Canvassing App vs. Property Data List for Roofing Leads: Which Actually Fills Your Pipeline

Most roofing owners frame this as one question with one answer: "What's the best canvassing app, or should I just buy a property data list?" After watching enough crews burn a season on the wrong tool, I'll tell you the framing is the problem. A canvassing app and a property data list are not competitors. They solve two different halves of the same job, and the contractors who win treat them as a pair — a targeting layer that tells you which doors are worth a knock, and a field layer that gets a body to those doors and records what happened.

The trouble is that vendors on both sides pretend their half is the whole. The canvassing-app salesperson shows you a beautiful map with pins and brags about route optimization, but the pins are every house in the neighborhood — including the 9-year-old roofs and the rentals you'll never sell. The data-list salesperson hands you a spreadsheet of 4,200 "high-intent homeowners" and walks away, leaving you with no way to assign, track, or measure a single knock against it. Buy only one and you've bought half a system.

What follows is the decision the way a sales manager who has actually run both should make it: what each tool genuinely does, where the data really comes from (and how wrong it is), the unit economics of cost per door versus cost per list record, the specific failure modes that quietly kill canvassing programs, and how to stitch a targeting layer and a field layer into one revenue motion instead of two disconnected purchases. I'll be specific about where roof-age and storm data help and where they're oversold, because honest targeting beats hopeful targeting every time.

The two things you're actually buying

Strip the marketing away and a roofing field-sales program has exactly two jobs:

1. Decide where to knock. Out of the 18,000 single-family homes in your service radius, which 1,500 are worth a rep's time this month? This is a targeting and data problem.
2. Get reps to those doors and capture what happens. Routing, territory assignment, the conversation, the disposition, the follow-up. This is a field-execution problem.

A property data list is a targeting input. It's a static (or periodically refreshed) export of addresses with attributes attached — owner-occupied vs. rental, estimated home value, year built, sometimes a roof-age estimate, sometimes a storm-exposure flag. You buy it, you load it, you knock it. On its own it does nothing in the field; it's a spreadsheet.

A canvassing app is a field-execution layer. It draws territories on a map, assigns reps, optimizes the order of stops, gives the rep a phone screen to log "not home / not interested / pitched / appointment set," and rolls those dispositions up into a manager dashboard. On its own it doesn't know which doors are good — it'll happily route a rep down a street of brand-new roofs.

Here's the part vendors blur: most canvassing apps include some targeting data, and most data lists include some light field capability. A canvassing app might let you color the map by home value or year built. A data-list provider might give you a mobile view to mark a door as knocked. Those overlaps are why owners think they're choosing between two versions of the same product. They aren't. One is built to move reps efficiently; the other is built to score addresses accurately. Whichever one you skimp on is the one that quietly caps your results.

What a property data list really is (and where the data comes from)

Before you can judge a list, you need to know what's underneath the attributes, because "roof age" and "homeowner" are not measured the way the sales deck implies.

Ownership and occupancy

The most reliable field on any list is owner-occupied vs. absentee/rental, and it's the one that matters most for canvassing. It's derived from county assessor and recorder records — does the owner's mailing address match the property address? When they match, it's almost certainly owner-occupied. This data is genuinely good, usually 90%+ accurate, because it comes straight from public tax rolls. If your reps are knocking rentals, the tenant can't authorize a roof replacement and the owner isn't there. Filtering to owner-occupied alone can lift your contact-to-conversation rate meaningfully.

Year built

Also assessor-sourced and reliable. "Year built" is the original construction date, which is not the same as roof age — and that distinction is where most targeting goes wrong. A 1985 house may have a 3-year-old roof from a re-roof in 2023; a 2015 house has a roof that's genuinely 11 years old. Year built is a useful floor (a 2024 build won't need a roof) but a terrible direct proxy for roof age on older homes.

"Roof age" estimates

This is the field that gets oversold hardest, so be clear-eyed. Almost no list has a measured roof installation date for an arbitrary home. There is no national "roof was replaced on this date" registry. What providers actually do is estimate roof age from some combination of:

• Year built (the lazy version — just assumes the original roof).
• Permit records, where the jurisdiction issues re-roof permits and makes them queryable. Permit coverage is wildly uneven county to county; plenty of re-roofs happen without a pulled permit.
• Aerial/satellite imagery analysis that infers material and apparent condition, sometimes change-detection between image dates to catch a recent replacement.

