Facebook Ads for Home Services: The Agency Playbook

If you run an agency serving roofers, HVAC companies, or remodelers, Facebook ads are still the highest-leverage demand engine you can build for them — and the one most agencies run badly. The gap between a mediocre account and a great one isn't a secret targeting hack. It's structure, offers, creative discipline, and what happens in the five minutes after a lead comes in.

This playbook covers the full system: campaign structure for local trades, offers that convert, creative that stops the scroll, budgets by market size, instant forms versus landing pages, and how to measure and maintain performance. Written for agencies — but a sharp owner-operator can run the same play.

Why do Facebook ads still win for home services?

Facebook ads win for home services because they create demand instead of waiting for it. Search captures the small pool of homeowners actively looking; Facebook reaches the much larger pool who need a roof, furnace, or kitchen but haven't started searching — at a cost per lead search platforms rarely match.

Think of the market for any trade as three layers: a thin layer actively searching right now (Google owns them, and every contractor in town is bidding on the same expensive clicks), a thicker layer that knows the roof is old or the AC is struggling but hasn't acted, and a thickest layer not thinking about it at all.

Facebook and Instagram are where you reach that second layer profitably. A well-built ad with a believable offer converts "I should deal with that eventually" into "I'll book the free inspection now." That's demand generation, and it's why in most residential markets Meta leads cost a fraction of comparable search leads — with the trade-off that they need better follow-up (more on that in the section most agencies skip).

There are three structural reasons Meta keeps winning for local trades specifically:

• The audience is genuinely there. Homeowners aged 30–65+ remain heavy Facebook users, and Instagram fills in the younger end. This is not a platform-decline story for this demographic.
• Visual proof sells trades. A before-and-after roof, a transformed kitchen, a crew on a ladder — home services are inherently visual, and a feed is where visual proof gets seen.
• The algorithm does the heavy lifting. Meta's delivery system has gotten very good at finding likely converters inside a geographic radius, which means simple accounts now outperform over-engineered ones.

The honest caveat: Facebook leads are cheaper because they're colder. Agencies that treat them like search leads — slow follow-up, one call, then quit — conclude "Facebook leads are junk." Agencies that build the full system conclude the opposite. Same platform, different outcome.

What campaign structure works for local trades?

The structure that works for local trades is simple: one campaign per core offer, one broad ad set per campaign targeting the service area, and three to six creatives inside it. Let Meta's algorithm find the buyers; spend your complexity budget on offers and creative, not audience slicing.

This runs against the instincts of media buyers trained on e-commerce or on the Facebook of five years ago, when stacked interests and lookalike ladders were the edge. In a geo-constrained local account today, hyper-segmentation mostly fragments budget, keeps ad sets stuck in learning, and starves the algorithm of conversion volume.

The baseline account build

• Campaign level: one campaign per offer (e.g., "Roof Inspection," "AC Tune-Up," "Kitchen Consult"). Use a lead-generation or sales objective optimizing for the actual conversion — lead submitted, not link clicks. Never clicks.
• Ad set level: one broad ad set. Geography (the real service area, not a lazy 50-mile circle), age floor around 28–30 to bias toward homeowners, and otherwise open targeting. Add Advantage+ expansion where it's available.
• Ad level: 3–6 creatives per ad set, deliberately varied — different hooks, different formats, different proof. This is where testing lives now.

Add a second ad set only when you have a real reason: a separate metro that needs its own budget guarantee, or a retargeting pool once traffic volume justifies it. A warm retargeting ad set — website visitors, video viewers, engagers — showing reviews and finished-job photos is cheap insurance and usually the second thing worth building, not the tenth.

Learning phase math, in plain terms

Meta's delivery system wants a meaningful number of conversions per ad set per week before it stabilizes. If a client's budget produces 40 leads a month and you've split it across six ad sets, no ad set ever gets enough signal and everything underperforms. Consolidation isn't a style preference — it's how the machine works. Fewer ad sets, more conversions each, better optimization for everyone.

