Roofing Lead Generation in 2026: Choose Channels by Profit, Not Lead Count

What counts as a roofing lead?

A useful roofing lead is a serviceable person-and-property opportunity with a legitimate reason for contact, enough information to route it, and an owner responsible for the next action. A web form, phone call, purchased record, postcard scan, or door conversation is an inquiry. Calling all of those “leads” makes vendors look comparable when they are selling fundamentally different things.

Use one funnel vocabulary across every channel:

Stage	Required evidence	Common false positive
Inquiry	Unique person/property reached the company through a recorded source	Spam, vendor, job applicant, duplicate call
Qualified opportunity	Service, geography, timing, property, and contactability meet written rules	Home outside the territory or service offered
Booked inspection	Real calendar slot, assigned owner, contact confirmation	“Requested appointment” with no accepted time
Kept inspection	Rep attended and recorded disposition/evidence	Reschedule or drive-by counted as complete
Estimate/opportunity	Defined work path and next decision date	Automated estimate with no homeowner engagement
Signed contract	Valid agreement under the company’s accounting rules	Contingent or canceled agreement counted as a win
Collected job	Revenue collected, job cost reconciled, gross profit known	Signed revenue treated as cash or profit

The distinction is not semantic. A marketplace may charge for a valid phone connection. An agency may report a form submission. A canvassing manager may count an appointment. The owner needs signed and collected outcomes, but each earlier stage still needs a fixed definition so the team can identify where quality or execution changes.

The minimum source-and-stage data contract

Every opportunity should retain the original source, latest meaningful source, campaign and creative identifiers, first-touch time, person and canonical property ID, service-area result, contact permission/provenance where relevant, assigned owner, stage timestamps, appointment outcome, loss/disqualification reason, contract value, job cost, cancellation, and collection status. Preserve the original source; do not overwrite it when a receptionist sends a reminder or a salesperson creates a new record.

Deduplicate at the person-property level while retaining touch history. If the same homeowner clicks a search ad, receives an earlier postcard, and later calls from a referral, the CRM should not create three opportunities or award three contracts. Store the touches, apply one declared attribution rule for routine reporting, and use controlled tests when the business question is incrementality. Attribution answers “which touch receives reporting credit?” Incrementality asks “what would have happened without this spend?” They are not the same question.

Which roofing lead-generation channel is best?

There is no universally best channel. The correct choice depends on current inbound opportunity, first-party records, territory density, time to signal, gross profit, sales capacity, and how much control the contractor needs. A channel with lower cost per inquiry can be worse when it creates long drives, duplicate competition, missed appointments, or poor-match work.

Channel	Intent at first touch	Time to signal	Control and ownership	Capacity it consumes	Best-fit condition
Referrals and partners	Usually high	Variable	High relationship control; uneven volume	Follow-up and partner care	Strong reputation and adjacent professional network
Google Business Profile/local organic	High when visible	Slow to compound	Owned presence on a rented platform	Reviews, content, accuracy, intake	Established service area and consistent operations
Paid search	High but auction-dependent	Fast	Strong query/geo control; platform dependency	Immediate call and inspection capacity	Search volume exists and offline outcomes are measurable
Local Services Ads	High local service intent	Fast after verification	Platform-defined lead product and dispute rules	Fast intake and profile/review operations	Eligible, verified operator with competitive unit economics
CRM reactivation	Warm/varied	Fast	High first-party control	Data cleanup, permission, follow-up	Meaningful historical customer and lead records
Targeted direct mail	Proactive attention	Weeks	High address/message/holdout control	Calls, inspections, print/postage cash	Defined residential territory and property-level selection
EDDM	Broad awareness	Weeks	Route/message control; broad selection	Potentially broad response	Route saturation or neighborhood awareness is the goal
Canvassing	Proactive attention plus live qualification	Days	High territory and rep control	Field labor, supervision, safety	Dense neighborhoods and repeatable field management
Events/community/retail	Trust/context dependent	Weeks to months	High local relationship control	Staff time and consistent presence	Community brand strategy with follow-through
Shared lead marketplace	Variable; often actively shopping	Immediate	Low exclusivity/data-source control	Rapid contact and competitive selling	Speed needed and tested close economics survive sharing

Choose for the company you have now

An owner-led roofer should first make the Business Profile accurate, answer the phone, ask for qualified referrals, preserve source data, and work the existing pipeline. One additional pilot is enough. Running five channels at once produces noise and follow-up debt.

A stable growth company with dependable inspection and production capacity can combine high-intent search capture with one proactive channel. If it has years of records, CRM reactivation is usually the most logical owned-data test. If the CRM is thin but the service territory is defined, compare targeted direct mail with territory-managed canvassing.

A multi-rep company can allocate channels by branch and capacity, but it needs territory governance, source persistence, appointment ownership, and cohort reporting before it adds volume. A company entering a new market has no local historical funnel; it should use conservative budgets, verify Business Profile eligibility and representation rules, build relationships, and treat early paid or outbound work as learning, not proof of a mature acquisition cost.

During a legitimate storm surge, capacity and safety become the constraints. Do not automatically open every channel because homeowners are calling. Use the storm-response operating plan to cap bookings, prioritize existing obligations, separate event reports from property diagnoses, and pause acquisition before service quality breaks.

Score channels before funding them

Rate each candidate from 1 to 5, then apply weights appropriate to the quarter:

channel score = expected contribution × weight + time-to-signal score × weight + control score × weight + capacity-fit score × weight + compliance confidence × weight + learning value × weight

Expected contribution should carry the largest weight, but uncertainty matters. A channel with a credible 4/5 expected return and high learning value may be a better pilot than one whose promised 5/5 return is based on a vendor’s unrelated national benchmark. Record the assumptions and confidence range; do not manufacture precision from unknown close rates.