So treat any "roof age" field as a band, not a date: recent / mid-life / due / overdue. A list that tells you a specific roof is "19 years old" is implying a precision that doesn't exist for most addresses. A list (or platform) that buckets a roof into "likely overdue" based on year-built plus a permit gap plus imagery is being honest about the uncertainty — and that band is still useful, because you're trying to skew the odds of a productive knock, not predict the future.

Storm exposure

Some lists tag whether an address sits inside a recent hail or high-wind footprint. The underlying data here can be solid — NOAA's Storm Prediction Center, the National Weather Service, and IBHS publish hail and wind event data, and several commercial vendors model hail swaths from radar. But a "storm-exposed" flag is a probability of having been hit, not evidence of damage on that specific roof. A 1.25-inch hail report two miles away does not mean the shingles on 412 Oak Street are bruised. Use it to prioritize a neighborhood for inspection, never to tell a homeowner their roof is damaged before anyone has looked at it.

Home value and demographics

Estimated market value, equity bands, length of residence, sometimes income. Mostly modeled, decent in aggregate, noisy per-address. Useful for de-prioritizing the bottom and top tails (you may not want the rental-heavy block or the gated estates with property managers) but not worth obsessing over.

The honest accuracy summary

If a list vendor won't tell you the source of a field, assume it's modeled and discount it.

How to interrogate a data vendor before you pay

List vendors are slick and most contractors buy on the demo. Don't. Run the same five questions past every provider and the weak ones fall away fast:

1. What's the source of each attribute, field by field? Make them separate the assessor-sourced facts (ownership, year built) from the modeled estimates (roof age, value, income). If they can't or won't, the modeled fields are doing more work than they admit.
2. How often is it refreshed, and what's the lag from the source? Assessor data updates on the county's schedule, not the vendor's. Ask when their last pull was for your counties. "Updated monthly" means nothing if your county only republishes the tax roll twice a year.
3. What's your match rate in my specific counties? Coverage is regional. A vendor that's excellent in suburban Dallas may be threadbare in rural Ohio because permit data isn't digitized there. Ask for a sample export of 200 addresses in your footprint and check it against what you already know.
4. How is roof age derived, and what's the confidence? If the answer is "year built," you're paying for a column you could pull from the assessor yourself. If it's permits plus imagery with a confidence band, that's worth money.
5. What are the use terms? Can you load it into your CRM and your canvassing app, or are you licensing a one-time view? Are appended phones marked for litigation/DNC scrubbing? A list you can't move into your field tool is half-useless.

The tell of a good vendor is that they volunteer the limits — "our permit coverage is thin in these three counties, lean on imagery there." The tell of a bad one is uniform confidence everywhere, which is statistically impossible across a real metro.

Sample size sanity check

Do the arithmetic before you buy a giant list. If your metro has 18,000 owner-occupied single-family homes and a realistic 22% of roofs fall in the due/overdue band at any time, that's about 3,960 genuinely-targeted addresses. With four reps knocking 100 doors a day, four days a week, you cover 1,600 doors a week — so a well-targeted set is roughly two-and-a-half weeks of knocking before you're re-knocking no-answers. Buying a 30,000-record "whole county" list when you can physically knock a fraction of it is how you end up with stale data you paid for and never touched. Buy what you can work in a refresh cycle, then refresh.

What a canvassing app really does

The field layer is where reps spend their day, so the value is in friction removed per knock, not in flashy features. The pieces that matter:

Territory drawing and assignment. Draw a polygon on the map, assign it to a rep or team, and lock it so two reps don't knock the same street. Good apps support hex-grid or block-level territories and let you see coverage — what's been knocked, what hasn't, what's stale.

The disposition screen. The single most important UI in any canvassing app. When a rep walks away from a door, how many taps to log the outcome? The fastest apps make "not home," "not interested," "callback," "pitched," and "appointment set" one tap each, with optional notes and a photo. If logging a door takes more than a few seconds, reps stop doing it accurately, and your data dies.

Routing. Order the stops so a rep isn't crisscrossing. Useful, but honestly oversold — most residential canvassing is dense enough that a rep walks a street regardless. Routing matters more for spread-out rural territories or for follow-up runs where you're hitting scattered callbacks.