One more structural note for agencies: build each client in their own ad account with their own pixel and their own domain. Shared infrastructure feels efficient until an account gets restricted and takes five clients down with it. Portfolio risk management is part of the job.

Which offers actually convert — and what makes them believable?

The offers that convert in home services reduce risk, create urgency, or reframe cost — free inspections, seasonal tune-up pricing, and monthly-payment financing framing. What makes any of them work is believability: a specific, plausible offer from a real local company beats a bigger discount that smells fake.

Homeowners are conditioned to distrust contractor advertising. "50% off!" from a company they've never heard of reads as a markup with a haircut, so the first rule of home-services offers is that believability beats size. An offer is believable when it's specific ("free 21-point roof inspection with photo report"), bounded ("first 15 homeowners this month" — only if plausibly true), and anchored to a reason ("storm season is coming").

The three offer archetypes, and when to use each:

• Risk-reversal offers — free inspection, free estimate, free design consult. These lower the barrier to a first conversation. They work because the homeowner's real fear isn't price; it's being pressured or ripped off. Sweeten with a tangible artifact: a photo report, a drone survey, a written scope. Best for high-ticket, low-frequency trades (roofing, remodeling).
• Financing-framed offers — "new AC system from $89/month" style framing (with real, compliant financing behind it). A $12,000 furnace replacement is a crisis; a monthly payment is a decision. This single reframe often outperforms any discount for big-ticket replacements. Never invent the number — pull it from the client's actual financing partner terms.
• Urgency and seasonal offers — pre-season tune-ups, post-storm inspections, end-of-season scheduling. Urgency converts when the reason is real. "Book your AC tune-up before the first heat wave" is honest urgency; a fake countdown timer is not, and homeowners can tell.

| Trade | Primary offer | Seasonal angle | Believability anchor | | --- | --- | --- | --- | | Roofing | Free inspection with photo report | Post-storm, pre-winter | Drone photos, insurance-claim guidance | | HVAC | Tune-up special or $X/mo replacement financing | Pre-summer, pre-winter | Fixed tune-up price, real financing terms | | Remodeling | Free design consult + written quote | "Done by the holidays" timelines | Portfolio of local completed jobs | | Windows/Doors | Buy-more-save-more or financing framing | Energy-bill season | Per-window pricing transparency | | Solar/Exteriors | Savings analysis or free exterior assessment | Rate increases, rebate deadlines | Utility-bill-based math, real rebate dates |

Test offers before you test creative. A new hook on a weak offer moves cost per lead a little; a stronger offer moves it a lot, and improves lead quality at the same time. When an account stalls, the offer is the first suspect, not the audience.

What makes home-services ad creative work?

Home-services creative works when it looks like it came from a real local company, not an ad agency. Real job-site photos beat stock imagery, hook-first copy beats clever copy, and local trust signals — recognizable neighborhoods, faces of the actual crew, real reviews — do more than any production polish.

The scroll is a trust filter. A glossy stock photo of a model in a hard hat triggers the same mental ad-blocker as a billboard. A slightly imperfect photo of a real crew tearing off a real roof three miles from the viewer's house stops the thumb, because it reads as evidence rather than advertising.

The principles that hold up across trades

• Real beats polished. Job-site photos, before/afters, owner-to-camera phone video. Ask clients for a weekly photo dump from the crews; it's the most valuable asset pipeline an agency can set up.
• Hook first. The first line of copy and first second of video carry the whole ad. Lead with the problem or the proof — "Your roof is trying to tell you something" / "This Maple Grove roof took one afternoon" — never with the company name.
• Local signals everywhere. City names in the copy, recognizable streets in the footage, "veteran-owned, based in [town] since 2009." Homeowners hire neighbors, not brands.
• One idea per ad. One offer, one CTA. Ads that list six services convert on none of them.
• Mix formats deliberately. Single-image proof shots, before/after carousels, 15–30 second owner videos, and simple text-on-photo ads for feed placements; vertical cuts for Stories and Reels. Format variety gives Meta's delivery system more inventory to match to more people.

