How to Pick Which Streets to Canvass for Roofing (A Field-Tested System)

Picking streets is the part of canvassing nobody trains. New reps get a script, a door-hanger, and a vague wave toward "that subdivision over there," then everyone wonders why the numbers are soft. The truth most sales managers learn the hard way: the street you choose decides 60 to 70 percent of your result before the first knock. A great pitch on a wrong street loses to a mediocre pitch on a right one. So if you want to fix canvassing economics, you fix street selection first.

The goal here is to give you a repeatable system for choosing where to knock, in the order a working contractor actually makes the decision: market type, roof age, storm exposure, neighborhood fit, competitive saturation, then route math. None of it requires expensive software to start. All of it gets sharper when you layer in data. By the end you should be able to look at a metro map and rank neighborhoods from "send the whole team" to "don't waste a Saturday," and explain why to a 22-year-old rep so they buy in.

First, decide what kind of canvasser you are

Before you pick a single street, get honest about which game you're playing, because the two games pick streets in opposite directions.

Storm/insurance restoration. You're chasing recent, documentable weather damage — hail, wind, fallen limbs — and helping homeowners document conditions so their insurer can decide coverage. Your street selection is driven by where the storm actually hit, hard, within a workable claim window. Freshness matters more than anything. A neighborhood that took golfball hail eleven days ago beats a wealthier one that's been quiet for two years.

Retail/age-out replacement. You're selling roofs that are simply worn out — end-of-life shingles, granule loss, brittle decking, the slow stuff. No storm required. Here your street selection is driven by housing age, original-builder shingle quality, and household ability to pay cash or finance. You want the 18-to-28-year-old asphalt roofs in neighborhoods where people fix things before they fail.

Most healthy companies run both, but a given canvassing day should commit to one. Mixing them confuses reps and wrecks your messaging. A storm route wants urgency and "we're already working three roofs on this block." A retail route wants trust, curb appeal, and "your neighbors are getting ahead of the next bad summer." Decide the game, then pick streets for that game.

The economic floor every street has to clear

Whatever the game, every candidate street has to clear a simple math floor or it's not worth the labor. Work it backward from a rep's day:

• A canvasser working a decent residential block knocks roughly 80–120 doors in a 4-hour shift.
• Contact rate (someone answers) on a weekday evening runs 25–40 percent; weekend midday can hit 40–50 percent.
• Of contacts, a strong rep books 8–15 percent into an inspection on a qualified street, far less on a bad one.

So a 4-hour shift on a good street might be 100 doors × 35% answer × 12% set = ~4 inspections. On a bad street, the same effort yields 100 × 25% × 4% = 1 inspection. Same labor, four times the output. That spread is the entire reason street selection deserves a system, not a hunch. Your job is to keep your crews on streets that produce the 4, not the 1.

Carry the math one step further to make the cost real. Say a fully loaded canvasser costs you about $25/hour all-in, so a 4-hour shift is $100 of labor before any commission. On the good street, four inspections at, say, a 25 percent inspection-to-job close and a $12,000 average job is one signed roof — roughly $100 of canvassing labor to originate a five-figure job. On the bad street, one inspection at the same close rate is a quarter of a job for the same $100 — your cost to originate just quadrupled, and that's before you count the morale tax of a rep who knocked 100 doors and set one appointment. Reps quit over bad streets far more than over hard scripts. Protecting the math protects your roster.

The practical takeaway: you don't need a fancy model to justify a selection system. You need to refuse to send crews onto streets that can't clear the floor, and a willingness to kill a street fast when the test numbers come back soft.

The seven signals that make a street worth knocking

Underneath all the tactics, you're scoring each street on seven signals. The rest of the playbook is just how to read each one cheaply and combine them.

1. Roof age — how close the typical roof is to end-of-life.
2. Storm exposure — whether a recent hail/wind event physically loaded this roof.
3. Roof type / pitch — whether it's the work you actually want and can price.
4. Homeownership & ability to pay — owner-occupied, household income, financeability.
5. Density & walkability — doors per hour, parking, dog/no-soliciting friction.
6. Competitive saturation — how many of your rivals' yard signs are already up.
7. Permit/recency — who already re-roofed, so you stop knocking dead doors.

A street that's strong on five-plus of these is a green street. Strong on three is a maybe — worth a test block. Strong on one or two is a no. Let's take each signal and turn it into something a rep can read from the curb or a manager can pull from data.