Leave-behind and capture. Per-door or per-rep QR codes on a door hanger, a mobile form, voice notes for the rep to dictate "older couple, wife handles decisions, roof looked rough on the north slope." The QR matters more than people think: it's the bridge from a no-answer door to an inbound lead two days later.

Live progress and accountability. As a manager, can you see who's where, how many doors per hour, set rate per 100 knocks, in real time? This is half of why you buy the app — not the routing, the accountability.

Hand-off to CRM. When a rep sets an appointment, does it create a lead in your CRM with the source intact, or does someone re-type it that night? Re-typing is where leads die.

Notice what a canvassing app does not do well on its own: tell you which doors are worth knocking. It'll show you every parcel. The targeting is your job — which is exactly why you also need a data layer.

The metrics a canvassing app should expose (and what each tells you)

The dashboard is where you manage the program, so know which numbers matter and which are vanity. Here's the set I actually watch.

Vanity metrics to ignore: total pins on the map, total doors "available," and anything that measures effort without tying to an outcome. A rep who knocks 140 doors a day and sets nothing is a problem, not a star.

The disposition taxonomy that actually pays off

Vague dispositions ruin the data. "Not interested" is a black hole — not interested why? The taxonomy I'd standardize on: Not home, Renter/wrong owner (so you can flag a list error), New roof / recently done (gold — feed this back to fix your roof-age model and stop knocking that house for years), Not now / callback (with a date), Pitched-no, Inspection set, and Hostile/no-knock (so nobody re-knocks and you stay clean on local ordinances). That "new roof" disposition is quietly one of the most valuable fields you'll ever collect, because it's ground-truth that corrects modeled roof age.

The core comparison

Here's the side-by-side, framed around the decision rather than feature checklists.

The pattern: the list makes each knock more likely to land; the app makes each rep do more knocks and proves what happened. Multiply those two and you get pipeline. Do one without the other and you've bought a denominator with no numerator, or vice versa.

Unit economics: cost per door vs. cost per record

Owners argue about monthly software price when they should be arguing about cost per productive knock. Let me work it with realistic-but-illustrative numbers so you can plug in your own.

The canvassing side

The app itself is cheap relative to labor. Say a canvassing app runs you a per-seat monthly fee for 4 reps. The real cost is the rep. A canvasser knocks somewhere around 20–30 doors an hour in dense residential, with maybe 30–40% answering. Over a 4-hour shift that's roughly 80–120 doors, of which 25–45 are conversations.

Now the targeting multiplier. If you knock an untargeted street, a big chunk of those conversations are wasted on 6-year-old roofs and renters. If you knock a targeted list — owner-occupied, roof in the due/overdue band, inside a recent storm footprint — a far larger share of conversations are with people who plausibly need a roof.

Suppose untargeted knocking sets 1 appointment per 100 doors and targeted knocking sets 2.5 per 100 (these are illustrative — your numbers will differ, but the direction is consistent across every disciplined crew I've seen). Same rep, same hours, same app cost. The targeting layer more than doubled output. That delta is the entire argument for spending on data.

The list side

Property lists price a few ways. Per-record pulls might run a few cents to a couple dimes per address. A 3,000-record targeted pull is real money but trivial against a rep's loaded labor cost for the days they'll spend knocking it. Subscription/platform pricing gives you ongoing refreshed access, which matters because lists rot.

The rot factor nobody budgets for

A property list starts decaying the day you buy it. People move (roughly 8–9% of Americans change residence each year, per Census data), roofs get replaced, ownership changes. A list that's six months old has meaningful drift; a year old and it's noticeably wrong. If you bought a one-time export and you're still knocking it nine months later, you're paying full labor cost to knock partly-dead data. This is the strongest argument for a living data source over a static spreadsheet.

Putting it together

The cheapest cost-per-appointment is almost always "both," because the list cuts the wasted-knock tax and the app lets you measure which list segments actually convert so next month's list is sharper.

When a canvassing app should be your first buy

Be honest about your situation. Lead with the app when:

• You already have decent targeting — a storm just hit a defined neighborhood, or you work a tight farm area you know cold. The "where" is solved; you need execution and accountability.
• You have 2+ reps and zero visibility. If you can't answer "how many doors did Marcus knock yesterday and what was his set rate," the app pays for itself in management leverage alone.
• Your problem is leakage, not targeting. Reps set appointments that never make it into the CRM, follow-ups fall through, no-answer doors never get a second visit. The field layer plugs those holes.
• Storm response. After a hail event, the whole neighborhood inside the swath is your target. You don't need a sophisticated list; you need to flood the area, track coverage, and not double-knock. App first.