Also distinguish controlled relationships and first-party outcomes (customers, permissioned contacts, referrals, and the company’s own source/outcome history), rented audience access (search profiles, ad platforms, and marketplaces), and proactive attention (mail or respectful canvassing). The company does not own a homeowner, a mailbox, a search result, or a social audience. A resilient portfolio develops direct customer relationships and portable measurement while using rented and proactive channels intentionally. Dependence on one platform is a business risk even when that platform works today.

What should a roofing company do first at each growth stage?

Fix the earliest broken conversion or capacity constraint, then add one channel whose audience, operating burden, and measurement match the company’s current stage. A new roofer needs trust and reachable intake. A growing retail company may need selected outreach. An established company may have recoverable opportunity in its CRM. A storm-loaded company may need to stop acquisition rather than increase it.

Company stage or condition	First growth work	Next channel to test	Metric that decides	Do not do yet
New owner-led roofer	Accurate Business Profile, referrals, reviews from real work, answered phone, simple CRM states, service area and job economics	One bounded local channel, often search capture, partnerships, or a small targeted-mail/field pilot	Qualified kept inspections and conservative contribution	Launch five channels or buy a long-term lead commitment without a baseline
Growing single-market retail roofer with spare capacity	Repair source, call, booking, and follow-up leakage	Targeted direct mail, CRM reactivation, canvassing, or paid search based on available audience/intent	Incremental contribution per dollar and inspection hour	Optimize cost per form while appointments or production are overloaded
Established local brand with historical records	Resolve CRM identities, then apply exclusions and identify relationship-specific segments	CRM reactivation plus selected net-new outreach	Incremental realized contribution versus holdout	Treat every prior customer as a replacement prospect
Proven operator entering a new market	Establish truthful local presence, call route, service/branch ownership, and conservative unit economics	Small paid-intent and targeted local cohorts	Market-specific funnel, not home-market average	Copy the mature-market budget or CAC into the launch forecast
Multi-rep/multi-branch company	Common source/stage definitions, territory, capacity, and branch ownership	Branch-specific channel portfolio with centralized measurement	Marginal contribution and capacity by branch/cohort	Let branches double-contact the same property or compete for attribution
Restoration/event surge	Existing customers, safety, verified event evidence, intake and production caps	Controlled readiness/post-event sequence only within capacity	Appropriate inspections, backlog, complaints, cash and quality	Open every channel because volume is visible
Capacity-constrained company	Repair the active queue, calls, inspections, estimates, production, documentation, or collections	No new acquisition until release/stop gates recover	Stage age, throughput, service level, cancellations, margin	Spend more because top-of-funnel reports look weak

A practical first-week diagnostic

Before choosing a lead vendor or increasing budget, reconstruct the last 50–100 inquiries if volume permits:

1. identify the original source and canonical property;
2. remove spam, duplicates, vendors, applicants, existing-customer service, and out-of-area work from the new-lead denominator;
3. trace qualified opportunities to attempted contact, booking, kept inspection, estimate, contract, cancellation, job, collection, and realized gross profit;
4. record response and stage times;
5. classify the first reason each lost opportunity stopped;
6. quantify current call, inspection, estimate, production, and cash capacity.

If 30% of qualified calls are not reached quickly, the next dollar belongs in roofing call coverage or response ownership. If appointments are plentiful but completion is weak, fix confirmation, routing, and delay. If inspections occur but estimates or follow-up age, buying inquiries adds avoidable rework. If the complete funnel is healthy and future capacity is empty, acquisition is the constraint.

New owner-led roofer: build trust before channel complexity

A new company may not have enough CRM history, reviews, branded search, or conversion data to support an elaborate portfolio. The owner’s first assets are trade competence, a clearly served geography, dependable communication, real customer proof, and a record of every lead and next action.

Referral and partner relationships can produce high-trust work but may be uneven. A correct Business Profile and basic local site capture active local intent over time. Paid search or Local Services Ads can test existing intent if eligibility, review profile, call coverage, and economics are ready. A small targeted direct-mail or canvassing pilot can earn attention under the company’s name when the owner has inspection time and a credible planning offer.

Do not use an aggressive discount or insurance fear to compensate for a new brand. Underwrite lower recognition, make the response professional, and keep the first cohort small enough to serve exceptionally.

Growing retail roofer: choose the audience the operation can convert

Once the company knows gross profit and can reliably move a qualified inquiry to a completed inspection, the channel question becomes allocation. If thousands of prior estimates, repairs, or customer properties have no current next action, work the eligible CRM reactivation segments. If the historical file is thin but the residential territory is clear, property-selected mail can create an address-level audience without paying a click auction. If dense routes and trained reps are available, canvassing can add live qualification. If high-intent searches exist, paid capture can remain part of the mix.

The better channel is the one with stronger marginal contribution after its specific capacity cost. Mail consumes call and inspection capacity. Canvassing consumes rep hours, drive, and management. Paid search consumes immediate response capacity. CRM reactivation consumes data review and relationship care. Treat those costs as part of acquisition.

Established brand: compound the installed base without overclaiming ownership

An established contractor controls more useful history: completed work, repairs, estimates, service interactions, responder sources, outcomes, and exclusions. That history can make reactivation and neighborhood programs more relevant. It does not turn every customer into permanent marketing inventory.

Use lifecycle-appropriate service for former customers, respect contact permissions and declines, and separate known relationships from unknown household mail. Nearby completed work can create recognition when used truthfully and with customer privacy. Preserve the original source and assisted touches so the company can see whether mail, branded search, referral, and call handling worked together without awarding four contracts to one job.

New-market expansion: learn local economics again

The home market supplies a method, not a guaranteed result. New-market recognition, housing stock, roof systems, permitting data, competition, weather patterns, media auction, drive time, labor, supplier access, and customer expectations differ. Establish a branch/service boundary and response owner before starting outreach.