Between three and six live creatives per ad set is the sweet spot: enough variety for the algorithm to work with, few enough that each gets real spend. When you add new creative, retire the worst performer rather than letting the ad set bloat.

Creative is a deep enough topic that we've broken it out into its own guide — see what actually makes home-services ad creative convert for hook formulas, shot lists to send clients, and format-by-format specs.

How should you set targeting and budget by market size?

Set targeting broad and let budget scale with market size: roughly $1,500–$3,000/month in small markets, $3,000–$6,000 in mid-size metros, and $6,000+ in large competitive metros. The budget's real job is producing enough weekly conversions for the algorithm to optimize — spend below that threshold underperforms per dollar.

Targeting first, because it's now the short conversation. Define the geography honestly — the towns the client will actually roll a truck to — set an age floor around 28–30, then stop. Interest stacks ("homeowners" + "home improvement" + "Zillow") mostly shrink delivery and raise costs in a local radius; conversion optimization already biases toward the people who submit forms, and those are overwhelmingly homeowners with real projects.

Budget is the more consequential decision, and the honest framing for clients is a floor, not a dial. In most residential markets:

• Small markets (under ~250k population): $50–$100/day. Costs per lead are lower, but so is inventory — watch frequency, because you'll saturate the audience faster and need fresher creative.
• Mid-size metros (~250k–1M): $100–$200/day. The comfortable middle: enough audience to scale, competitive but not brutal auctions.
• Large metros (1M+): $200+/day to be more than a rounding error in the auction. Below that, delivery gets thin and results get noisy — sometimes the right advice is a smaller geographic slice of the metro rather than a thin layer over all of it.

Two budget rules worth enforcing as an agency. First, fund one offer properly before launching a second — two campaigns at half-budget each is worse than one at full budget. Second, scale in steps, not leaps: raise budgets roughly 20–30% at a time and let performance settle for several days, because doubling a budget overnight resets learning and usually spikes costs before they recover.

Seasonality is a strategy, not an excuse

Every trade has a demand curve — HVAC spikes with the first heat wave, roofing after storm season, remodeling in late winter as homeowners plan. Weak agencies let clients pause ads in the slow season and restart cold. Strong agencies budget-shift instead: heavier spend into the ramp weeks before peak season (when auctions are still cheap and homeowners are starting to think about it), lighter maintenance spend plus retargeting in the trough. The pipeline never goes to zero, and the client never restarts from a dead pixel.

Instant forms or landing pages: which should you use?

Use instant forms when the priority is volume and low cost per lead; use landing pages when the priority is lead quality and pre-sold intent. Most home-services accounts should start with instant forms plus aggressive follow-up, then graduate high-ticket offers to landing pages once creative is proven.

This is the most common fork in the road, and the trade-off is real, not mythical. Instant forms (Meta's native lead forms) convert at higher rates because the homeowner never leaves the app and the form pre-fills. That means cheaper leads — and colder ones, because a three-tap submission requires almost no commitment. Landing pages add friction: a click, a load, a scroll. Fewer leads, but each one has read your pitch, seen your reviews, and typed their own phone number. They arrive warmer and book at higher rates.

A practical decision framework:

• Start with instant forms for new accounts, tune-up/inspection offers, and any client whose economics depend on volume. Harden the form: use "higher intent" mode, add 1–2 qualifying questions (own or rent, timeline), and replace at least one pre-filled field with a typed one.
• Move to landing pages for high-ticket offers (full replacements, remodels), financing-framed campaigns that need explanation, and clients whose sales teams are drowning in unqualified volume.
• Run both in mature accounts — instant forms feeding volume at the top, a landing-page campaign feeding the sales team pre-sold appointments — and compare them on cost per booked job, not cost per lead.