Signal 1: Roof age — the single best predictor for retail

For age-out work, roof age beats everything. Asphalt three-tab and architectural shingles — the overwhelming majority of U.S. residential roofs — carry useful service lives that cluster in a predictable band. Three-tab tends to give 15–20 years in a hot climate, architectural 20–30, depending on ventilation, color, slope, and install quality. The sweet spot for retail canvassing is the back half of that band: roofs that look tired but haven't failed, where the owner already half-knows it's coming.

The problem is you usually can't get a precise install date per house without pulling records one at a time. So you triangulate.

How to estimate roof age without knocking

Subdivision build year. Tract neighborhoods were built in waves, and most homes got their original roof from the same builder in the same year. Find the build-out year and you've got the original roof age for the whole block. County assessor records and parcel viewers list "year built" for free in nearly every county. If a 1999 subdivision never re-roofed, those original roofs are pushing 25-plus years — a green street for retail, all day.

The re-roof wave. Here's the part rookies miss: subdivisions re-roof in waves too. The first roofs go around year 18–22, a cluster follows as neighbors copy neighbors, and within a few years a big chunk of the street is freshly done. So a 1999 build that already had a hail event in 2014 may be half new roofs now — looks old on paper, knocks dead in reality. You have to layer build year against any known re-roof trigger (a documented storm, a wave of permits) to find the streets that are old and still original.

Curb tells, for reps on foot. Train reps to read a roof in three seconds from the sidewalk: dark streaks and bald patches (granule loss), curling or cupping shingle edges, a roof line that looks wavy or sagged, moss on the north slope, and mismatched patches. A roof that reads "tired but intact" on a block of similar-age homes is the exact door you want.

Street View time-travel. Before sending a crew, walk the street in Google Street View and check the capture date. If imagery from a few years back shows the same weathered roof, that roof has been aging in plain sight. If a house shows a bright new roof in recent imagery, flag it as a skip.

Worked example: scoring two retail streets

Maple Ridge is the street. It's old enough to be due, has no storm or permit wave that would've already replaced the roofs, and the curb confirms it. Oakwood is a beautiful neighborhood you should not be knocking for retail roofs — you'd burn a shift to find the one outlier.

Reading the climate adjustment

The age bands above are national averages, and they move with climate. A roof in Phoenix or Las Vegas bakes under UV and thermal cycling that ages asphalt faster — the same architectural shingle that gives 25 years in a mild Midwest climate may give 18 in the desert. Coastal salt air, freeze-thaw in northern winters, and high humidity with algae growth in the Southeast all pull the useful-life number down. So adjust your retail sweet-spot band to your market: in a harsh-sun metro, start prospecting subdivisions at 14–16 years, not 18. In a mild climate, you can push the band later.

Ventilation and color matter too, and they're things a rep can read. Dark roofs on poorly ventilated attics cook from below and age faster; you'll see them fail earlier than the light-colored, well-vented house next door of identical build year. When two same-age streets diverge in how tired their roofs look, attic ventilation and shingle color are usually the reason — and the more tired street is the better knock.

The "original spec" tell

One more retail nuance worth training into reps: builder-grade three-tab versus upgraded architectural shingle. Production builders in the late 1990s and early 2000s often installed cheap three-tab shingles to hit a price, and those roofs age out faster and uglier than later architectural installs. A subdivision built in 2003 with builder three-tab can be more due than a 2000 build that went architectural. From the curb, three-tab reads as a flat, uniform, shadow-line pattern; architectural reads as dimensional and chunky. A street of aging three-tab is a green-plus street — old spec, old roof, and homeowners who already dislike how it looks.

Signal 2: Storm exposure — the engine of restoration canvassing

For storm work, exposure outranks age. A two-year-old roof that took severe hail is a legitimate, documentable claim; a worn-out roof that hasn't seen weather in five years is a retail conversation, not a restoration one. The mistake crews make is canvassing a whole ZIP code after a storm when the damaging core only crossed a few square miles.

Hail and wind damage is shockingly localized. A single supercell can drop 1.75-inch hail on one subdivision and pea-sized hail two miles away. Wind damage tracks even tighter — microbursts and downburst corridors can flatten one street's ridge caps and leave the next street untouched. If you canvass off the actual damage footprint, your inspection-to-approval rate craters because half the roofs have nothing to document.

How to map the real damage footprint

Start with the public storm record. The NOAA Storm Prediction Center publishes daily storm reports including hail size and wind gusts with locations, and the National Weather Service issues local storm reports (LSRs) after events. The SPC storm report archive lets you pull the hail and wind points for the exact date. This tells you where and how big, for free, from a source no one can argue with.

Layer radar-derived hail estimates. NWS radar products and Maximum Estimated Size of Hail (MESH) data turn the radar sweep into a hail-size grid. Commercial hail-verification reports (the kind adjusters reference) build on this. The grid shows you the swath — the line the core actually tracked — so you can draw your canvass boundary on the swath, not the whole county.