Run small, comparable cohorts and report the new market independently. A blended company CAC can hide an unprofitable launch because the mature market carries it. Use conservative cash limits and explicit exit gates. A new market that cannot fulfill warranties or customer updates after the launch is not a growth market.

What an “owned” or controlled audience actually means

Marketing shorthand often calls email, CRM, or direct mail “owned.” Use a narrower definition. A roofer can control:

• its customer relationship and service promise;
• its lawful, permissioned contact records and company exclusions;
• its website, phone experience, creative, offers, and field conduct;
• its canonical property/opportunity record;
• its original-source, touch, contract, cancellation, collection, and gross-profit history;
• its ability to export that history from vendors under contract.

It does not own a person, a purchased address, the mailbox, an ad platform, a Business Profile, a search query, or an audience on another network. Direct mail is powerful because the roofer can choose and measure an address cohort and put its own brand in the market, not because the address grants permission for cold calls or texts.

The durable advantage is portable learning. A company that knows which selected properties, prior relationships, messages, routes, and calls created collected gross profit can allocate the next cohort without trusting a platform’s lead score. That is why PorchRocket emphasizes property identity, exclusions, experiments, and writeback alongside campaign execution.

The next-dollar test

For each eligible channel, forecast downside, base, and upside:

marginal contribution = incremental realized jobs × realized gross profit per job − next-cohort media/outreach − incremental intake, inspection, and sales cost

Then ask whether the company can safely absorb the resulting calls, inspections, estimates, production starts, working capital, and customer commitments. Fund the channel only if the downside is survivable, the base clears the required return, the learning is identifiable, and stop rules are observable.

This test can favor direct mail because selection excludes poor-match properties and the company retains cohort-level outcome history. It can favor search because local intent is already active. It can favor CRM because prior acquisition work created context. The answer should change with the company’s evidence. A channel guide that always chooses its own product is not an allocation model.

Should a roofer capture active intent or earn new attention?

Work active intent first when qualified homeowners are already searching, calling, referring, or sitting unworked in the CRM. Use outbound outreach when the company has spare capacity, a serviceable property audience, a truthful reason to reach out, and the patience to measure incrementality.

The two jobs use different channels and expectations:

Acquisition job	Typical channels	Primary risk	First diagnostic
Capture active local intent	Business Profile, organic, paid search, Local Services Ads, referrals	Paying for inquiries that cannot reach or book the team	Missed calls, impression/lead availability, booking capacity
Recover known opportunity	CRM reactivation, prior estimates, customer/service lifecycle	Treating stale history as blanket permission or current need	Identity, exclusion, segment, next action
Earn targeted attention	Direct mail, canvassing, community/partner outreach	Paying to interrupt poor-match properties or overclaiming a signal	Serviceable universe, reason, holdout, capacity
Build future preference	Local content, reviews, partnerships, completed-job presence	Slow/uncertain attribution and inconsistent execution	Audience need, proof, production cadence, branded-search trend

Do not evaluate an outbound channel by search-lead response expectations. A homeowner who was not shopping before a postcard or knock starts at a different stage. Conversely, do not use a six-month brand metric to excuse a paid-search campaign that cannot connect calls this week.

Audit active-intent capture before increasing bids or budget

For each high-intent source, verify:

• the public phone/form works on mobile and after hours;
• service areas and offered services reflect actual operations;
• local profiles are accurate and comply with platform representation rules;
• calls/forms retain source and property identity;
• spam, existing customers, applicants, vendors, and duplicates are separated;
• real availability exists for the service/territory shown;
• response reaches an owner with a due time;
• qualified and converted outcomes return to ad systems where appropriate;
• reviews and public replies follow a truthful, noncoercive policy.

Google’s service-area guidance should be checked for current representation requirements; a service-area list is not a license to claim locations the company does not operate.

Earn the right to start outbound outreach

Before outbound spend, require:

1. a defined eligible property/relationship universe;
2. a specific, evidence-bounded homeowner offer;
3. exclusions and cross-channel contact pressure;
4. call, inspection, sales, production, and cash capacity;
5. a treatment/holdout or credible counterfactual;
6. a fixed contribution model and maturation window;
7. quality, complaint, and stop thresholds.

Direct mail and canvassing can make a business less dependent on auctions. They can also create a large queue of inspections with weak fit if selection and messaging are poor. The direct-mail guide and canvassing guide own those execution details.

Track branded search and direct response as outcomes, not proof

Outbound attention may later arrive as a branded search, direct visit, main-line call, referral mention, or neighborhood conversation. Preserve exposure and self-reported discovery where useful. A rise in branded search can support a story about awareness; it does not prove which channel caused the rise without a design.

Use geographic or address-level holdouts, staggered rollouts, and stable time windows where practical. Report multi-touch paths, but do not split one contract into arbitrary percentages that sum neatly while avoiding the counterfactual question.

Should roofers buy leads from lead-generation companies?

Buy leads only when you can define the product, verify its origin and consent, route it immediately, and prove that the resulting contracts clear your contribution threshold. Do not buy a label called “roofing lead.” A purchased lead can be useful, but its economics are especially sensitive to duplication, intent, geography, job type, and speed.

Common commercial models include:

• Exclusive inquiry: promised to one contractor, though the homeowner may independently contact others.
• Shared inquiry: sold or distributed to multiple contractors.
• Live transfer: a call is connected after basic screening; billability rules matter.
• Booked appointment: a time is requested or accepted; define confirmation, serviceability, and no-show treatment.
• Qualified opportunity: supposed to meet filters; the contract must state every filter and evidence.
• Performance-based engagement: payment is tied to a later event; define attribution, cancellation, audit rights, and prohibited incentives.

