How to Target Old Roofs Door to Door: A Field-Tested System for Knocking the Right Houses

Walk any residential street with a roofer who has been doing this for fifteen years and watch their eyes. They are not looking at the houses. They are looking at the roofs. Curling tabs on the south-facing slope. Granule wash streaking the gutters black. A patch of three-tab shingles on an architectural roof where somebody fixed a leak and gave up matching. That practiced read is the difference between a crew that knocks 120 doors to set two appointments and a crew that knocks 60 to set six.

The problem is that the read does not scale. You cannot clone the fifteen-year eye and hand it to a 22-year-old you hired last Tuesday. And even the veteran is guessing on the doors where the damage is subtle, where the roof is mid-life, or where the wear is on the slope you cannot see from the truck. Targeting old roofs door to door is really two problems stacked on top of each other: knowing which roofs are actually aging out before you walk up, and converting the conversation once you are standing on the porch. Most teams are weak at the first and pour all their training into the second. That is backwards. The smartest thing you can do to a door-knocking program is make sure your reps are spending their hours in front of the right houses.

What follows is the system I would hand a new sales manager who has a four-person canvass team, a territory, and a quota. It covers how to read roof age from the ground and from above, how to score a neighborhood before anyone laces up boots, how to build routes that respect both the data and the math of a sales day, what to actually say at the door, the legal and safety rails you cannot skip, and how to measure whether any of it is working. There is real first-hand operational detail here, the kind you only get from watching a lot of crews succeed and fail. Where modern aerial-imagery tooling changes the game, I will say so plainly, including where it does not.

Why "old roof" is the wrong target on its own

Let me start by breaking the premise a little, because the contractors who win at this think more precisely than "find old roofs."

Roof age matters because of what it predicts: a roof that is near or past the end of its service life is more likely to be insurable as a full replacement after a storm, more likely to be failing in ways a homeowner can feel (leaks, energy bills, daylight in the attic), and more likely to belong to someone who already knows in the back of their mind that this is coming. According to the National Roofing Contractors Association, most roof systems are designed for roughly 20 years of useful service, though actual life span swings widely with climate, ventilation, material quality, and installation. The Asphalt Roofing Manufacturers Association and NRCA guidance puts basic three-tab shingles in the 15-to-25-year band and architectural (laminated) shingles in the 25-to-35-year band, but field reality compresses those numbers hard in high-UV, high-hail, or poorly ventilated situations.

So "old" is a proxy. What you are really hunting for is due. A roof is due when its remaining service life is short enough that replacement is the rational call, and due-ness is a function of three things at once:

1. Age relative to material and climate. A 16-year-old three-tab roof in West Texas sun is functionally older than a 16-year-old architectural roof in coastal Oregon shade.
2. Storm exposure. A roof that took 1.75-inch hail eighteen months ago has aged faster than its years. Impact bruises the mat, knocks granules loose, and accelerates the UV breakdown the granules were protecting against.
3. Condition and installation history. Patches, mismatched shingles, prior layovers (shingle-over-shingle), poor ventilation, and bad flashing all pull the due date forward.

When you target purely on visible age, you knock a lot of doors where the roof is old but the homeowner has no trigger, no budget, and no intention of moving. When you target on due-ness, you knock doors where age, exposure, and condition line up — and those conversations are night-and-day easier. The whole system below is built to find due, not merely old.

Reading roof age and condition from the ground

Before any data tool, your reps need to be able to read a roof from the sidewalk in five seconds. This is a teachable skill, and you should drill it. Here is the field checklist I teach, roughly in order of how much signal each item carries.

The high-signal tells

• Granule loss in the gutters and at downspout splash blocks. Asphalt shingles shed their protective granules as they age and after impact. Black, sandy sediment in the gutters or a dark fan at the bottom of the downspout is one of the most reliable ground-level signs that a roof is well into its second half of life.
• Curling and cupping tabs. When shingle edges lift or the centers cup, the mat has dried out and lost flexibility. This is a late-life sign and reads clearly even from across the street, especially in raking morning or evening light.
• Surface mismatch and patches. A rectangle of slightly different shingles, or three-tab patches on a laminated roof, tells you the roof has already needed intervention and the owner has lived with a halfway fix.
• Bald spots and shiny patches. Where granules are gone, the black asphalt mat shows through. Under sun it can look almost shiny. These are the spots that fail first.
• Sagging ridge or deck plane. A ridge line that dips or a roof plane that looks wavy can indicate deck or structural issues — a bigger conversation, and often an older home.