Respect hail-size thresholds. As a rough field rule, asphalt shingles usually need around 1-inch (quarter-size) hail or larger before you reliably find functional damage — bruising, mat fracture, granule displacement. Below that you'll find cosmetic marks and a lot of "there's nothing here" inspections. Wind damage on asphalt tends to show up in the 50-plus mph gust range, more so where shingles were already aging or poorly sealed. Use the public data to canvass inside the threshold, not outside it.

Read the wind direction. Hail and wind come in at an angle, so damage concentrates on the storm-facing slopes — typically the south and west faces for a classic plains supercell tracking northeast. When you scout, the give-away is collateral: dented gutters, dinged downspouts, spatter marks on AC fins, cracked window screens, and pockmarked fence tops and decks. If the soft metals on a street are clean, the shingles probably are too. If the gutters are peppered, go knock.

Storm timing windows

Storm canvassing is a race, but not only against competitors. Homeowners generally have a window to file after a covered event — commonly stated in policies as a "prompt" or one-year notice requirement, though it varies by carrier and state, so never quote a hard deadline to a homeowner as fact. What this means operationally: the highest-value canvass days are the first two to three weeks after the storm, while damage is fresh, undisputed, and the homeowner is motivated. Map the footprint within 48 hours, get crews on the hardest-hit streets first, and work outward as the dense core gets saturated.

The 48-hour storm-response checklist

Speed is the whole game on storm work, but speed without targeting just means you're first to the wrong streets. Run this sequence in the two days after a significant event:

1. Confirm the event was real and severe. Pull SPC storm reports and the NCEI Storm Events Database for the date. Look for hail at or above 1 inch and wind gusts at or above 50 mph in your service area. If the public record shows only pea-sized hail, the storm probably isn't worth a full deployment, no matter how loud it sounded.
2. Draw the swath. Overlay the radar-derived hail grid on a street map and outline the band where hail size cleared your damage threshold. This is your hard canvass boundary.
3. Rank subdivisions inside the swath by build year, roof spec, and storm-facing slope. Older, three-tab, south-and-west-facing streets go to the top.
4. Build the door order. If you have per-roof storm modeling, sort doors by modeled exposure. If not, sort by storm-facing slope and visible age.
5. Scout for collateral on the top two subdivisions — dented soft metals confirm the radar before you commit a team.
6. Deploy the next morning on the confirmed core, with a dead-door skip list and clean documentation discipline.

Wind-only events are different

Not every storm is hail. Straight-line wind, downbursts, and the occasional weak tornado damage roofs without a hailstone in sight, and they canvass differently. Wind damage is about lift and seal failure: missing or creased shingles, lifted ridge caps, exposed nail heads, and debris impact. It concentrates on edges, ridges, and the windward slope, and it's spottier than hail — one street loses shingles, the next is fine. For wind events, lean harder on the visible-damage drive-by than on a radar grid, because gust data is point-sampled and patchy. If you can see torn shingles and displaced ridge caps from the street, that's your signal; if the roofs read intact, the gust that triggered the warning may have missed this block entirely.

Signal 3: Roof type and pitch — knock the work you want

Not all roofs are jobs you want. A street full of complex slate, tile, or steep cut-up architecture may be impressive, but if your crews are an asphalt shop, you'll write estimates you can't competitively fill or safely produce. Conversely, a street of simple 4:12 to 6:12 walkable asphalt is fast to inspect, fast to price, and fast to build.

From the curb, train reps and route-planners to favor:

• Walkable pitch (roughly 4:12–7:12) — faster, safer inspections and lower production cost than steep-slope.
• Asphalt shingle if that's your core trade — highest match between what you scope and what you install.
• Simple geometry — fewer valleys, dormers, and penetrations means tighter, more confident estimates and fewer change orders.

This signal is also a quiet safety filter. Steep and complex roofs raise fall risk, and fall protection is the most-cited safety issue in residential roofing. Choosing walkable streets is partly a margin decision and partly keeping your people off the roofs most likely to hurt them.

Signal 4: Homeownership and ability to pay

A roof can be perfectly due and still be a bad door if the person who answers can't authorize or fund the work. Two filters matter.

Owner-occupied vs. rental. Renters can't approve a roof, and absentee landlords are slow, price-driven, and often out of area. Owner-occupancy rates vary enormously block to block. You can infer it: homestead exemption flags in assessor data, mailing-address-equals-property-address in parcel records, and curb signals (personalized landscaping, kids' toys, well-kept entries read owner; bare, transient, or multiple-mailbox setups read rental). For retail especially, prioritize high owner-occupancy streets.