Questions to put in the lead contract

Ask the vendor, in writing:

1. What exact event makes a lead billable?
2. Is it exclusive, and what does “exclusive” exclude?
3. What site, ad, phone number, affiliate, or partner generated it?
4. What did the consumer see and affirm before data was transferred?
5. Can the vendor produce the consent record and applicable disclosure?
6. Which ZIP codes, property types, services, languages, and timing filters are enforced?
7. How are duplicates, existing customers, agents, renters, wrong numbers, spam, and out-of-area records handled?
8. What is the replacement/dispute window, evidence standard, and response time?
9. May the vendor or its affiliates resell, retarget, or retain the record?
10. What reporting and record-export rights survive cancellation?
11. Are there minimums, auto-renewal, price-change, exclusivity, or long-term commitments?
12. Who is responsible for Do Not Call, channel permission, calling-hour, recording, text, and state-law compliance?

The FTC’s advertising resources make the central principle simple: marketing claims must be truthful, non-deceptive, and appropriately supported. A contractor cannot outsource reputational or legal risk merely by buying a data feed. Telemarketing, texts, recordings, and state requirements need qualified review for the actual workflow; this guide is operational guidance, not legal advice.

Shared-lead economics: a sensitivity example

Assume a shared inquiry costs $140. Of 100 purchased inquiries, 70 are valid/contactable, 45 are serviceable and qualified, 25 book, 18 keep an inspection, and 3 sign. Purchased lead cost is $14,000. If incremental intake/inspection/sales cost is $5,400 and gross profit per contract is $5,250, the cohort creates $15,750 of gross profit against $19,400 of variable acquisition expense: negative $3,650 campaign contribution before fixed overhead.

At five signed contracts with the same costs, gross profit becomes $26,250 and contribution becomes $6,850. Nothing about the vendor CPL changed; the economics changed because contract yield moved from 3% to 5%. That is why the decision sheet must vary validity, kept-inspection rate, contract yield, cancellation, gross profit, and internal handling cost. The worked numbers are illustrative, not PorchRocket or industry benchmarks.

Red flags include guaranteed revenue, a source or consent path the vendor will not disclose, no duplicate policy, vague geography, automatic credit denial, long lock-in before a pilot, reports that end at “lead delivered,” and pressure to call or text in ways your compliance owner has not approved.

Define billable, accepted, qualified, and disputed as separate states

A vendor may bill a lead before the contractor knows whether it fits. Build an intake ledger:

State	Required evidence
Delivered	Vendor lead ID, timestamp, source/site/campaign, payload, consent/disclosure artifact reference
Billable under contract	Delivery met the exact commercial definition
Accepted for review	Technical delivery succeeded and entered the contractor’s system once
Serviceable/qualified	Written geography, service, property, intent, and contact rules pass
Duplicate/existing	Matched to a prior person-property/opportunity under the contract window
Disputed	Reason code, evidence, submission time, vendor response due
Credited/rejected	Final commercial result and credit amount
Worked outcome	Attempts, booking, kept inspection, contract, cancellation, collection

Do not discard disputed leads from the performance denominator without showing them. A source with 30% credits may have an attractive net CPL while consuming substantial office time and producing unstable volume.

Price dispute labor and speed requirements

If the contract allows 48 hours to dispute a lead, the company needs weekend and absence coverage. Calculate internal review minutes, evidence collection, and vendor follow-up. Add that cost to the source.

Similarly, if lead value depends on immediate contact, price the coverage needed to respond. A cheap shared lead delivered at 10:30 p.m. is not cheap if no approved handler exists until morning.

Test with a cancellable cohort

Cap spend, dates, geography, service, and lead count. Use the same scripts and intake rules as production. Predefine:

• maximum invalid, duplicate, out-of-area, and unverifiable-source rates;
• minimum contact, qualification, kept-inspection, and retained-contract outcomes;
• maximum cost per mature retained contract and per constrained inspection hour;
• complaint/permission and customer-experience guardrails;
• dispute turnaround and data-export acceptance;
• automatic termination or no-renewal unless explicitly approved.

Do not sign a long exclusivity or minimum-volume commitment because the first ten leads looked good. Early lead order and small samples can be unrepresentative.

Preserve the consumer-facing evidence

For lead sources that rely on a form, call, affiliate, or partner, retain or obtain the page/script/disclosure version, timestamp, source URL/phone, fields, affirmative action, and transfer chain appropriate to the workflow. The vendor’s summary field consent=true is not the underlying artifact.

Qualified counsel should review calling, text, prerecorded/artificial voice, recording, state, and seller/vendor obligations. The FTC’s Telemarketing Sales Rule guidance is one federal starting point, not a complete legal opinion.

Direct mail, canvassing, PPC, or CRM reactivation?

Compare all four on the same final outcomes: incremental contribution per deployed dollar and per constrained sales hour. Then use channel-specific metrics to diagnose the path. Cost per click, response rate, doors per hour, and contact rate are useful operating measures, but none is the finish line.

Targeted direct mail

Direct mail is appropriate when the company can define serviceable properties, wants address-level control, can wait through production and response windows, and has a reasoned inspection offer. It is also unusually testable: randomized address holdouts can reveal lift without relying entirely on last-touch attribution.

In PorchRocket’s 128,000-mailer experiment, high-priority personalized mail produced a 0.750% contract yield compared with 0.109% for broad generic mail. The contract-yield risk ratio was 6.86 (95% CI 4.81–9.77). Under disclosed assumptions of a $15,000 average signed contract and 35% gross margin, the high-priority personalized arm produced $1,071,470 in campaign contribution after outreach and incremental inspection/sales cost. Those are study results, not a forecast. Selection, market, creative, delivery, capacity, margin, and measurement all affect transferability. The direct-mail buyer’s guide contains the complete arm definitions, economics, sensitivity analysis, and limitations.

CRM reactivation

CRM reactivation is often the fastest owned-data test for a contractor with years of estimates, repairs, customers, and stalled opportunities. It still requires identity resolution, serviceability, exclusion, appropriate contact basis, accurate messaging, and a durable writeback to the CRM.

