can a roofer legally say about insurance at the door?

You may inspect, document damage with photos, write an accurate repair estimate, and hand that documentation to the homeowner; you can state facts about your own scope. You may not handle, negotiate, or adjust the claim for them (that's unlicensed public adjusting), interpret their policy or coverage, promise a payout or approval, promise the deductible is waived or absorbed, advertise a "free roof," or represent them against their insurer. The safe frame: "I'll document the damage and write you an estimate that's yours to keep; if you file, you file it and your insurer decides what's covered."

Which canvassing metrics should I track to find wasted time?

Don't track only doors and jobs — the waste hides in the middle. Track doors knocked, contact rate, buyable rate (real candidates per contact, which is your dead-door meter), conversation rate, inspection booking rate, inspection-to-estimate rate, estimate-to-signed rate, doors per signed job, and cost per signed job. Read them together: high buyable rate with low conversation rate is a rep coaching issue; low buyable rate is a route problem. Without the buyable rate you can't tell whether your reps are weak or your streets are.

Should I work my own CRM before cold canvassing?

Yes, almost always. Your aged estimates and past customers are the warmest "cold" doors you have — prior contact, a name, sometimes a measurement, and a roof that's now years older than when you last quoted it. Enrich those addresses with a current roof-age and storm read to see which have aged into your window or taken a storm since, then run a phone-and-door pass on them before spending on cold canvassing. Most operations never work this list, which is leaving easy money idle.