Ability to pay / financeability. You're not looking for the richest neighborhoods — you're looking for the right ones. Ultra-high-end homes often have property managers, complex roofs, and architects in the loop. The canvassing sweet spot for retail is solid middle and upper-middle owner-occupied housing: people with home equity, a stake in curb appeal, and either cash or clean credit for financing. The U.S. Census Bureau's American Community Survey publishes median household income, owner-occupancy, and home value by census tract and block group, free, and it's the cleanest public way to grade a neighborhood's payment profile before you ever knock.

For storm work this signal softens — the insurer's decision, not the homeowner's bank account, drives most of the funding — but it never disappears, because deductibles, upgrades, and code items still come out of pocket.

Reading a census tract before you knock

The American Community Survey publishes, free, at the census-tract and block-group level: median household income, median home value, owner-occupancy rate, year-structure-built distributions, and even median year moved in. That last one is quietly useful — a tract where owners moved in long ago and stayed tends to have stable, invested homeowners who maintain the property. Pull three numbers for any candidate tract: owner-occupancy above ~70 percent, median income in your serviceable range, and a structure-age distribution that skews toward your target build years. A tract that hits all three is a strong retail bet before a single rep sets foot on it.

Avoid the trap of treating income as the only filter. A modest-income tract of long-tenured owner-occupants with 22-year-old roofs and strong financing uptake can out-convert a high-income tract of young move-ins with new roofs. You're selling to people whose roof is due and who can act — not to a tax bracket.

Signal 5: Density, walkability, and friction

This is the boring signal that quietly decides your cost per door. Two streets can have identical roofs and incomes, but one yields twice the knocks per hour.

Favor:

• Tight lot spacing. Suburban tract neighborhoods with 50–80-foot lot widths let a rep cover doors fast on foot. Rural acreage and long driveways destroy doors-per-hour.
• Grid or short-loop street layouts. Easy to work systematically without backtracking. Sprawling cul-de-sac mazes waste walking.
• Sidewalks and on-street parking. Reps move faster and your route vehicle stays close.

Avoid or de-prioritize:

• HOAs with no-soliciting enforcement and gates. Gated communities are often a legal and practical dead end for door-knocking. Some municipalities also require a solicitation permit — check the city clerk before you deploy, because a citation in front of a prospect kills the whole street.
• Apartment and townhome density (for retail) — shared roofs, HOA decisions, no single homeowner to sell.

Signal 6: Competitive saturation

Drive a candidate street before you commit a crew and count yard signs and fresh roofs. Three readings:

• No new roofs, no competitor signs: virgin territory — great, assuming the other signals hold.
• A few fresh roofs, a competitor sign or two: a storm or age wave is breaking here. This is often the best moment — social proof is building ("the neighbors are doing it") and the street isn't tapped yet. Lean in with neighbor references.
• Many fresh roofs, signs everywhere: late. The wave already crested; you're knocking dead doors and fighting on price. Skip or send one rep to skim, don't commit the team.

Knowing where a street sits on that curve is the difference between riding a wave and cleaning up after it.

Signal 7: Permit data — stop knocking dead doors

The most underused free signal. When a roof is replaced, a permit is usually pulled. Many cities and counties publish issued building permits, often searchable or downloadable, sometimes via open-data portals. If you can pull re-roof permits for a subdivision, you get a near-exact map of which houses are already done — the dead doors — and, by inversion, which are still original.

Use it two ways:

1. Exclusion list. Mark every recently permitted address as a skip so reps don't waste knocks on brand-new roofs.
2. Timing read. A spike in re-roof permits in a subdivision two or three years ago tells you a storm or age wave already broke there — confirming Signal 6's "late" read. A subdivision that's old by build year but shows almost no re-roof permits is the jackpot: due, and still original.

Permit coverage and searchability vary wildly by jurisdiction — some are beautiful open portals, some are a clerk and a filing cabinet — so treat it as a bonus layer where available, not a universal step.

How to actually pull permit data

The access path differs by city, but the usual options, easiest first:

• Open-data portal. Larger metros publish issued permits on a city or county open-data site, often filterable by permit type and date and downloadable as a spreadsheet. Filter for roofing/re-roof permits, export, and you have an address-level skip list in minutes.
• Permit-search web app. Many mid-size cities run a permit-lookup tool on the building department's site where you can search by address or subdivision. Slower, but workable for a focused target list.
• Public-records request. Where nothing's online, a short records request to the building department often gets you a permit export. Worth doing once for your top recurring territories.