When a property data list should be your first buy

Lead with data when:

• You're spread across a metro with no obvious farm. Knocking randomly in a big service area is a slow death. You need the list to concentrate reps where roofs are actually aging.
• You're light on storms. Retail/age-based roofing (no recent hail to ride) lives or dies on roof-age targeting. The whole game is finding the due and overdue roofs before a competitor does.
• You're also running direct mail. A list isn't only for knocking — it's the audience for mailers, microsites, and follow-up. One good targeted list feeds multiple channels.
• You have one or two reps and can't afford wasted days. With limited field hours, every wasted knock hurts more. Targeting is the highest-leverage dollar.

The failure modes that quietly kill canvassing programs

This is the part the vendor demos never cover, and it's where most money leaks. Watch for all of these.

1. Buying a list and knocking it blind. No app, no tracking. Reps work the spreadsheet on paper or in their heads, set rate is invisible, and you can't tell which segments converted. You'll never know if the "overdue roof" band beat the "storm-exposed" band because you never recorded it. You bought targeting and threw away the learning.

2. Buying an app and pointing it at every parcel. The map shows 18,000 pins, reps knock the nearest ones, and your conversion looks like random door-knocking because it is random door-knocking with a nicer interface. The app didn't fail; you skipped the targeting layer.

3. Treating a roof-age estimate as gospel. A rep tells a homeowner "our records show your roof is 22 years old and needs replacement." The homeowner re-roofed four years ago. Now you've burned credibility and maybe invited a complaint. Roof age is a band and a reason to inspect, never a claim about their specific roof. Train reps to say "homes in your neighborhood are at the age where roofs start failing — mind if I take a look at yours?"

4. Disposition decay. Week one, reps log every door perfectly. Week six, they batch-log "not home" for the whole street at lunch because the screen is slow. Your data is now fiction. Fix it by choosing a fast disposition UI and by managing to it — pull the door-level log and spot-check.

5. No second touch. A no-answer is not a dead lead; it's an untimed one. Most doors are no-answers. If your program has no systematic re-knock at a different time of day, plus a leave-behind with a QR that lets the homeowner come to you, you're throwing away the majority of your reach.

6. The CRM gap. Appointment set on a doorstep at 6 p.m., re-typed (or not) at 9 p.m., first-touch source lost. Three weeks later nobody knows that win came from canvassing, so you can't compute cost per win, so you can't decide whether to expand canvassing. Source attribution has to be automatic.

7. Compliance blind spots. No-knock registries and local solicitation ordinances are real; some municipalities require a permit to canvass. If your list came with appended phone numbers and your reps start cold-calling them, you've stepped into TCPA territory. And on the door, the script matters — promising outcomes you can't control is how you end up with a regulator's attention. More on that below.

The script and the line you cannot cross

Because storm and insurance work is where the money is, it's also where roofers get themselves in trouble. The targeting tool and the canvassing app don't break the law — the script does. Here's the bright line, and it applies whether the lead came from a list or a knock.

A roofing contractor may: inspect a roof, document damage thoroughly with photos and measurements, write an accurate, Xactimate-aligned estimate to repair their own scope of work, hand that estimate to the homeowner, and state facts about the work they propose. That's your lane and it's a wide one.

A roofing contractor may not, for compensation, do the things that constitute unlicensed public adjusting in most states: negotiate or "handle" the insurance claim on the homeowner's behalf, interpret the policy or what's covered, promise a specific payout or that the claim will be approved, promise to "waive," "absorb," or "eat" the deductible, advertise a "free roof," or otherwise represent the homeowner against their insurer. State insurance regulators and the NAIC are explicit about this, and several state DOIs (Texas's TDI among the most active) have gone after contractors for crossing it.

So the safe canvassing frame after a storm is: "There was a hail event in this area. We'll inspect and document your roof for free and, if there's damage, give you a written estimate to repair it. You file the claim with your insurer; they decide coverage." You document, you estimate, the homeowner files, the insurer decides. Never "we'll get your roof approved," never "we handle the insurance," never "the deductible's on us." Teach your reps the do-not-say list explicitly — it protects your license and your reputation, and it's a competitive edge when the contractor down the street is making promises they can't keep.