The medium-signal tells

• Moss and dark streaking. Streaking is usually algae (Gloeocapsa magma) and is cosmetic more than structural, but heavy moss holds moisture against the shingle and shortens life. Read it as "older, north-facing, damp" context, not as proof of failure.
• Exposed or rusted fasteners and flashing. Rusty pipe-boot collars, lifted step flashing, and exposed nail heads point to an aging install and likely entry points for water.
• Pipe boots cracked at the rubber collar. These fail around the 10-to-15-year mark regardless of the shingle and are a great low-threat opener ("those rubber boots around your vent pipes crack right about now and that is usually the first leak").

The context tells

• Neighborhood build-out year. Tract neighborhoods get roofed within a year or two of construction and then re-roofed in waves. If you know a subdivision went up in 2004 and used builder-grade shingles, the whole pocket is statistically due. More on neighborhood scoring below.
• Visible re-roofs nearby. When you see new roofs scattered through a street, the original cohort is hitting end of life and some neighbors have already pulled the trigger. The hold-outs are warm.

Train reps to call these out loud to each other on a ride-along until it is automatic. A simple drill: park at the end of a street, have each rep rank the worst five roofs they can see and explain why. Then walk it and check who was right. Two ride-alongs of this and a new rep's read improves dramatically.

Reading age by shingle type

One refinement that separates a sharp rep from an average one: read the shingle type before you read the wear, because the same amount of wear means very different things on different products.

• Three-tab (flat, single-layer, uniform tab pattern). These are the builder-grade shingle of the 1990s and early 2000s and the shortest-lived of the common asphalt products. On a three-tab roof, curling and granule loss show up earlier and the end-of-life window is tighter. A worn three-tab roof in a 2002 subdivision is almost certainly original and due.
• Architectural / laminated (dimensional, shadowed, staggered look). Thicker, two-layer, and longer-lived. Wear shows up later and the failure is often at the laminate bond or in granule loss on the storm-facing slope rather than wholesale curling. A 2010 architectural roof can still have years left; a 1998 one is past due.
• Designer / luxury laminates and impact-rated (Class 4) shingles. Rarer on tract homes, more common on higher-value or rebuilt-after-storm homes. Class 4 impact-rated product resists hail far better, so do not assume a hail event totaled it — and a homeowner who paid for Class 4 may already be sophisticated about roof condition.

Knowing the type also sharpens the pitch. "Those builder-grade three-tabs in this neighborhood start shedding right around now" lands as expertise; "your roof is old" lands as a guess.

The ceiling on ground reading, though, is hard. You only see one or two slopes from the street. The slope that took the hail, or the slope baking in afternoon sun, is often the one you cannot see. That is exactly where overhead data earns its keep.

Estimating roof age from above

The overhead view solves the biggest blind spot of ground reading: you can see all the slopes, the full granule-loss pattern, repairs, and the storm-facing aspect. The trick is knowing what aerial imagery can and cannot tell you.

What overhead imagery shows you

• Whole-roof granule loss and weathering pattern. From above you can see the mottled, lighter, granule-depleted areas across every plane, not only the street-facing one. A roof that looks fine from the curb can be cooked on the back slope.
• Repairs and mismatches across the whole roof. Patches you would never spot from the ground are obvious from above.
• Slope and aspect. Which planes face south and west (UV) and which face the prevailing storm track. This is how you separate two same-age roofs by real exposure.
• Tree cover and shading. Heavy canopy ages the shaded slope differently (moss, moisture) and protects others.
• Imagery history. Comparing imagery captured a few years apart can show when a roof was last replaced — if a roof looks new in 2015 imagery and weathered in 2024 imagery, you have bracketed its install window.