Whatever the source, normalize what you pull into two lists: a skip list of recently re-roofed addresses, and a permit-density read per subdivision (count of re-roof permits in the last few years). High recent density means a wave already broke — late. Near-zero density on an old subdivision means original roofs — jackpot. You're using the same dataset to both avoid dead doors and time the wave.

Where roof-by-roof data changes the math

Everything above is doable by hand, and you should know how to do it by hand. But the manual version has a ceiling: it's slow, it's coarse (you're scoring whole subdivisions, not individual roofs), and it leans on triggers you can see rather than the actual condition of each roof. Reading build year off the assessor tells you the original roof age for a tract, but it can't tell you which specific houses already re-roofed quietly without a visible sign, or which roofs a storm physically loaded harder than their neighbors.

This is the gap roof-intelligence tools are built to close, and where a platform like RoofPredict fits the workflow. Instead of scoring a subdivision, it scores roofs address by address from aerial imagery and storm modeling, and it does two specific things that are hard to do manually at scale:

• Roof-age range per address. From aerial and historical imagery it estimates an age range for each roof — not a precise install date, and that distinction matters. A range like "18–24 years" is honest about the uncertainty in visual estimation, and it's exactly enough to rank a street: it surfaces the genuinely old-and-likely-original roofs and quietly drops the ones that already look freshly done. You stop knocking dead doors without pulling a single permit by hand.
• Storm physics modeled per roof. Rather than a county-wide hail blob, it models the storm's effect on each individual roof — size, exposure, slope facing — to rank which specific homes the event most likely wore out. That turns a fuzzy "this ZIP got hit" into a ranked door list inside the swath, so your first, most valuable post-storm shifts land on the hardest-hit roofs first.

The honest limits, because they matter to how you use it: an age range is an estimate from imagery, not a roof-coring report — it ranks and prioritizes, it doesn't certify a roof's exact age, and your rep still has to verify condition on site. A storm model outputs odds and exposure, not proof of damage — it tells you which roofs to inspect first, and the physical inspection is what actually documents the condition. Used that way — to rank streets and order doors, then confirm on the roof — it compresses the hours of assessor-plus-radar-plus-Street-View work into a prioritized list and frees your reps to spend the day knocking instead of researching. It does not knock for you, and it does not replace the inspection. It just makes sure the knocks land where the roofs are actually due.

The framing to keep clean, especially on storm work: the roofer documents conditions and writes the estimate, the insurer decides coverage, and the homeowner owns the claim. Data that ranks which roofs to look at first is a routing tool, not a coverage promise — never let a rep present a model output as a guarantee that a roof is damaged or that a claim will pay.

Putting it together: a street scorecard

Here's a simple weighted scorecard you can hand a sales manager. Score each candidate street 0–5 on each signal, multiply by the weight, sum. It forces consistency and lets you compare neighborhoods objectively instead of by gut.

Multiply, sum, and rank. A retail street scoring above ~75 percent of its possible weighted points is a send-the-team street. 50–75 percent is a test-block street — send one rep for two hours and read the result before committing. Below 50 percent, don't.

The weights are a starting point; tune them to your market. In a hail alley, storm exposure dominates. In a stable metro with old housing stock and no recent storms, roof age and originality carry the day.

A filled-in scorecard example

To make it concrete, here's a retail street run through the scorecard. "Birchwood Estates," a 2002 build, no storm history, middle-income owner-occupied, tight suburban lots, a handful of fresh roofs starting to appear:

That's 87 percent of the maximum possible weighted score — a send-the-team street. The low storm-exposure score barely dents it because the retail weights make storm nearly irrelevant. Run the same street through storm weights and it would score poorly, which is exactly right: it's a great retail street and a bad storm street. The scorecard keeps that distinction honest instead of letting a rep's enthusiasm blur it.

When to override the scorecard

The number is a guide, not a tyrant. Override it in two situations. First, a referral cluster: if you just signed two jobs on a block, knock the whole block regardless of score, because active social proof temporarily spikes conversion far above what the static signals predict. Second, a production constraint: if your crews are an asphalt shop and a high-scoring street turns out to be tile or steep-slope, the roof-type score should veto the total — a street you can't profitably build isn't a street worth selling. Treat those as deliberate, logged exceptions, not as license to ignore the system whenever a street "feels" good.

The week-by-week canvassing workflow

Knowing how to score a street is half of it. The other half is a repeatable weekly motion so street selection actually happens before crews roll, instead of in the truck at 4 p.m.

