Google Ads for home services & trades: calls, emergency intent, and booked jobs

Direct answer — how should home services run Google Ads?

Wire phone (and chat where used) as money conversions with duration floors, separate emergency from scheduled and branded intent, coordinate Local Services Ads with Search instead of double-counting, pace by metro when economics differ, and only then let Smart Bidding or Performance Max spend — always judged on cost per booked job. Home services PPC fails when forms are primary while dispatch closes on the phone, when “AC repair near me” shares a Target CPA with brand, and when LSA + Search + PMax all claim the same booked job.

This insight owns Google Ads for plumbers / HVAC / home services & trades. Measurement: multi-path lead tracking, call tracking guide, US conversion tracking. Structure: intent-based account structure. UK LSA decision: LSA vs Search. Map pack: Local SEO, US Local SEO. Vertical hub: Local services. Service page: Google Ads for local services.

Working rule: The diary is the KPI. Calls are the conversion. Emergency and scheduled are different products.

What “good” looks like after 90–120 days: Ads sees the phone jobs it was missing; emergency and scheduled CPAs are readable separately; LSA (if used) has CRM tags; metros with margin get budget; Health Score recovers; cost per booked job falls even if form CPA “looks worse.”

If you only remember three moves: track calls properly, split emergency vs scheduled vs brand, optimize to booked jobs — not form theatre.

Proof anchor: US home services Google Ads — four metros, ~$19k/month, 2.9× → 6.4× ROAS in 120 days after tracking-first rebuild.

What home services Google Ads actually buys

Plumbers, HVAC, electricians, drain, roofing-adjacent, and similar trades buy:

• Emergency intent — “AC repair near me,” “burst pipe,” same-day urgency; high CPC; short decision cycle; phone-first.
• Scheduled / maintenance intent — tune-ups, installs, quotes; longer cycle; forms and calls mix.
• Brand / navigational — people who already chose you; must not be judged on cold CPA.
• Map-adjacent demand — map pack and GBP still own much of “near me”; paid does not replace Local SEO.
• Optional LSA — screened lead product beside Search where eligible (LSA vs Search UK; US categories similar in spirit).

Aggregators (Angi-style, Checkatrade-style, HiPages-style depending on market) raise CPCs and pollute lead quality. Negatives, lander trust, and call quality scoring matter more than “more keywords.”

Out of scope as full essays: automotive package ladders (automotive insight); franchise multi-location OS (multi-location insight); Meta lead forms as a primary trades channel (usually weak for cold phone jobs).

The measurement stack (non-negotiable)

1. Call conversions with duration floors

Most US/UK trades still close on the phone. Dynamic numbers on every money page; call conversions into Google Ads as primary when duration clears spam (often 60–120 seconds depending on spam mix). Pocket dials and 8-second hang-ups are not booked jobs (call tracking, call assets).

In the US home services case, invisible phone + CRM paths were roughly 38% of attributable conversions before rebuild — Smart Bidding was blind to them.

2. Offline “job scheduled” / closed-won imports

Import CRM stages that mean money. Form submit alone is observation, not the training signal (OCI, bad conversion data).

3. Form hygiene

Keep forms secondary (or carefully dual-primary only with a written rule). Never leave pageviews or “click to call” button clicks as primary without duration.

4. Deduplication across LSA, Search, and PMax

One booked job must not become three primaries. Tag LSA in CRM; design Search call vs LSA lead rules (multi-path). Score Lead Path Coverage before raising Target CPA.

5. Capacity-aware success definition

If the van calendar is full, buying more emergency leads only burns fees and damages close rate. Media cannot outrun staffing.

Self-check: Tracking Trust, Google Ads Health Score.

Intent structure that survives high CPCs

| Layer | Examples | Treatment | | --- | --- | --- | | Brand | Brand + city | Exact/phrase; always funded; separate CPA story | | Emergency | Repair, leak, no heat, same day | Dedicated Search; call assets; tight geos; highest supervision | | Scheduled / category | Install, tune-up, replacement quote | Separate themes; landers with proof and financing if relevant | | Competitor | Competitor brand modifiers | Optional; ruthless quality review | | Research / DIY | How-to, parts, DIY | Negate aggressively |

Do not put emergency and brand in one Max Conv campaign and celebrate blended CPA. Emergency will look “expensive” while brand subsidizes the dashboard.

Multi-metro operators (Phoenix / Dallas / Denver / Tampa in the published case): split by metro (or portfolio with honest scopes) when allowable CPA differs. One national Target CPA hides winners and overfunds losers.

Local Services Ads beside Search

Where eligible, LSA is a screened lead channel — not a Search replacement. Run it when you can answer phones fast and dispute junk; keep brand Search; attribute separately; compare cost per booked job after disputes, not raw lead fees.

UK detail: LSA vs Google Ads Search. US operators: same CRM discipline — the US case coordinated LSA with Search so they stopped competing blindly for the same intent.

Landing pages and message match

Home services paid traffic is mobile-heavy and impatient.

• Click-to-call above the fold on emergency landers.
• Service-area proof per city/metro.
• Licence, insurance, and review trust where claimed in ads.
• Emergency ads must not land on a generic homepage that talks only about annual memberships.

The US case moved from ~2.4% service-page CR toward local-services medians after CTA and message-match work (message match, RSA for services, landing page gap).

Emergency vs membership copy: If the ad promises same-day repair, the first screen must confirm hours, call path, and service area — not a long brand story. Membership and financing offers belong on scheduled / replacement landers. Mixing those messages is how Quality Score and conversion rate both sag while CPCs climb.