Whichever you choose, the capture mechanism matters less than what happens in the next five minutes. An instant-form lead worked in two minutes outperforms a landing-page lead called tomorrow. We've written a full breakdown of the trade-offs, form-hardening tactics, and page templates in landing pages vs. instant forms for home services.

What happens after the lead? (The part most agencies skip)

After the lead is where most Facebook ad accounts are actually won or lost. Speed-to-lead decides everything: a homeowner contacted within five minutes is dramatically more likely to book than one contacted hours later. Without instant, persistent follow-up, even great ads produce "bad leads" — because nobody worked them.

Here's the uncomfortable truth: you can build a flawless account and still get fired, because the client's front desk called each lead once, at 4:45 PM, from an unknown number, and left no voicemail. Multiply by forty leads and the client's conclusion is "Facebook doesn't work" — and your invoice is the line item that gets cut.

Facebook leads are interruption leads: the homeowner tapped your ad between a cousin's vacation photos and a recipe video. Their interest is real but perishable. In most markets, contact rates fall off a cliff after the first half hour; after a day you're effectively cold-calling.

The minimum viable follow-up system

• Instant response, under five minutes, 24/7. A text within a minute of form submission — because leads come in at 9 PM and on Sundays, this realistically means automation or AI, not a human rota.
• Text-first, call-backed. Homeowners screen calls from unknown numbers; they read texts. Open with a text that references the specific offer ("Hey Sarah, saw you requested the free roof inspection — is tomorrow morning or afternoon better?"), then call.
• Persistence with a spine. A multi-day sequence — several touches across the first 48 hours, tapering over two weeks. Most booked appointments in home services come after the second contact attempt, not the first.
• Qualification and booking in the conversation. The goal of follow-up isn't "reach the lead," it's a booked, confirmed appointment on the client's calendar, with own/rent, timeline, and address captured on the way.

This is exactly the layer Peak Logic OS was built for — AI follow-up that texts every lead instantly, answers questions, qualifies, and books the appointment, so the ads you built actually get credit for the jobs they create. But whatever tooling you use, the principle stands: an agency that only delivers leads is easy to fire; an agency that delivers booked appointments is not. The full playbook — sequences, scripts, and how to sell this layer to clients — is in speed-to-lead for home services.

Diagnose before you touch the ads

When a client says "the leads are bad," audit the follow-up before the ad account. Pull ten recent leads and answer three questions: How fast was first contact? How many total attempts? Was there a text or only calls? In our experience, most "bad lead" complaints trace to follow-up gaps, not audience quality. Fixing the ads when the problem is the phones is how agencies burn months — and clients.

How do you measure what actually matters?

Measure cost per booked job, not cost per lead. Cost per lead tells you what Meta charged; cost per booked job — spend divided by appointments that actually happened — tells you whether the client is making money. Report the full funnel: lead → contacted → booked → shown → sold.

Cost per lead is the metric agencies love because it's the one the ad platform hands them, and it's the metric that gets agencies fired, because it can improve while the business gets worse. Drop CPL from $40 to $22 by loosening the form, and if booking rate falls by half, the client is paying more per job than before — while your report shows a win.

The funnel every client report should show:

• Cost per lead — fine as a diagnostic, never as the headline.
• Contact rate — what share of leads were actually reached. This exposes follow-up failures before they get blamed on targeting.
• Cost per booked appointment — the first number that reflects the whole system, ads plus follow-up.
• Show rate and cost per completed appointment — no-shows are a solvable problem (confirmation sequences), but only if you're measuring them.
• Cost per sold job and estimated ROAS — even with lag and estimation, tie spend to revenue. A roofing client who spends $4,000 and closes three jobs doesn't care what the CPL was.

Yes, close rates lag — a remodel sold in October traces to a July lead. Report what's known now and update trailing months as jobs close. An honest "here's pipeline value, here's closed value so far" beats a precise-looking CPL dashboard every time, and it changes the renewal conversation from "leads got 10% cheaper" to "we generated $86k in closed work." For the reporting templates and the attribution conversation with clients, see how to prove marketing ROI for home-services clients.