In PorchRocket’s 72,000-record trial, 24,000-record personalized and generic arms were compared with a 24,000-record holdout. Personalized outreach produced 161 contracts (0.671% yield), generic produced 62 (0.258%), and holdout produced 10 (0.042%). Personalized versus generic contract-yield RR was 2.60 (95% CI 1.94–3.48). Under the study assumptions, personalized outreach produced $742,505 in contribution and $468,800 more than generic. The CRM reactivation guide explains data reduction, segment outcomes, permissions, neighborhood/monitoring experiments, and causal limits.

Canvassing

Canvassing is attractive when neighborhoods are dense, reps can be trained and supervised, safe/legal working windows are clear, and the company values live qualification. It consumes labor immediately, so contracts per paid field hour and drive/duplicate waste matter as much as doors knocked.

In PorchRocket’s 90,000-door crossover study, standard routes produced 150 contracts across 30,000 doors (0.500%), ranked routes produced 318 (1.060%), and ranked routes plus a property brief produced 450 (1.500%). Ranked-plus-brief versus standard contract-yield RR was 3.00 (95% CI 2.50–3.61), while contracts per 100 paid hours rose from 8.7 to 29.7. These results do not make a map a causal roof diagnosis and do not guarantee another crew’s result. See the canvassing software and territory guide for crossover design, route metrics, economics, and confounders.

Paid search and Local Services Ads

Search reaches people expressing immediate intent, but it operates in a local auction and can become expensive precisely when search volume is most obvious. Landing-page promise, geography, call coverage, negative keywords, job filters, and offline outcome feedback determine whether clicks become profitable work.

Google’s current documentation distinguishes a raw lead from a qualified lead and a converted lead, and supports importing offline stages so campaigns can be evaluated deeper in the funnel. As of June 15, 2026, Google also documents a Data Manager API migration for certain offline-conversion and enhanced-conversion-for-leads uploads. Implementation details can change; use the current Google Ads conversion documentation rather than an old integration tutorial.

Google Local Services Ads currently lists roofing as eligible, positions the product as pay per lead rather than pay per click, and requires screening/verification. “Pay per lead” does not mean “pay per signed job.” Apply the same qualified, kept, signed, canceled, collected, and contribution waterfall used for every other source.

Do channel combinations create lift or attribution fog?

Combine channels only with a written mechanism. Examples: a preapproved direct-mail exposure might increase recognition for a safely assigned canvassing visit; search may capture interest that a community sponsorship created; receptionist coverage may prevent paid calls from being lost. Define the expected interaction, preserve touch history, and test staged geographies or address cohorts. Do not keep adding touches until every vendor claims the same contract.

How should a roofing company build a lead-channel portfolio?

Build a portfolio around different acquisition jobs, time horizons, and dependencies, not an arbitrary percentage for each channel. A resilient program usually includes a base of owned reputation/relationships, active-intent capture, a controlled first-party recovery motion, and one measured outbound channel when capacity permits.

Assign each channel a role and decision horizon

Portfolio role	Example	Decision horizon	Evidence needed
Always-on capture	Profile, referrals, organic, bounded search	Weekly health; monthly/quarterly economics	Delivery, qualification, offline outcomes, local search volume
Recoverable inventory	CRM segments and triggers	Cohort/trigger maturation	Eligibility, holdout, segment contribution
Scalable outbound	Targeted mail or territory canvassing	Pilot then score-band/territory scale	Incremental contracts, marginal contribution, capacity
Experimental	New partner, creative, audience, platform	Fixed budget/sample	Predeclared hypothesis and stop rule
Compounding brand asset	Reviews, local evidence, useful content, community trust	Quarterly/annual	Branded search/direct response, assisted paths, market holdouts where possible

Do not require every channel to produce the same payback date, but do require each to have a falsifiable role. “Brand awareness” without a defined audience, behavior, indicator, and review date becomes a permanent exception to accountability.

Reserve capacity before allocating money

Convert each channel budget into low/base/high expected inquiries, calls, bookings, inspections, estimates, contracts, and production starts. Put them on the same weekly calendar. Include the existing pipeline and organic inquiries.

Example: paid search may run continuously, a 20,000-mailer campaign may concentrate response within an in-home window, and canvassing may create appointments during evening shifts. The portfolio fits only if front desk, calendar, estimators, and production can absorb the combined high case.

Avoid correlated platform risk

Paid search, Local Services Ads, Business Profile visibility, and organic search may all depend on one platform ecosystem and the same local search volume. Treat them as distinct tactics but not fully independent risk. Likewise, direct mail and canvassing can share the same property signal and saturate the same neighborhood.

Record vendor, platform, data source, audience, and operational dependencies. Build continuity for lost account access, profile suspension, tracking outage, data-provider failure, or call routing disruption. Owned customer/property records and source history should remain usable if a platform changes.

Fund learning explicitly

Separate a controlled experimental budget from proven-channel release. An experiment is successful if it produces a reliable stop or scale decision, even when its immediate contribution is negative within the approved downside. This prevents teams from hiding failed tests inside blended marketing spend, or refusing to test because every dollar is expected to hit a mature CAC on day one.

Every experiment needs an owner, hypothesis, population, comparison, primary outcome, guardrails, maximum loss, maturation, and decision date. Do not run ten underpowered experiments in parallel.

How much should a roofing company spend on lead generation?

Do not begin with a universal percentage of revenue. Begin with serviceable capacity, gross profit per contract, kept-inspection rate, close rate, working capital, and the contribution the company needs after acquisition. A percentage can be a planning cross-check, but it cannot tell whether the next dollar creates or destroys value.