Monday: build the target list

1. Pull a metro map and your category: retail or storm for the week.
2. For storm: pull the latest SPC/NWS storm reports and radar hail swaths, draw the footprint, list the subdivisions inside it.
3. For retail: pull assessor build-years, flag subdivisions 18–28 years old, cross-check re-roof permits where available, and drop the ones with a re-roof wave.
4. Run the candidate streets through the scorecard. Rank them.
5. If you have roof-by-roof data, sort the doors inside the top streets so the oldest-range / hardest-hit roofs are knocked first.

Tuesday–Thursday: scout and test

1. Drive the top 3–5 streets. Confirm curb signals, count competitor signs and fresh roofs, check for gates/no-soliciting, confirm any solicitation-permit requirement with the city.
2. Run a two-hour test block on the borderline streets with one strong rep. Log doors knocked, contacts, and inspections set. Kill streets that underperform the math floor.

Friday–Saturday: deploy the team to proven streets

1. Commit the crew to the streets that cleared scoring and the test. Saturday midday is your highest contact-rate window — spend it on your best street, never your worst.
2. Assign tight territory so reps don't overlap or skip. Hand each rep the dead-door exclusion list (permitted/new roofs).

Continuously: log everything and let data compound

1. Track per street: doors knocked, contact rate, inspections set, inspections-to-jobs. Over a season this builds your own private map of which neighborhoods convert — the most valuable asset a canvassing operation owns.

The metrics that tell you a street is working

Log at the street level, not only the rep level, because you're trying to grade territory as well as people. Track at minimum:

When a street's set rate runs healthy but the close rate tanks, the problem isn't street selection — it's your inspection or estimate process. When the set rate itself is low on a street that scored well, re-examine the scoring inputs; usually a re-roof wave already broke or competitors saturated it. Separating those two failure modes is the entire value of street-level logging. Without it, you blame reps for territory problems and territory for rep problems, and fix neither.

Mapping tools and route discipline

You don't need enterprise software to run this, but you do need every door tracked on a shared map so reps don't overlap or skip. Canvassing apps let a manager draw territory, assign streets, and have reps tag each door's outcome (not home, not interested, inspection set, new roof/skip) in real time. The payoff compounds: after a season, the map is your institutional knowledge. A new rep inherits it instead of relearning the metro from scratch, and a roof-data layer that flags due roofs slots straight into the same map as a prioritization overlay.

A worked storm scenario, start to finish

It's a Thursday in late spring. Overnight, a line of storms crossed the north side of your metro. Here's the selection sequence a sharp storm lead runs.

Hour 1 — footprint. Pull SPC storm reports for last night: hail points showing 1.75-inch stones reported near two suburbs, wind gusts to 70 mph along a corridor. Pull the radar-derived hail swath — a 3-mile-wide band tracking northeast across four subdivisions. That band, not the county, is the canvass zone.

Hour 2 — rank inside the band. Inside the swath, the four subdivisions differ. One is a 2019 build (young roofs, but 1.75" hail still loads them — keep it). Two are early-2000s builds (older, more vulnerable — prioritize). One is a gated HOA (deprioritize, soliciting blocked). If you've got roof-by-roof storm modeling, you now have a ranked door list across all four ordered by modeled exposure; if not, you rank by build-year and storm-facing slope.

Hour 3 — scout for collateral. Drive the top subdivision. Dented gutters, spatter on the AC units, cracked screens on the south and west faces — real exposure, confirmed. The give-aways line up with the radar. Green light.

Day 1–2 — deploy on the core. Crews knock the two early-2000s subdivisions first, working the storm-facing streets, leading with documentation help and neighbor activity. Reps log every door and every soft-metal observation — that collateral evidence is what supports an honest, well-documented inspection.

Week 2–3 — work outward. As the core saturates, expand to the swath edges and the younger subdivision, then to the gusts-only corridor where wind (not hail) is the story. By the time competitors fan out across the whole county, you've already worked the streets the storm actually hit hardest.

That's the entire advantage: not knocking more, knocking where the storm was, in the order of exposure.

A worked retail scenario, start to finish

No storm in months, but you need to keep crews busy and the board full. Here's the retail selection sequence.

Step 1 — find the age band. Pull the assessor build-year layer for your metro and filter to subdivisions built 1998–2006 — old enough that original roofs are due, adjusted earlier if you're in a harsh-sun climate. You get a dozen candidate subdivisions.

Step 2 — drop the already-waved. Cross-check re-roof permit density. Three of the twelve show heavy recent re-roof permits — a wave already broke there (likely an old hail event). Drop them; they'll knock dead. Two show almost zero re-roof permits despite being 22 years old — flag those as priority.