A targeting list and a canvassing app help you find the roofs likely to qualify (age plus storm exposure) and run the documentation workflow cleanly. They are not, and should not pretend to be, claim-handling tools.

The same list also feeds your mail and digital — don't waste it on knocking alone

One of the biggest misreads in this whole debate is treating the targeting list as a canvassing input only. It isn't. A good ranked list of due-roof, owner-occupied homes is the audience for three channels at once, and the channels compound.

Direct mail to the same addresses. The due-roof list is exactly who should get a mailer. The cheapest, highest-trust version is a personalized piece — the homeowner's neighborhood, the roof-age angle (handled honestly), a clear reason to act — with a tracked response path. When you mail and knock the same target set, the knock lands warmer because the homeowner has seen your name. Sequence it: mail drops, then reps knock that segment three to five days later, then the no-answers get a second mail touch. That's a multi-touch farm, not a one-shot blitz, and it's why a static spreadsheet you bought once is the wrong tool — you want a target set you can re-touch across mail and door over weeks.

A personalized landing page / microsite per home. The mailer and the door hanger both carry a QR. Scanning it shouldn't dump the homeowner on your generic homepage — it should land them on a page about their roof: a roof profile, the storm history for their address, the cost-of-waiting framing, and a single form or call button. That's what turns the 60-70% of doors that are no-answers into inbound leads on the homeowner's timeline instead of yours.

The follow-up CRM record. Every channel — knock, mail response, QR scan — should land in the same lead record with source intact, so you're not running three disconnected programs against one list. This is the strongest practical argument for a platform over a pile of point tools: the list, the mail, the microsite, and the field knock all reference the same address and the same lead.

The math on this is simple. If you've already paid to identify and rank a due-roof home, knocking it once and walking away is leaving most of the value on the table. Touch it across mail, door, and a QR-driven microsite, and your cost-per-acquired-customer on that list drops because you're amortizing the targeting spend across more shots on goal.

A buyer's checklist before you sign anything

Run every vendor — list or app — through this. If they fail three or more, keep looking.

For a property data list / targeting layer:

• They'll tell you the source of every attribute, separating assessor facts from models.
• Roof age is expressed as a band with a confidence signal, not a fake exact date.
• They'll provide a sample export in your counties to spot-check match rate.
• Storm flags cite their underlying event/radar source and are framed as exposure, not damage.
• You can export/license the data into your CRM and field app, rather than only viewing it.
• Appended contact data is DNC/litigation-scrubbed and dated.
• It's refreshable on a schedule, not a one-time spreadsheet that rots.

For a canvassing app:

• One-tap disposition screen with a sane taxonomy (incl. "new roof" and "no-knock").
• Territory draw, assignment, and locking so reps don't double-knock.
• Real-time manager view of doors/hour and set rate by rep.
• Set rate sliceable by list segment (this is the learning loop — non-negotiable).
• Per-door / per-rep QR leave-behind that routes to a real capture page.
• Automatic lead creation into your CRM with immutable first-touch source.
• Offline mode that syncs — reps lose signal in cul-de-sacs and basements.
• A funnel view to cost per appointment and cost per win.

For an integrated platform (both halves in one):

• The ranked target list and the field app share one address/lead record.
• The funnel ties wins back to the target segment that produced them.
• CRM sync is two-way to the specific CRM you already run.
• The vendor is honest about data limits (band not date, odds not proof).

A real workflow: turning a list and an app into pipeline

Let me lay out the actual operating rhythm of a crew running both layers well. This is the thing demos never show — the boring weekly cadence that produces results.

Step 1 — Define the target before you draw a territory

Don't start on the map. Start with the filter. Decide the audience: owner-occupied, single-family, roof in the due or overdue band, and — if you're storm-riding — inside the most recent qualifying hail/wind footprint. De-select rentals, brand-new builds, and the streets you've worked in the last 90 days. You now have a ranked address set, not a neighborhood.

Step 2 — Rank, don't just filter

A flat list of 3,000 "qualifying" homes is still too many to knock at once. Rank them — by an opportunity score that blends roof-age band, storm exposure, and home characteristics — so reps work the top of the list first. The top 500 should be measurably better doors than the bottom 500, and your tracking will eventually prove (or disprove) that.