What overhead imagery cannot tell you

Be honest with your reps about the limits, because over-claiming gets them in trouble at the door. Imagery cannot tell you the exact install date, the shingle's warranty class, the attic ventilation, or whether a soft hail bruise exists that has not yet shed granules. Resolution, capture date, sun angle, and image age all introduce error. The output of any age-from-imagery process is an estimate expressed as a range, never a precise date. A roof might read as "likely 18 to 24 years old," not "installed March 2007." Treat anyone or any tool that claims a precise install date from a photo with suspicion.

Building your own age-from-imagery read

You can do a rough version of this manually with free and low-cost imagery before paying for anything:

1. Pull current high-resolution satellite or aerial imagery for the target area.
2. Pull a historical image from several years prior (many imagery providers retain dated historical layers).
3. For each roof, look for the new-roof signature in the older image (uniform color, sharp granule field, no streaking) versus weathering in the newer one. The transition brackets the install date.
4. Cross-reference county assessor data, which often lists "year built" and sometimes effective-year or permit history, to anchor the original roof cohort.
5. Combine with the visible weathering severity to land on a range.

This is slow and skilled work to do by hand across thousands of parcels. It is exactly the kind of thing that aerial-imagery analysis tools now automate at scale — which is the bridge to the next section.

Layering storm history onto roof age

Age alone undersells due-ness. The accelerant is storm exposure, specifically hail and high wind, and you should be layering storm data on top of your age read.

The Insurance Institute for Business & Home Safety has spent years testing how hail damages asphalt shingles, and their work is the cleanest public reference on the subject. IBHS impact testing uses laboratory-manufactured hailstones at 1.5-inch and 2.0-inch diameters and documents the damage modes that matter: granule loss, mat cracking, and tears, all of which accelerate the underlying aging the granules were there to slow. A roof that took a real hail event is, in service-life terms, older than its install age suggests. Roughly 80 percent of installed U.S. residential roofs are asphalt shingle, per IBHS, so this applies to almost every door you knock.

For the storm-history layer, the authoritative public sources are NOAA's Storm Prediction Center and the NOAA/NWS Storm Events Database. The SPC publishes daily and archived severe-weather reports including hail size and wind, and the Storm Events Database holds the longer record by county going back to 1950. Local NWS offices also issue Local Storm Reports during and after events. None of these were built for roofers, so they take work to translate into "which streets got hit how hard," but they are the ground truth that every commercial hail-map product is ultimately derived from.

Here is the practical way to use storm data to target:

• Pull the hail and wind history for your territory by date and footprint. You want to know which neighborhoods sat under which storm cells, and the estimated stone size and wind speed.
• Cross the storm footprint with roof-age cohorts. A neighborhood that is both old enough to be near end of life and took 1.5-inch-plus hail is your highest-priority pocket. Old-and-hit beats old-or-hit every time.
• Mind the clock. A homeowner's right to file a storm claim is governed by their policy and state law, and many policies set deadlines (often one to two years) from the date of loss. Document conditions and let the homeowner and their insurer handle the claim — your job is to find the worn roofs and document what you observe, not to make coverage promises.

A hard guardrail here: a storm forecast or a storm history is odds, not proof. The fact that a cell with hail signatures passed over a neighborhood raises the probability that roofs there have impact damage. It does not prove any specific roof was damaged. Train reps to talk in terms of "worth a free inspection because your area took hail in that April storm," never "your roof was definitely damaged." The honest, probabilistic framing is also the legally safer one.

Scoring a neighborhood before anyone knocks

The single biggest leverage point in a door-to-door program is choosing where to knock. A great rep in a bad neighborhood loses to an average rep in a due one. Build a repeatable neighborhood score so this decision is not left to whoever feels lucky.

Here is a scoring model you can run on a spreadsheet today. Score each candidate neighborhood 1 to 5 on each factor, weight, and sum.

Where to get each input:

• Roof-age cohort: county assessor "year built" as a baseline, refined with imagery weathering and any age-from-imagery tooling.
• Storm exposure: NOAA SPC reports and the NWS Storm Events Database, or a commercial hail-map layer derived from them.
• Homeownership rate, value band, density: the U.S. Census Bureau's American Community Survey publishes owner-occupancy, home value, and housing-age data down to the census-tract and block-group level for free.
• Permits: most county and city permit portals are searchable; re-roof permit clusters tell you a neighborhood's replacement wave is live.