Reverse the funnel from the desired contracts

Suppose the company wants 12 incremental signed contracts in a month. Its qualified-lead-to-booked rate is 70%, booked-to-kept is 75%, kept-to-signed is 30%, and inquiry-to-qualified is 65%.

required kept inspections = 12 ÷ 0.30 = 40

required bookings = 40 ÷ 0.75 = 53.3, round to 54

required qualified opportunities = 54 ÷ 0.70 = 77.1, round to 78

required inquiries = 78 ÷ 0.65 = 120

The plan therefore needs roughly 120 incremental inquiries and 40 incremental inspection slots. If the team has only 20 open slots, buying 120 inquiries is not a growth plan; it is a backlog and reputation plan.

Set a maximum acquisition cost from contribution

Let:

• GP = expected gross profit per signed and retained contract
• I = incremental intake, inspection, sales, and financing cost per contract not included in GP
• B = contribution buffer reserved for fixed overhead, risk, and profit
• CACmax = maximum external acquisition spend per contract

Then:

CACmax = GP − I − B

If expected gross profit is $5,250, incremental internal selling/inspection cost is $900, and the company requires $2,000 to remain after acquisition, maximum external CAC is $2,350. At a 4% inquiry-to-retained-contract yield, the maximum external cost per inquiry is $94. At 2%, it is $47. The channel’s allowable “lead price” changes drastically with qualification and execution.

Three sensitivity cases

Case	100 inquiries cost	Signed, retained contracts	Gross profit at $5,250	Internal variable handling	Campaign contribution
Fragile funnel	$9,000	2	$10,500	$3,600	-$2,100
Base funnel	$9,000	4	$21,000	$4,800	$7,200
Strong funnel	$9,000	7	$36,750	$6,600	$21,150

These are sensitivity cases, not market benchmarks. The lesson is that conversion and internal handling can move the spend ceiling more than a modest CPL negotiation. Estimate with retained contracts and realized job-level gross profit when possible, not signed revenue.

Add cash-flow guardrails

Acquisition spend leaves before many roofing projects collect. Account for deposits, material orders, subcontractor terms, permit timing, financing funding, claim timing where applicable, cancellation windows, seasonality, and production backlog. Set a weekly cash exposure limit and a pause condition. A positive modeled contribution does not solve a working-capital shortfall.

Budget at the margin, not from blended CAC

Suppose the first $10,000 of search spend creates five retained contracts and the next $10,000 creates three because higher-volume queries/territories are less efficient. Blended external CAC at $20,000 is $2,500 across eight contracts. The marginal CAC of the second tranche is $3,333. If the company’s maximum external CAC is $2,350, the historical average hides an unprofitable next dollar.

For every scalable channel, build bands by spend, score, geography, route, audience, or time:

Release band	Incremental spend	Incremental retained contracts	Marginal external CAC	Marginal contribution	Capacity effect
Proven core	Set target	Set target	Set target	Set target	Set target
Expansion 1	Set target	Set target	Set target	Set target	Set target
Expansion 2	Set target	Set target	Set target	Set target	Set target

Release the next band only when its conservative case clears the threshold. The best first segment of a mailing list does not validate every lower-ranked address. The best paid-search exact terms do not validate broad expansion. The densest canvassing area does not validate a distant route.

Watch saturation in operations as well as media

Marketing saturation appears as rising media cost, falling response, weaker audience depth, or creative fatigue. Operational saturation appears as slower answers, longer appointment delay, lower kept rate, estimator overload, reduced follow-up, production backlog, and lower realized margin. The second can make every source appear to decay at once.

Plot channel volume against downstream service levels. If close rates fall only after weekly inspections exceed 45, capacity, not source quality, may set the efficient frontier. Hiring or process improvement could raise the ceiling, but it must be proven rather than assumed in the budget.

Set downside and liquidity limits

For each channel, calculate maximum cash deployed before a mature decision, expected worst-case variable loss, and obligations created if the campaign succeeds. Direct mail is prepaid; paid media can be throttled daily; lead contracts may have minimums; canvassing adds payroll and supervision; contracts create material and production cash needs.

The owner should know: “If this test produces zero contracts, cash loss is X. If it produces the high case, we need Y inspection hours, Z material/production cash, and this many crew starts.” Both tails matter.

How should roofers measure lead quality and ROI?

Use a cohort waterfall that follows each source from first inquiry through collection and gross profit, while showing speed, capacity, quality, and complaint guardrails. Keep the denominator visible on every percentage.

Measure	Definition	Review cadence
Valid/contactable rate	Non-spam inquiries with usable identity/contact ÷ inquiries	Daily/weekly
Serviceable qualified rate	Opportunities meeting written service/geo/job rules ÷ inquiries	Weekly
Speed to first owned action	Time from arrival to accountable human/system action; show median and 90th percentile	Daily
Booking and kept rates	Booked or kept inspections ÷ a fixed upstream cohort	Weekly by mature cohort
Contract and retained yield	Signed or retained contracts ÷ inquiries and ÷ kept inspections	Weekly/monthly
Collected gross profit	Reconciled gross profit from the cohort	Monthly/quarterly as jobs mature
External CAC	Channel spend ÷ retained contracts	Monthly
Fully variable acquisition cost	Channel + incremental intake/inspection/sales ÷ retained contracts	Monthly
Campaign contribution	Collected or expected gross profit minus variable acquisition expense	Monthly with maturity label
Capacity/quality	Slot fill, backlog age, no-show, cancellation, rework, complaint, opt-out	Daily/weekly

Cohorts prevent premature conclusions

Group inquiries by arrival week or campaign and give them the same maturity window. A search lead received yesterday should not be compared with a direct-mail cohort that has had 60 days to sign and build. Show “immature,” “mature,” and “closed” rather than automatically labeling every unresolved record lost. Lock historical source and stage timestamps so later data cleanup does not rewrite the experiment.

Measure lift where possible

Randomized address or record holdouts are powerful for direct mail and reactivation. For channels that cannot use individual randomization, stage comparable geographies or time windows, document differences, and avoid claiming causality from a simple before/after chart. Search volume, weather, staffing, competitor behavior, and seasonality can move simultaneously.