Step 3 — Cut territories from the ranked set

Now draw the map. Carve the ranked addresses into rep-sized territories — dense enough to walk, balanced so each rep has comparable opportunity. Assign and lock them in the app so nobody double-knocks.

Step 4 — Knock with a documentation-first script

Reps work the door with the inspect-and-document frame. Fast dispositions on every door: not home, not interested, callback, inspection set. Voice note the color commentary. Leave a door hanger with a QR on every no-answer so the house can convert later.

Step 5 — Inspect, document, estimate

For every inspection, photograph the roof methodically (every slope, flashing, penetrations, soft metals, gutters, collateral), measure, and produce a clean written estimate of your repair scope. If it's a storm claim, the homeowner files; you've handed them documentation, not a coverage opinion.

Step 6 — Capture the lead with source intact, automatically

The appointment or inspection should create a lead in your CRM from the field, stamped with the canvassing source, the rep, the territory, and the address — no nightly re-typing. That immutable first-touch source is what lets you measure the channel later.

Step 7 — Second-touch the no-answers

The morning after, the no-answer doors get a re-knock at a different time block and the QR-driven inbounds get worked like any other lead. The majority of your reach is in this layer; programs that skip it leave most of the value on the street.

Step 8 — Read the funnel and re-cut next week's list

At week's end, look at the full funnel: doors knocked → conversations → inspections → estimates → wins, with set rate per 100 by rep and by list segment. Which band converted? Did storm-exposed beat age-only? Feed that back into next week's ranking. This learning loop is the whole point of running both layers — and it's impossible if you knocked a spreadsheet blind.

A worked example: one month, two approaches, side by side

Numbers make the abstract argument concrete. These are illustrative, not benchmarks — plug in your own — but the structure of the comparison holds across every disciplined crew I've watched. Take one rep, four field days a week, four weeks, 100 doors a day: 1,600 doors for the month.

Approach A — canvassing app, untargeted parcels. The app routes the rep through whatever's nearby. Of 1,600 doors, contact rate runs ~35%, so ~560 conversations. But a big share are renters and recent roofs, so the qualified-conversation rate is low. Say the set rate lands at 1.0 appointment per 100 doors → 16 appointments. Of those, half show and inspect → 8 inspections; estimates written on 6; close a third → 2 jobs. The app cost is a rep seat plus the rep's loaded labor. Cost per job is dominated by all those wasted knocks on bad doors.

Approach B — same app, same rep, but knocking a ranked due-roof, owner-occupied, storm-aware list. Same 1,600 doors, same ~35% contact rate (~560 conversations), but now most conversations are with plausible buyers because the renters and 6-year-old roofs were filtered out before the rep ever walked the street. Say the set rate climbs to 2.5 per 100 → 40 appointments. Show/inspect ~20, estimates on ~16, close a third → 5-6 jobs.

Same labor, same hours, same app. The only difference is the targeting layer, and it roughly tripled jobs. The list cost is a rounding error against the labor you already spent — that's the entire economic case. And because Approach B logged set rate by segment, next month you can lean harder into whichever band converted best. Approach A learned nothing it can act on.

Now flip it: imagine Approach B's list without the app — knocked blind off a spreadsheet. You'd still get better doors, but you couldn't see set rate by rep, couldn't slice by segment, couldn't catch the rep batch-logging "not home," and couldn't prove which jobs came from canvassing. You'd have the targeting and throw away every ounce of the learning. That's the symmetry: the list without the app is unmeasured, the app without the list is unfocused.

What separates the pros from everyone else

After the tooling decision, the gap between crews that crush canvassing and crews that quit it comes down to a handful of operating habits that have nothing to do with which app you bought.

• They knock the same farm repeatedly, not once. A territory worked four times over a season — knock, mail, re-knock, microsite follow-up — outperforms four different neighborhoods hit once each. Familiarity converts.
• They protect the script line religiously. No promises about claims, payouts, or deductibles. The crews that get sued or fined are the ones whose reps freelance the storm pitch. The disciplined ones document and estimate, full stop.
• They treat "new roof" dispositions as a data gift. Every rep-confirmed recent roof gets fed back so the targeting model stops sending anyone there. Over a season this sharpens the list more than any vendor refresh.
• They manage to set rate, not door count. A rep knocking 80 good doors at a 3% set rate beats one knocking 140 random doors at 0.5%. Pay and praise on outcomes.
• They close the loop in the CRM. Source is automatic, cost per win is computed monthly, and the canvassing budget is decided on that number — not on vibes about whether "door-knocking still works."