A worked example. Suppose you are weighing two subdivisions:

• Maple Crest: built 2003-2005, architectural shingles, took 1.75-inch hail fourteen months ago, 82% owner-occupied, mid-market values, wide streets, few permits pulled yet. Roof-age cohort 5, storm 5, ownership 4, value 4, density 4, permit 4. Weighted score ≈ 4.6.
• River Walk: built 1998, mixed re-roofs already done, last hail four years ago and minor, 60% owner-occupied with rentals mixed in, narrow streets, lots of recent permits. Roof-age cohort 3 (many already replaced), storm 2, ownership 3, value 3, density 2, permit 2. Weighted score ≈ 2.6.

Maple Crest is where you send the team. River Walk's wave has largely passed and its storm signal is stale. Without scoring, a rep might pick River Walk because it "looked older." The model catches what the eye misses.

Reading the assessor and Census data correctly

A few practical notes on the public data, because people misread it and target badly as a result.

Year built is not year roofed — but it is your best free anchor. Assessor "year built" tells you when the cohort of original roofs went on, give or take a year. The catch is re-roofs: a 1995 home may have been re-roofed in 2012, which resets its roof clock to roughly 2012 even though the assessor still says 1995. That is exactly why you layer imagery on top — the imagery tells you whether the original roof is still up there. Some assessors also publish permit history or an "effective year built," which can flag major work. Use year built to find candidate cohorts, then let imagery and permits correct it.

Census ACS is published in multi-year estimates. The American Community Survey publishes 1-year and 5-year estimates; for small areas like census tracts and block groups you will usually be reading 5-year estimates, which are averages over a window and lag the present. That is fine for owner-occupancy and housing-age structure, which change slowly, but do not treat a five-year-old estimate as today's exact number. You are using it to rank neighborhoods, not to count individual doors.

Owner-occupancy is a quiet multiplier. A neighborhood at 85% owner-occupied converts very differently from one at 55%. Renters cannot authorize a roof and will send you to a landlord who is price-driven and slow. When two neighborhoods tie on age and storm, the higher owner-occupancy pocket wins, and it is worth more weight than people give it.

Where RoofPredict fits in the targeting stack

Everything above can be done by hand. The reason most teams do not do it well is that doing it by hand across a whole territory — pulling imagery, bracketing install windows, cross-referencing assessor data, overlaying storm footprints, scoring neighborhoods, and then sequencing routes — is days of skilled desk work that no one on a four-person crew has time for. That gap is the practical case for a targeting platform.

RoofPredict exists to compress that work. It analyzes aerial imagery to estimate a roof-age range per address — a range, not a date — and models storm physics per individual roof so you get a per-house read on both how old the roof likely is and how hard the weather has actually worked it. Instead of "this neighborhood took hail," you get a house-by-house ranking of which roofs are most likely due, so your reps walk up the porches where age and exposure line up and skip the ones that were re-roofed last year. It is built to rank doors and routes, not to sell you leads — you still knock, you still inspect, you still earn the job at the door.

What it does well is the part humans are slow and inconsistent at: reading every slope of every roof at scale, bracketing ages from imagery history, and overlaying modeled storm exposure so the list you hand a rep is sorted worst-first. What it honestly does not do — and you should not expect any tool to — is replace the inspection. The age figure is a range with real uncertainty. The storm model gives you odds that a roof is worn, not proof any specific roof is damaged. It will not tell you the attic ventilation, the exact shingle warranty class, or whether a soft bruise is hiding under intact granules. You confirm all of that on the roof. Think of it as a way to point your reps' hours at the right houses, not as a substitute for the read you still do in person. Used that way — targeting and sequencing, with the inspection still doing the proving — it turns the multi-day desk exercise above into a sorted list, and that is where the time savings and the conversion lift come from.

Turning a scored neighborhood into a route

A sorted list of due roofs is useless if a rep zig-zags across town to hit them. Route construction is its own discipline, and it is where the math of a sales day collides with the data.