Step 3 — grade the demographics. Pull ACS for the remaining tracts. Keep the ones above ~70 percent owner-occupancy and in your serviceable income range. One candidate is mostly rental — drop it. Two are long-tenured middle-income owner-occupied — keep, high priority.

Step 4 — confirm spec and pitch on the drive-by. Two of the survivors are aging builder three-tab on walkable pitch — ideal. One is upgraded architectural that still looks decent — lower priority. One is steep cut-up architecture you can't competitively build — drop on roof-type veto.

Step 5 — score and rank. Run the survivors through the scorecard. The 22-year-old, low-permit, three-tab, owner-occupied, walkable street tops out near 90 percent. That's your Saturday street.

Step 6 — test the borderline, deploy on the winner. Send one rep two hours midweek on the architectural borderline street; if it clears the math floor, add it to rotation. Put the whole team on the top street Saturday midday with a permit skip list in hand.

Same discipline as storm work, different lead signal: there, the storm footprint orders the doors; here, age-and-originality does.

What pros get wrong about street selection

After enough seasons you see the same expensive mistakes repeat. Watch for these.

Canvassing the whole ZIP after a storm. The single most common error. The damaging core is usually a few square miles; the ZIP is fifty. Canvassing outside the swath buries your reps in "nothing here" inspections and tanks morale. Map the footprint first, every time.

Chasing the richest neighborhood. High-end streets feel like the prize and usually aren't: complex roofs, property managers, architects, slow decisions, price-shopping. The middle and upper-middle owner-occupied streets convert faster and build faster.

Ignoring the re-roof wave. An old subdivision by build year can be half-replaced already if a storm or age wave broke there years ago. Build year alone is a trap. Always cross-check a re-roof trigger — permits, fresh roofs on the drive-by, or roof-age data — before you call a street "old."

Treating every street the same. Reps sense it when management sends them somewhere on a hunch. A scorecard and a test-block discipline tell reps you've done the homework, which is half of why they knock with conviction.

Burning a Saturday on a weak street. Your highest contact-rate hours are scarce. Spending them on an unproven street is the most expensive mistake in the system, because you can't get the time back. Prove streets midweek; deploy the team on Saturday only to winners.

Knocking dead doors. No exclusion list means reps waste knocks on brand-new roofs and freshly-done houses, which is demoralizing and slow. Build the skip list from permits and recent imagery before deployment.

No logging, no learning. Crews that don't record doors/contacts/sets/jobs per street rebuild their territory knowledge from scratch every season. The companies that log compound a private conversion map that no competitor can copy.

Confusing storm and retail on the same shift. Sending a crew out with a storm-urgency script onto an age-out retail street (or vice versa) produces mush. The homeowner senses the mismatch — there was no storm, but the rep is talking storm — and trust evaporates. Commit each shift to one game and brief the messaging to match.

Treating a model output as proof. On storm work especially, a rep who tells a homeowner "our data shows your roof is damaged" before anyone has inspected it is making a claim the company can't stand behind and a regulator won't like. Per-roof storm modeling outputs odds and exposure to decide which roofs to inspect first — it is not evidence of damage. The inspection documents condition; the model just routes the truck. Train that distinction hard, because crossing it invites trouble and erodes the homeowner trust that makes a street referable.

Over-scouting and never deploying. The opposite failure: managers who research endlessly and never commit a crew. Selection is meant to be fast — a Monday morning, not a week. Pull the data, score, test the borderlines midweek, deploy Saturday. Analysis that delays knocking is just procrastination with a spreadsheet.

Compliance and courtesy that keep streets open

A street is only worth knocking if you can keep knocking it. Two practical guardrails.

Solicitation permits and no-knock rules. Many municipalities require a door-to-door solicitation permit, and some maintain no-knock registries; the FTC and state consumer-protection rules also govern door-to-door sales, including cancellation-right disclosures on contracts signed at the home. Check the city clerk and your state attorney general's guidance before deploying in a new jurisdiction. A reputation for clean, permitted canvassing keeps neighborhoods — and referrals — open to you.

Honest, non-pressure messaging. Especially post-storm, the line between helpful and predatory is watched closely by regulators and homeowners alike. Keep reps documenting conditions and offering inspections, not promising claim outcomes, "free roofs," or covered deductibles. The roofer documents and estimates; the insurer decides coverage; the homeowner owns the claim. Streets stay open to companies that play it straight, and close fast to the ones that don't.

Bringing it home

Picking which streets to canvass isn't a vibe — it's a scoring problem with seven readable signals and a math floor every street has to clear. For retail, lead with roof age and originality. For storm, lead with the real damage footprint and rank doors by exposure. Layer in ownership, density, saturation, and permits, then prove borderline streets with a midweek test block before you spend a Saturday on them.