None of that is a feature you buy. It's a program you run, and the tools exist to serve it.

Where RoofPredict fits this exact decision

The reason I framed this as "two halves of one job" is that the cleanest version of this program is one platform that does both the targeting and the field execution, so the learning loop in Step 8 actually closes. That's what RoofPredict is built to be, and it's worth being concrete about which capabilities map to which half of the problem.

The targeting half (the "property data list," but living and ranked). RoofPredict scores every home in your service area by roof-age band — recent, mid-life, due, overdue — combined with per-roof storm exposure and an opportunity score, and hands you a ranked target audience, house by house, with a "why this home" evidence chain so a rep knows why a door made the list. You draw territory with a hex-map, import your own addresses by CSV, and filter to storm-hit areas. Crucially, it's a living source rather than a one-time export that rots in your truck — which directly attacks the cost-of-stale-data problem from the economics section. And the framing is honest by design: roof age is a band, not an install date; storm exposure is odds of being hit, not proof of damage. It tells you which roofs likely qualify, not what's wrong with a specific roof.

The field half (the canvassing app). From that ranked list, RoofPredict builds door-knock routes, assigns canvassers, and runs a mobile field app with next-stop, fast outcome forms, voice notes, and leave-behind QR codes — per-home or lookup QR for the door hanger — plus live route progress so a manager sees coverage and set rate in real time. The no-answer-becomes-inbound problem is handled by the QR: every targeted home also gets a personalized microsite and PDF report (roof profile, storm history, cost-of-waiting) with a lead-capture form, so the house can convert two days after the knock.

The part that makes both worth it — measurement and CRM. Every field outcome flows into a lead pipeline (new → contacting → appointment → inspected → won/lost) with an immutable first-touch source, so canvassing-sourced wins stay attributed. It syncs two-way to the CRM you already run — JobNimbus, AccuLynx, ServiceTitan, HubSpot, Roofr, SalesRabbit, CompanyCam and others — so nobody re-types a doorstep appointment at 9 p.m. And the results funnel shows delivered → views → form → calls → leads → wins with cost-per-lead and cost-per-win, actual vs. estimate vs. benchmark, broken out so you can finally answer "did the overdue-roof band beat the storm-exposed band?" — which is exactly the Step 8 learning loop that a bought spreadsheet plus a generic canvassing app can't close on their own.

The honest pitch: if you buy a standalone list and a standalone canvassing app, you can absolutely make it work — plenty of good crews do — but you'll be the integration glue, manually reconciling which list segment produced which win. RoofPredict's reason to exist is that the targeting, the field app, the QR/microsite capture, the CRM sync, and the funnel measurement are one system, so the loop closes by default instead of by spreadsheet.

If you only have budget for one this quarter

A real answer, not a hedge:

• Storm-driven shop with a defined footprint and 2+ reps and no visibility? App-side first. Your targeting is the storm map; your bleeding is execution and tracking. Get reps measured and stop leaking appointments.
• Retail/age-based shop spread across a metro? Data-side first. Without targeting you're knocking noise, and no field app saves you from working bad doors efficiently.
• Can swing one integrated platform? That's the lowest cost-per-win path, because you stop paying the integration tax and you get the learning loop. The whole argument of this piece is that the two halves multiply, and the multiplication only happens cleanly when they share one source of truth.

Whatever you choose, instrument it. The contractors who win this aren't the ones with the prettiest map — they're the ones who can tell you, at the end of the month, their cost per win by source and by target segment, and who use that number to re-cut next month's list. That feedback loop, far more than any single tool, is what separates a canvassing program from a bunch of guys knocking doors.

The bottom line

A canvassing app and a property data list are not rivals; they're the field half and the targeting half of one customer-acquisition motion. The list (or, better, a living, ranked targeting layer) decides where the productive doors are and keeps you off the renters and the new roofs. The app gets bodies to those doors fast, captures every outcome, and proves what happened. Skip either and you cap your results — wasted knocks on one side, invisible performance on the other.

Buy with your situation in mind, keep the data honest (roof age is a band, storm exposure is odds, neither is a claim about a specific roof), keep your script inside the document-and-estimate lane, and above all close the measurement loop so each month's targeting gets sharper. Do that, and the answer to "canvassing app or property data list" stops being a choice and starts being a system.