The economics you are optimizing

A door-to-door rep's day has a fixed number of productive hours — call it five to six hours of actual knocking after travel, setup, and breaks. Two numbers determine the day's output:

• Doors per hour (DPH): how many doors the rep physically reaches. Walking density, gates, dogs, and travel between targets all hit this.
• Contact and conversion rates: of doors knocked, what share answer, and of those, what share convert to an inspection.

The mistake new managers make is optimizing only conversion (better pitch) while ignoring DPH. A rep who knocks 40 doors at 8% conversion sets 3.2 inspections. A rep who knocks 70 doors at 6% conversion sets 4.2. Density and routing often beat pitch polish. The best programs win both, but you cannot win conversion if your reps are in trucks instead of on porches.

Building the route

1. Cluster, do not scatter. Take your due-roof list and group targets into tight walkable clusters of 60-100 doors. A rep should be able to work a cluster on foot without returning to the truck.
2. Sequence for daylight and door-answer patterns. Contact rates rise late afternoon into early evening when people are home. Front-load the day with the lighter-density or drive-heavy clusters and save the dense, high-due cluster for the 4-to-7 p.m. window when answer rates peak. Respect local solicitation hours.
3. Include the surround, not only the bullseye. When you knock a due roof and set an inspection, knock the four-to-eight neighbors immediately. They share the build year and the storm footprint, the social proof is real ("I'm doing an inspection for your neighbor two doors down"), and the marginal travel cost is zero. This neighbor-knock discipline is one of the highest-ROI habits in canvassing.
4. Plan for callbacks and not-homes. Roughly 60-80% of doors will be not-home or no-decision on the first pass. Build a second-pass route for the same cluster on a different day and time so you catch the people who were out.
5. Geo-fence the rep. Assign each rep a defined cluster for the day so two reps do not double-knock and so you can hold them accountable to a finished area.

A sample day plan

That is roughly 85-105 doors in a day with the densest, best-targeted block placed where contact rates are highest.

Seasonal and territory-level planning

Routes are the day; territory planning is the quarter. A few patterns worth building around:

• Chase the storm season, but do not abandon the age book. Hail and wind seasons concentrate restoration demand into bursts, and when a real event hits your area you should swing the team onto the fresh footprint fast, while the event is top of mind for homeowners and within policy deadlines. But the age-driven "due roof" work is the steady baseline that keeps the crew producing in the quiet months. The best programs run both: storm-reactive when there is a storm, age-proactive the rest of the year.
• Sequence neighborhoods by ripeness, not only score. A neighborhood whose replacement wave is just starting (a few new roofs, most still original) is riper than one where half the roofs are already done. Permit clusters and visible re-roof counts tell you where in the wave a neighborhood sits. Hit the early-wave pockets first; the late-wave ones have fewer doors left to win.
• Protect your best neighborhoods from over-knocking. If you flood a high-due neighborhood with every rep for three days straight, you exhaust it and you irritate it. Rotate. Work a pocket, set what you can, move on, and bring a second pass back later for the not-homes. Treat your top neighborhoods as a renewable resource, not a strip mine.
• Map competitor activity. Yard signs, branded dumpsters, and door-hangers tell you who else is working a pocket. A neighborhood three competitors are already saturating may be worth less to you than a quieter one with the same age and storm profile. Sometimes the second-best-scoring neighborhood that nobody else has found is the best place to send your team.

What to actually say at the door

Targeting puts the rep on the right porch. The conversation has to do the rest. The biggest error reps make is leading with the roof's problem and the sale, which trips every "salesperson at my door" alarm a homeowner has. The better structure earns a few seconds, gives a real reason, and lowers the threat.

The opener

Keep it short, specific, and low-pressure. A structure that works:

1. Name the area reason. "Hi, I'm [name] with [company] — we're doing roof inspections over here in Maple Crest because the neighborhood took hail back in that April storm."
2. Give a specific observation if you have one. "From the street it looks like your roof is right around the age where these builder-grade shingles start shedding granules."
3. Make a small ask, not a big one. "I'm not here to sell you anything today — I'd just take a quick look and let you know honestly whether you've got a real issue or whether you're fine for another few years."