A weekly channel decision card

For every source, show cohort size and maturity, spend, valid and qualified counts, bookings, kept inspections, contracts, retained contracts, expected and realized gross profit, variable handling cost, CAC, contribution, sales hours, complaints/opt-outs/refunds, and backlog age. Add a short confidence note.

Use predeclared actions:

• Scale carefully: mature cohort clears contribution floor; capacity and quality remain inside guardrails; tracking reconciles.
• Hold and learn: economics remain uncertain, sample is small, or one diagnosed stage has a repairable issue.
• Stop or redesign: mature contribution is below floor, consent/source is unverifiable, complaint or quality threshold breaks, or fulfillment capacity is exceeded.

Never scale solely because response or CPL improved. A cheap channel can consume the most expensive resource in the company: a qualified estimator’s time.

Diagnose the first funnel break

Pattern	Inspect first	Do not do yet
Spend/impressions but few inquiries	query/audience, ad/list/offer, service area, page/phone delivery	Blame sales close rate
Many inquiries, low validity	spam, source/affiliate, duplicate rules, tracking, billable definition	Increase response staffing blindly
Valid but low serviceability	targeting geography, job/property filters, public service representation	Call faster and hope
Serviceable but low booking	answer/wait, offer, appointment availability, intake burden	Buy more inquiries
Good bookings, high no-show	confirmation, expectation, source intent, delay, route	Call every lead more often
Kept inspections, low estimates	qualification, inspection standard, service/job mismatch	Declare marketing unqualified
Estimates, low contracts	scope/price/presentation, follow-up, financing, rep and source mix	Change CPL target only
Signed jobs, low retained/collected	cancellation, production promise, financing/claim/cash, job quality	Report signed ROAS as final
All channels weaken together	intake, capacity, sales staffing, season/weather, production	Replace every vendor
Last-touch source looks strong, holdout shows little lift	brand/organic capture, attribution stealing, concurrent campaigns	Scale on platform ROAS alone

Attach root-cause codes to fixes and observe whether the next cohort changes. A dashboard that names the weak stage but never assigns a corrective owner becomes reporting theater.

Run a quarterly allocation review

Use mature fixed cohorts and forward capacity. For each channel, bring:

• purpose and current dependency risk;
• spend and internal variable cost;
• marginal, not just blended, CAC and contribution;
• realized gross profit and cash maturity;
• constrained inspection/sales hours consumed;
• quality, complaint, cancellation, and rework;
• sample uncertainty and incrementality design;
• next expansion band and downside exposure;
• changes in market, platform, data, or operations.

Then choose: protect base, expand marginally, hold, redesign, test, or exit. Reallocate gradually unless a safety, compliance, cash, or tracking defect requires immediate stop. Large quarterly swings can destroy learning and partner/channel continuity.

Reconcile platform reporting to CRM and finance

Platform conversions, calls, and leads are operational signals. The contractor’s CRM should own qualification, appointment, contract, and job linkage; finance/job cost should validate collected economics. Reconcile counts by stable lead/click/call IDs and documented match rules.

Google’s qualified and converted lead guidance supports feeding deeper lead stages back to Google Ads. That can improve bidding and reporting, but it does not make platform attribution an incrementality study. Audit privacy, identifiers, stage definitions, duplicates, and uploads under the current platform documentation.

What must happen after the lead arrives?

Every inquiry needs immediate acknowledgement, correct triage, a visible booking path, one owner, and an exception queue until it reaches a documented disposition. Speed matters, but “call within five minutes” is not a complete operating system. A rushed call to the wrong person with no available appointment is still failure.

The intake path should:

1. Capture name, callback, property address, service requested, immediate safety/emergency context, source, and preferred timing.
2. Verify serviceability without making a remote roof, damage, coverage, or price diagnosis.
3. Detect existing customer/property records and preserve the new touch.
4. Route emergencies and existing obligations under written rules.
5. Offer only calendar capacity the assigned team can honor.
6. Send a truthful confirmation with expectations and change/cancel path.
7. Create an owner, next action, due time, and escalation if unanswered or incomplete.
8. Record outcome and disqualification reason using governed options.

If inbound coverage is the problem, use the roofing AI receptionist guide to compare AI, live answering, employees, and hybrid escalation. If leads enter but age, skip stages, or disappear between departments, use the roofing operations guide to define lifecycle entry/exit rules and exception ownership.

The no-lead-left-behind queue

Build saved queues for unassigned inquiries, failed call/message delivery, no available appointment, duplicate needing review, missing address/service, reschedule without new time, no-show without disposition, estimate without next decision date, and stale signed-but-incomplete jobs. Each exception needs an owner, age, next action, and escalation. Automation should surface and route exceptions, not silently mark them complete.

When is PorchRocket the right growth partner?

PorchRocket fits when a roofing company wants to create and convert opportunities using first-party records, property/territory prioritization, direct outreach, call and office coordination, and closed-loop measurement around its existing CRM. It is not a generic SEO/PPC agency, a shared-lead marketplace, an insurer, a public adjuster, or a promise of automatic jobs.

PorchRocket is most credible for a residential contractor with a defined service area, meaningful gross profit per appropriate job, enough call/inspection/production capacity to run a controlled pilot, a CRM or usable records, and an internal owner willing to maintain process and measurement. It is a poor fit for a company demanding guaranteed contracts, unable to service the response it purchases, unwilling to use exclusions or lawful messaging, or unwilling to reconcile outcomes beyond lead count.