You can run the whole system with free public data — assessor build-years, Census tract demographics, SPC/NWS storm reports, radar hail swaths, and city permit portals — and you should learn it that way so you understand what the data means. When you're ready to compress the research and rank roofs door by door instead of subdivision by subdivision, roof-age-range and per-roof storm modeling tools pick up exactly where the manual method hits its ceiling. Either way, the principle holds: keep your crews on the streets where the roofs are genuinely due, in the order the data says, and let the pitch do its job on doors that were worth knocking in the first place.

FAQ

How do I find out the age of roofs on a street before knocking?

You triangulate. County assessor records give the subdivision's year built, which usually equals the original roof age for the whole tract. Cross-check re-roof permits (where the city publishes them) and drive-by curb signals — granule loss, curling, streaking — to find which roofs are still original. Roof-age tools that estimate an age range per address from aerial imagery can do this door by door, though a range is an estimate that ranks roofs, not a certified install date you'd verify on site.

Should I canvass the whole ZIP code after a hail storm?

No — that's the most common and most expensive mistake. The damaging core of a storm is usually a few square miles, while a ZIP can be fifty. Pull NOAA Storm Prediction Center reports and radar-derived hail swaths to draw the actual footprint, then canvass inside the swath, ordered by hail size and storm-facing slope. Canvassing outside the footprint buries reps in inspections that find no damage.

What hail size actually damages an asphalt shingle roof?

As a field rule, asphalt shingles generally need around 1-inch (quarter-size) hail or larger before you reliably find functional damage like bruising, mat fracture, or granule displacement. Smaller hail tends to leave cosmetic marks and a lot of empty inspections. Use public hail-size data to canvass inside that threshold, and remember the inspection — not a storm model — is what documents actual damage.

Are wealthy neighborhoods the best to canvass for roofing?

Usually not. High-end streets bring complex roofs, property managers, architects, slow decisions, and price-shopping. The canvassing sweet spot for retail is solid middle and upper-middle owner-occupied housing — homeowners with equity, curb-appeal motivation, and cash or clean credit. Census tract income and owner-occupancy data help you grade a neighborhood's payment profile before you knock.

How many doors should a canvasser knock per shift, and what's a good conversion rate?

A rep working a tight residential block typically knocks 80–120 doors in a 4-hour shift. Contact rate runs 25–40 percent on weekday evenings and 40–50 percent weekend midday. Of contacts, a strong rep sets 8–15 percent of qualified doors into inspections. On a good street that's roughly 4 inspections per shift; on a bad street, closer to 1 — which is exactly why street selection matters more than knock volume.

How do I stop reps from knocking houses that already have new roofs?

Build a dead-door exclusion list before deploying. Pull recent re-roof permits from the city or county portal and mark those addresses as skips. Cross-check recent aerial or Street View imagery for visibly new roofs. Roof-age data that flags freshly-done roofs per address does the same job at scale. Handing reps a skip list saves wasted knocks and protects morale.

Do I need a permit to canvass door to door for roofing?

Often yes. Many municipalities require a door-to-door solicitation permit and some maintain no-knock registries, and the FTC plus state consumer-protection rules govern door-to-door sales, including cancellation-right disclosures. Check the city clerk and your state attorney general's guidance before deploying in a new jurisdiction. Clean, permitted canvassing also protects your reputation and keeps neighborhoods open to you.

What's the difference between picking streets for storm work versus retail?

Storm canvassing leads with the recent damage footprint — hail size, wind corridor, freshness — because exposure outranks age and there's a claim window. Retail leads with roof age and originality, owner-occupancy, and ability to pay, with no storm required. A given canvass day should commit to one game, because the messaging, urgency, and street choices pull in opposite directions.

How does roof-by-roof data improve over scoring whole subdivisions by hand?

Manual selection scores whole tracts off build year and visible storm triggers, which is coarse — it can't see which specific homes quietly re-roofed or which roofs a storm loaded hardest. Roof-by-roof data estimates an age range and models storm exposure per address, producing a ranked door list inside your chosen streets. It compresses hours of assessor-plus-radar research, but it ranks and prioritizes — it doesn't replace the on-site inspection that documents condition.

What's the best day and time to canvass for roofing?

Weekend midday — typically Saturday — gives the highest contact rate, often 40–50 percent, because people are home. Weekday early evenings run 25–40 percent. The strategic rule: spend your scarce high-contact hours only on streets that have already cleared your scorecard and a midweek test block. Never burn a Saturday on an unproven street, because you can't get those hours back.