Notice what that does. It explains why you are on this street specifically (storm + area), it demonstrates competence (the granule observation), and it asks only for a free look with an honest-either-way promise. "You might be fine" is disarming and it is true — and saying it builds the trust that converts the homes that are not fine.

Handling the common objections

• "I already had someone look at it." "Totally fair — a lot of folks here have. Did they give you anything in writing on the condition, or just a quote? I'll leave you a documented read either way so you've got a second opinion in your file."
• "My roof is fine." "It might well be, and I'll tell you straight if it is. The thing is the back slope and the storm-facing side are where the wear shows up first, and you can't see those from the ground. Two minutes and you'll know."
• "How much is this going to cost me?" "The look is free. If you do have damage, your roof's age and what your policy covers drive everything from there — that's a conversation for after we actually know what's up there. I'm not going to quote you a number before I've seen it."
• "Are you going to do my insurance claim?" "No — I document the conditions and give you an estimate of the work. You own the claim and your insurer decides coverage. I'll make sure you've got clear documentation to work with." This is the honest answer and it keeps you clear of the line you must not cross.

The legal line at the door

This matters and reps get companies in trouble by ignoring it. You document conditions and provide estimates. The homeowner owns their claim and the insurer decides coverage. Do not promise a claim will be approved, do not promise a "free roof," do not promise to cover, waive, rebate, or absorb a deductible, and do not present a storm forecast or storm history as proof that a specific roof was damaged. Many states explicitly regulate public-adjusting activity and deductible handling by contractors; crossing into "we'll get your claim approved and your deductible covered" can expose the company to real liability. Keep the rep's job to its lane: find the worn roof, document what is there, write an honest estimate.

The compliance and safety rails you cannot skip

A door-to-door program that ignores the rules is a lawsuit and an injury waiting to happen. Build these in from day one.

Door-to-door sales law

The FTC's Cooling-Off Rule gives consumers the right to cancel certain sales made at their home (and a few other locations) within three business days for sales of $25 or more at the home. If you sign contracts at the door, you must give the buyer notice of that cancellation right and the cancellation forms, and honor a timely cancellation. Many states layer additional home-solicitation requirements on top. Beyond the federal rule, most municipalities require solicitation permits and set allowed hours (commonly something like 9 a.m. to dusk, with local variation). Check and comply per city; carrying the permit and respecting hours also makes reps less likely to draw nuisance complaints.

Respect "No Soliciting" signs and Do-Not-Knock lists. Skipping a posted house costs you nothing and a complaint costs you a lot. Train reps to log and avoid these.

Honest-claims conduct

Reiterate the rule from the door section as policy: no claim-approval promises, no deductible handling, no "free roof," no forecast-as-proof. Put it in writing in your rep handbook and your scripts so it is not left to improvisation.

Roof and ladder safety

When the inspection happens, OSHA's fall-protection standard (29 CFR 1926.501) requires protection from falls at six feet or more above a lower level in residential construction — guardrails, safety nets, or personal fall-arrest systems, or a compliant alternative plan where conventional protection is infeasible. Ladder setup, footwear, weather (never a wet or frosted roof), and a no-solo-climb policy belong in your safety program. A canvasser is not necessarily climbing, but your inspectors are, and a single fall ends careers and companies.

Measuring whether your targeting actually works

If you cannot measure it, you cannot improve it, and most canvass programs measure almost nothing useful. Track the funnel at the level of the targeting decision, not only the rep.

The core funnel metrics

The two numbers that reveal targeting quality specifically are doors per inspection and revenue per 100 doors. Pitch quality moves inspection-set and sold rates; targeting quality moves how few doors it takes to get there and how much revenue each hundred doors generates. If you improve targeting and hold the pitch constant, doors-per-inspection should fall and revenue-per-100-doors should rise. That is the experiment to run.

Running the targeting A/B

To prove your targeting changes work, split it cleanly:

1. Take two comparable reps (or the same rep on alternating days).
2. One works neighborhoods chosen by your old method (gut, or "looks old"); the other works neighborhoods chosen by the scoring model and a due-roof list.
3. Hold the pitch and the hours constant.
4. Compare doors-per-inspection and revenue-per-100-doors over two to three weeks.