Model	Primary deliverable	Data/control	Internal ownership required	Best when
Marketing agency	Media/content/campaign execution	Varies by contract and platform	Brand, sales, tracking	Search/content/media is the primary constraint
Lead marketplace	Inquiries, transfers, or appointments	Usually lower source ownership	Fast intake and competitive sales	Immediate volume is worth duplication/variability
Point software	Tool used by internal team	Usually higher if export/API is sound	Strong process/admin owner	Team already knows the workflow
Managed revenue operations	Opportunity selection through coordinated handoff and measurement	High first-party/operational emphasis	Executive owner plus departmental operators	Data and handoff gaps cross channels and teams

Ask any provider to show the lead definition, data provenance, required systems, implementation owner, reporting denominator, holdout/test design, contribution formula, cancellation rights, data export, and what happens when capacity fills. A top-tier proposal says what it cannot infer and what the contractor must still own.

A 90-day roofing lead-generation plan

Days 1–15: establish truth

• Define stages and denominators from inquiry through collected job.
• Reconcile the last 90–180 days by original source, maturity, and gross profit where available.
• Audit phone/form delivery, Business Profile representation, service areas, calendar capacity, duplicate handling, and source persistence.
• Calculate maximum CAC from company-specific gross profit, internal variable cost, and contribution floor.
• Name one owner for each exception queue.

Days 16–30: repair conversion before buying volume

• Fix unanswered/abandoned calls, failed forms, incorrect routing, unavailable calendar slots, missing confirmations, and stale follow-up.
• Clean exclusion, serviceability, loss reason, and contact-permission data.
• Send qualified and converted outcomes back to paid platforms where appropriate, using current platform documentation and privacy/security review.
• Establish the weekly channel card and capacity pause rules.

Days 31–60: recover owned opportunity

• Segment existing estimates, prior repairs, customers, stalled opportunities, and appropriate maintenance/monitoring records.
• Run a compliant, controlled reactivation pilot with a holdout where feasible.
• Keep messaging tied to known history; do not manufacture urgency or imply a remote diagnosis.
• Measure response through retained contracts and variable contribution.

Days 61–90: pilot one incremental channel

• Choose targeted mail, canvassing, or paid search based on territory, intent, time, and capacity.
• Predeclare cohort, budget, run length, success threshold, quality guardrails, and stop rule.
• Preserve control geography/records when feasible.
• Review mature contribution and constrained-hour productivity, not a blended dashboard of every touch.
• Scale one step, redesign, or stop. Add another channel only after the first creates a trustworthy learning loop.

Roofing lead-generation FAQs

What is the best source of roofing leads?

The best source is the one that creates incremental, serviceable, retained contracts above your contribution threshold without breaking capacity or quality. Referrals and local intent are often high quality but limited; CRM reactivation can surface known opportunity; mail and canvassing can earn new attention; paid platforms add speed. Measure your cohorts rather than borrowing a universal ranking.

How can I get roofing leads for free?

No channel is truly free. Referrals, reviews, local profiles, community relationships, old-lead follow-up, and organic content may avoid media fees, but they consume time, software, service quality, and management. Treat those resources as costs and prioritize actions that compound owned reputation and data.

How much should a roofing lead cost?

There is no responsible universal number. Maximum inquiry cost equals allowable acquisition cost per retained contract multiplied by expected inquiry-to-retained-contract yield, adjusted for internal handling cost. Use your gross profit and cohort conversion, not signed revenue or a vendor’s national average.

Are exclusive roofing leads worth it?

Sometimes. “Exclusive” must be contractually defined, source and consent should be verifiable, filters must match the operation, and the cohort still has to clear its contribution threshold. Exclusivity cannot prevent a homeowner from independently contacting other roofers.

Should a new roofing company buy leads?

Only with a small, cancellable pilot and strong intake discipline. A new company has little local conversion history, so borrowed close-rate assumptions are risky. Build accurate local profiles and relationships, instrument the funnel, preserve cash, and learn from a tightly bounded cohort.

Should roofers spend more on leads after a storm?

Only if conditions are safe, the event data is reviewed, the company has real call/inspection/production capacity, messaging is truthful, and working capital supports the workload. Visible homeowner interest is not permission to overbook or claim a specific property has damage.

Do referrals need tracking?

Yes. Record the referring customer or partner, original relationship, property, dates, reward where lawful, and downstream outcome. This reveals which relationships create appropriate work and prevents every unattributed contract from being called a referral.

How long should a channel test run?

Long enough for the predeclared cohort to mature through the relevant sales cycle, but no longer than quality, cash, or compliance guardrails allow. Set sample, budget, maturity window, and stop rules before launch. A fixed 30-day rule is inappropriate for every market and channel.

Should I hire a roofing marketing agency?

Hire an agency when its specific capability is the diagnosed constraint and the contract preserves account access, data, definitions, export, truthful claims, and outcome measurement. Do not hire one merely because leads are down while calls, inspection availability, or follow-up are broken.

Methodology and limitations

This guide is an original decision framework built from funnel economics, current first-party platform and regulator documentation, and three PorchRocket field-study owners. It does not assert a universal lead cost, conversion rate, marketing percentage, or channel ranking.

The direct-mail, CRM, and canvassing results are reported on their dedicated pages so definitions, randomization boundaries, denominators, confidence intervals, costs, and limitations remain attached to the claims. Cross-study comparisons are directional because populations, interventions, time windows, and constrained resources differ. Campaign contribution means estimated contract gross profit minus outreach and incremental inspection/sales cost; it is not revenue, cash, net income, or EBITDA.

Platform features and implementation requirements can change. Google sources were checked July 22, 2026; confirm current eligibility, billing, conversion, privacy, and API documentation inside the relevant account before purchase or implementation. FTC sources establish federal principles, not a complete state-by-state legal review. Calling, texting, recording, solicitation, licensing, referral, insurance, and privacy requirements vary. Use qualified counsel for the company’s markets and workflow.

Property, CRM, marketing, and weather signals can prioritize outreach but cannot remotely establish roof condition, damage, insurance coverage, claim validity, or a homeowner’s need. PorchRocket does not guarantee leads, contracts, claim payment, revenue, or profit. Every forecast should disclose company-specific assumptions and uncertainty.