FAQ

Is a canvassing app or a property data list better for roofing leads?

Neither alone — they do different jobs. A property data list is a targeting layer that decides which doors are worth knocking (owner-occupied, roof in the due/overdue band, storm-exposed). A canvassing app is a field-execution layer that routes reps to doors, captures dispositions, and reports set rates. The lowest cost per appointment comes from running both: the list cuts wasted knocks, and the app measures which list segments actually convert so next month's targeting is sharper.

How accurate is the roof-age data on a property list?

Treat it as a band, not a date. There is no national registry of roof installation dates, so 'roof age' is modeled from year built, re-roof permit records (coverage varies wildly by county), and sometimes aerial imagery. A list that claims a specific roof is exactly 19 years old is overstating its precision. A platform that buckets roofs into recent / mid-life / due / overdue is being honest, and that band is still useful for skewing the odds of a productive knock.

What does roofing canvassing cost per door?

The app seat fee is minor; the real cost is rep labor. A canvasser knocks roughly 20-30 doors an hour in dense residential, with 30-40% answering. The variable that moves cost per appointment most is targeting: knocking an untargeted street wastes conversations on new roofs and rentals, while a targeted, owner-occupied, due-roof list can materially raise the set rate for the same labor. Optimize cost per productive knock, not the software's monthly price.

How fast does a bought property list go stale?

Quickly. Roughly 8-9% of U.S. households move each year (Census data), roofs get replaced, and ownership changes, so a one-time export drifts within months and is noticeably wrong by a year. If you're still knocking a spreadsheet you bought nine months ago, you're paying full labor cost to work partly-dead data. A living, periodically refreshed targeting source beats a static list for exactly this reason.

Can I tell a homeowner their roof is too old based on the list?

No — say it carefully. Roof-age data is a band and a reason to inspect, not a fact about their specific roof. A homeowner may have re-roofed recently in a way the data missed, and telling them 'our records show your roof is 22 years old' when it's 4 years old destroys credibility. Train reps to say 'homes in your area are at the age where roofs start failing — mind if I take a look at yours?' and let the inspection establish the facts.

How do storm-exposure flags work and can I rely on them?

Storm flags model whether an address sits inside a recent hail or high-wind footprint, using NOAA Storm Prediction Center, National Weather Service, and IBHS event data plus radar-based hail modeling. That's a probability of having been hit, not evidence of damage on a particular roof. Use it to prioritize a neighborhood for inspection — never to tell a homeowner their roof is damaged before anyone has physically looked at it.

What can a roofer legally say about insurance when canvassing after a storm?

Stay in the document-and-estimate lane. You may inspect, photograph and document damage, and write an accurate repair estimate for your own scope, then hand it to the homeowner. You may not, for compensation, negotiate or handle the claim, interpret coverage, promise a payout or approval, waive or absorb the deductible, or advertise a 'free roof' — those cross into unlicensed public adjusting in most states. The safe frame: you document and estimate, the homeowner files, the insurer decides coverage.

Do I need a canvassing app if I only have one or two reps?

If your problem is visibility or leakage — appointments not reaching the CRM, no-answer doors never re-knocked — yes, even at one or two reps the accountability and capture pay off. But if you're a very small retail shop spread across a metro, your highest-leverage first dollar is usually the targeting data, because with limited field hours every wasted knock hurts more and targeting prevents the most waste.

How do I keep canvassing dispositions from becoming fiction?

Pick an app with a one-tap disposition screen so logging a door takes seconds, then manage to it. The common failure is reps batch-logging 'not home' for a whole street at lunch because the UI is slow, which poisons your data. Pull the door-level log periodically and spot-check against rep claims. Accurate dispositions are what let you compute set rate by segment and improve targeting — without them, you can't learn anything.

How does this connect to my CRM so I can measure cost per win?

The field tool needs to create a lead with an immutable first-touch source the moment an appointment is set, ideally syncing two-way to your CRM (JobNimbus, AccuLynx, ServiceTitan, HubSpot, Roofr, SalesRabbit, CompanyCam and similar) so nobody re-types it at night. With source preserved, you can run the full funnel — doors to conversations to inspections to wins — and compute cost per lead and cost per win by source and by target segment, which is how you decide whether to expand canvassing.