If scored targeting wins — and in practice it usually does, sometimes dramatically — you have your business case to systematize it. If it does not, your scoring inputs are wrong and you tune them. Either way you are deciding with data instead of folklore.

A worked funnel example

Numbers make the leverage obvious. Say a rep knocks 100 doors in a day in a poorly targeted neighborhood: 30% contact (30 conversations), 10% of contacts set an inspection (3 set), 70% of those complete (2 completed), 35% of completed sell (about 0.7 sold). At a modest job value, that is roughly seven sold jobs per 1,000 doors.

Now target the same rep into a high-due, storm-hit neighborhood and hold the pitch constant. Contact rate is similar at 30%, but inspection-set rate climbs to 16% because more of these roofs genuinely warrant a look and the storm reason is real (about 4.8 set), completion holds at 70% (3.4 completed), and sold rate rises to 45% because more of these roofs are actually due (about 1.5 sold). That is roughly fifteen sold jobs per 1,000 doors — more than double the output from the same reps, same hours, same pitch, purely from where they spent their day. That gap is the entire argument for taking targeting seriously, and it is why doors-per-inspection and revenue-per-100-doors are the numbers to watch.

Logging discipline

None of this works without clean dispositions. Every door gets a status: not-home, no-soliciting, not-interested, inspection-set, callback. Reps log on the phone at the door, not from memory at 8 p.m. Without disposition data your second-pass routing and your funnel metrics are both fiction.

A 30-day rollout plan

To make this concrete, here is how I would stand up the system in a month with an existing four-person team.

Week 1 — Build the target list. Pull assessor year-built data and Census housing-age and ownership data for your territory. Overlay NOAA storm history. Identify your top three to five due neighborhoods with the scoring model. If you are using an aerial-targeting platform, load the territory and generate the per-address due ranking now; if not, do the imagery bracketing by hand for your top neighborhoods only.

Week 2 — Train the read and the pitch. Two ride-along sessions on ground roof-reading drills. Script the opener and the four core objections. Drill the honest-claims language until it is reflexive. Confirm solicitation permits and document local allowed hours and no-knock rules.

Week 3 — Run tight routes and measure. Cluster the due list into 60-100-door walkable blocks. Assign geo-fenced clusters. Enforce neighbor-knocks around every set. Log every disposition. Place the densest, highest-due cluster in the 4-to-7 p.m. answer window.

Week 4 — Review and tune. Pull doors-per-inspection and revenue-per-100-doors by neighborhood. Run the targeting A/B if you have not. Re-score neighborhoods with what you learned. Build second-pass routes for the not-homes. Decide which inputs to trust more next month.

The mistakes that quietly kill door-to-door programs

A closing list of the failures I see most, so you can pre-empt them.

• Targeting on visible age instead of due-ness. You knock old roofs whose owners have no trigger and skip mid-life roofs that took hail. Layer storm and condition onto age.
• Optimizing pitch and ignoring density. Reps in trucks instead of on porches. Cluster tightly; doors-per-hour is half the equation.
• Skipping neighbor-knocks. The cheapest doors you will ever knock are the ones next to a set inspection. Make it mandatory.
• Treating storm history as proof. It is odds. Say "worth a free look," never "your roof is damaged." The honest framing also keeps you legal.
• Claiming exact roof ages. Age from imagery is a range. Reps who say "your roof is exactly 19 years old" get caught and lose trust. Speak in ranges.
• Crossing the claims line. No approval promises, no deductible handling, no free-roof promises. Document conditions, write estimates, let the homeowner own the claim and the insurer decide.
• No disposition logging. Without it, your routing and your metrics are guesses, and you can never prove your targeting works.
• Ignoring permits, hours, and no-knock signs. Cheap to comply with, expensive to violate.

The through-line of the whole system is simple: spend your reps' limited hours in front of the roofs that are actually due, be honest about what you know and what you only suspect, and measure the targeting decision itself. Crews that do those three things knock fewer doors and sell more jobs — which is the entire point of taking targeting seriously instead of leaving it to whoever feels lucky on the drive over.