How often should you refresh creative?

Refresh creative on evidence, not a calendar: when frequency climbs past roughly 3–4 and cost per lead trends up for a week or more, the ad is fatiguing. In practice, most local accounts need one or two new creatives a month, with small markets burning through creative fastest.

Local audiences are finite: in a small market, a decent budget shows your ad to the same few thousand homeowners again and again, and even great creative wears out. The fatigue signature is consistent — frequency rising, CTR sliding, CPL climbing — over a sustained stretch. Two noisy days are weather; ten days of drift is fatigue.

A refresh cadence that works without burning the team out:

• Iterate before you reinvent. A winning ad with a new first line, a new thumbnail, or a swapped photo often buys another month. New concepts are expensive; variations are cheap. Refresh winners, replace losers.
• Keep a bench. Maintain 2–3 tested-but-unlaunched creatives per client so a fatigue dip is answered same-day, not after a two-week production cycle.
• Rotate seasonally by default. Even without fatigue signals, creative should track the season — storm-damage angles in spring, tune-up angles pre-summer, "done by the holidays" in fall. Seasonal relevance is a free CTR boost.
• Never touch everything at once. Swap creative into a working ad set incrementally. Nuking a stable ad set to relaunch "fresh" resets learning and turns a creative problem into a delivery problem.

The agencies that win long-term treat creative as a pipeline, not a project: a standing photo/video feed from each client's job sites, a monthly production rhythm, and a testing habit that always has the next ad ready before the current one dies.

FAQ

How much should a roofing company spend on Facebook ads?

Anchor the budget to job economics, not a generic number. In most residential markets, $2,000–$4,000/month is a workable starting range for a single-market roofer — enough conversion volume for Meta's optimization to function. Since one roof replacement often covers a month of spend, the real question is break-even math: if the account books even two or three jobs a month, the budget defends itself. Small markets can start lower; competitive metros usually need $5,000+ to matter in the auction.

Are Facebook leads lower quality than Google leads?

They're colder, not worse. Google leads are actively searching, so they close faster with less follow-up. Facebook leads responded to an interruption, so they need fast, persistent follow-up to convert — but they typically cost far less, and with a real speed-to-lead system the cost per booked job is often comparable or better. The mature answer is both channels: search to capture existing demand, Facebook to create it at scale.

How long before Facebook ads work for a home-services client?

Expect leads within days and a fair verdict in 60–90 days. The first 2–4 weeks are learning: the algorithm calibrating, you reading offer and creative signals. Weeks 4–8 are iteration. Judging by booked jobs takes longer because sales cycles lag — a week-two lead may become a week-eight sale. Set that expectation with clients before launch; accounts killed at day 21 die of impatience more often than strategy.

Should agencies use Advantage+ or manual campaigns for local trades?

Lean into automation for delivery, keep manual control where local knowledge matters: geography, offer, and creative. Broad targeting with Advantage+ expansion generally outperforms hand-built interest stacks in a local radius. But don't hand the algorithm decisions it can't reason about — service-area boundaries, seasonal offers, which crew photos to run. Automation is a delivery engine, not a strategist.

What's a good cost per lead for home-services Facebook ads?

It varies too much by trade, market, and offer for a universal number — instant-form tune-up leads can run quite cheap, while landing-page leads for full remodels cost multiples more. The better question is cost per booked job against the client's average job value and close rate. A $60 lead that books is a bargain; a $15 lead nobody answers is pure waste. Benchmark each account against its own trailing performance and its own margins.

Can a home-services company run Facebook ads without an agency?

Yes — the platform keeps getting simpler, and an owner with good job photos and a real offer can get leads flowing. Where DIY usually breaks is everything around the ads: follow-up speed, creative refresh discipline, measurement past cost per lead, and time. Most owners who succeed either dedicate real weekly hours to it or pair simple self-run campaigns with software that handles follow-up and booking automatically.