FAQ

How can I tell a roof's age from the street without going up on it?

Read the high-signal tells first: granule loss in the gutters and at downspout splash blocks, curling or cupping shingle tabs, surface patches and color mismatches, and bald spots where the black mat shows through. Add context from the neighborhood's build year and visible re-roofs nearby. Ground reading only shows you one or two slopes, so treat it as a screen, not a verdict — the storm-facing or sun-baked slope you cannot see from the curb is often where the real wear is.

Is it better to target old roofs or storm-damaged roofs door to door?

Target the overlap. A roof that is both near end of service life and took recent hail or high wind is far more likely to be a full replacement and an easier conversation than one that is only old or only hit. Old-and-hit beats old-or-hit. Use roof-age cohorts to find aging neighborhoods, then overlay NOAA storm history to find which of those pockets actually took a significant event.

How accurate is estimating roof age from aerial imagery?

Aerial imagery gives you an age range, not an exact install date. By comparing current imagery against dated historical imagery you can bracket when a roof was last replaced, and weathering severity narrows the range further. But resolution, capture date, and sun angle all introduce error, and imagery cannot see attic ventilation, shingle warranty class, or a soft hail bruise hiding under intact granules. Always speak in ranges and confirm on the roof.

What is the best way to find old roofs in a specific neighborhood?

Start with the county assessor's year-built data to find the original roof cohort, then refine with imagery weathering and any age-from-imagery tooling. Cross-reference U.S. Census American Community Survey housing-age and ownership data, check permit portals for re-roof clusters that show whether the replacement wave has started or passed, and overlay NOAA storm history. Score each neighborhood on roof-age cohort, storm exposure, ownership rate, value band, and density before sending anyone to knock.

How many doors should a roofing rep knock in a day?

After travel, setup, and breaks a rep gets roughly five to six productive knocking hours. With tight, walkable clusters of 60-100 doors and the densest block placed in the 4-to-7 p.m. peak-answer window, 85-105 doors a day is realistic. The number matters less than density and targeting: a rep who knocks 70 well-targeted doors usually beats one who knocks 40 scattered ones.

What should I say at the door when targeting an aging roof?

Lead with the area reason (storm or neighborhood age), add one specific observation if you have it, and make a small ask: a free, honest look with a 'you might be fine' promise. Avoid leading with the sale. Have scripted, low-pressure answers ready for the common objections, and keep the honest-claims language reflexive: you document conditions and write estimates, the homeowner owns the claim, and the insurer decides coverage.

What legal rules apply to door-to-door roofing sales?

The FTC Cooling-Off Rule gives consumers three business days to cancel certain home sales of $25 or more, and you must provide the cancellation notice and forms when you sign at the door. Most cities require a solicitation permit and set allowed hours, and you must honor 'No Soliciting' signs and do-not-knock lists. Many states regulate public-adjusting and deductible handling by contractors — never promise claim approval, deductible coverage, or a free roof.

How do I prove my neighborhood targeting is actually working?

Track doors-per-inspection and revenue-per-100-doors, because those two numbers isolate targeting quality from pitch quality. Run a clean A/B: have one rep work neighborhoods chosen by gut and another work neighborhoods chosen by your scoring model and a due-roof list, hold the pitch and hours constant, and compare over two to three weeks. If scored targeting lowers doors-per-inspection and raises revenue-per-100-doors, you have your business case.

Can I tell a homeowner their roof was definitely damaged by a recent storm?

No. Storm history and forecasts are odds, not proof. The fact that a hail-producing cell passed over a neighborhood raises the probability that roofs there have impact damage, but it does not prove any specific roof was damaged. Frame it as 'your area took hail in that storm, so a free inspection is worth it,' and let the actual roof inspection determine what is really there.

Does software like RoofPredict replace the in-person roof inspection?

No, and you should be skeptical of anything that claims it does. Aerial-targeting tools estimate a roof-age range per address and model storm exposure per roof so you can rank and sequence which doors to knock first, which saves the days of desk work that targeting otherwise takes. But the age figure is a range with real uncertainty and the storm model gives odds, not proof. You still confirm condition, ventilation, and hidden damage on the roof itself.