Answer-rate SLAs and dispatch truth

Paid media for trades is an operations product. If average answer time is minutes during booked hours, you are buying leads for competitors and spam bots. Set a written SLA (e.g. answer or return within X minutes on emergency campaigns), route overflow to an answering service with CRM tags, and pause or cap themes when SLA fails. Google Ads cannot fix an empty van calendar or an unstaffed phone — it only accelerates the failure.

Performance Max for home services

PMax after:

• Call + CRM signal is clean
• Brand and emergency Search foundations exist
• City/metro asset groups (not one national black box)
• Brand exclusions / search themes reduce cannibalization

PMax before that stage chases soft forms and claims brand. Read attribution inflation, UK PMax lead quality UK, US PMax lead quality US, and PMax Readiness. PMax does not replace map pack or LSA ops.

When PMax is live, keep Search for query diagnostics and emergency control. Do not starve brand Search because PMax ROAS “looks better” — that is usually cannibalization theatre.

Negatives and spam (trades-specific)

Weekly search terms. Junk families:

• DIY / how-to / parts-only
• Jobs / careers / salary
• Wholesale / trade-only if you are retail residential
• Wrong trade (“electrician” on a plumber account)
• Out-of-area tourism when geo interest is loose
• Insurance spam and tire-kickers (duration + CRM quality)

Search terms & negatives. Typical local-services waste benchmarks in our audits run high (~32% typical, large share recoverable in ~30 days with hygiene) — the US case cut ~22% of monthly spend in the first month post-cleanup.

Shared account negatives for DIY/careers/wrong-trade; campaign-level lists for emergency themes so you do not block commercial investigation language that still books (“AC not cooling,” “water heater age”). Review AI Max / broad expansions the same week they go live (AI Max, broad + Smart Bidding).

Market notes — US, UK, AU, APAC

US: Call-heavy; aggregator competition; LSA common; multi-metro bleed. Start from US Google Ads tracking and the US home services case. Location: US Google Ads.

UK: Consent Mode, call-recording disclosures, LSA eligibility, Checkatrade-style SERPs. UK tracking, LSA compare, UK Local SEO.

Australia: High CPCs; HiPages / ServiceSeeking comparison behaviour; call-first trades. Suburb landers + licence proof (AU Google Ads).

APAC WhatsApp markets: Instrument chat if that is how jobs book (WhatsApp guide) — still keep call floors where phones close.

Sequencing playbook (90–120 days)

1. Days 1–14 — Audit. Health Score; path coverage; soft primary purge; LSA/Search double-count map.
2. Days 15–30 — Tracking rebuild. Call DNI everywhere money; duration primaries; OCI job scheduled; forms secondary.
3. Days 30–50 — Structure. Brand / emergency / scheduled / competitor by metro; LSA coordination; lander CTAs.
4. Days 50–75 — Smart Bidding on clean signal. Enable after ~7+ clean days; freeze thrash (learning phase, Smart Bidding playbook).
5. Days 75–120 — Controlled PMax / budget shifts. City asset groups; move budget to metros that book; keep weekly CRM reconcile.

Weekly ops (30–45 min): search terms; cost per booked job by metro and intent; answer-rate SLA; LSA disputes; Ads conversions vs CRM jobs.

If the weekly reconcile shows Ads “conversions” exceeding booked jobs by a wide margin, stop scaling — fix goals and spam, then resume. Scaling on fiction is how $19k months look busy and cash-poor.

Proof — US four-metro home services (full case)

• ~$19k/month across Search, LSA, PMax; dashboard ROAS ~2.9×; health 47.
• Calls not primary; polluted conversions; blended emergency/brand; weak landers (~2.4% CR).
• Rebuild: call + OCI primaries; metro + intent split; LSA coordinated; lander CTA fix; waste cleanup.
• 120 days: 6.4× ROAS (near 6.5× top-quartile local-services benchmark); +31% visible conversions; health 78; cost per booked job −34% MoM once bidding trained on revenue events; Phoenix and Dallas earned budget.

Transfer the sequence, not the screenshot ROAS.

Capacity, seasonality, and emergency spikes

Heat waves and freeze events spike emergency CPCs and lead volume. Cap spend when answer rates collapse. Strengthen scheduled capture in shoulder seasons. Do not permanently raise Target CPA during a storm week without CRM quality — you will lock in junk economics.

Common failure patterns

• Forms primary while phones book jobs.
• Emergency + brand in one campaign.
• One Target CPA across four metros with 2× different economics.
• LSA and Search double-counted.
• PMax national bucket day one.
• No duration floor — spam “conversions” teach Max Conv.
• DIY negatives never maintained.
• Judging success on form CPA while cost per booked job rises.
• Ignoring map pack while overpaying Search for the same “near me” demand.
• Raising budgets when the diary is full.

How this fits the cluster

• This insight (CP-037): home services & trades Google Ads OS.
• Automotive (CP-036): sibling vertical with package tiers instead of emergency/scheduled.
• Call / multi-path / US tracking: measurement layer.
• LSA compare: UK (and conceptual US) channel sequencing.
• Local SEO insights: map pack sibling.
• Home services & trades Google Ads (CP-037). SaaS LTV bidding (CP-038). Multi-location & franchise (CP-039).

FAQ

What is the best Google Ads strategy for plumbers?

Track calls with duration floors, separate emergency from scheduled and brand, fund brand defense, coordinate LSA if eligible, and optimize to booked jobs. Fix landers so emergency ads show click-to-call immediately.

Should HVAC companies use Performance Max?

After call + CRM tracking and Search foundations. Not as a day-one dump for four cities.

How do I track phone leads for home services ads?

Dynamic numbers, call conversions with duration thresholds, and preferably OCI for job scheduled. See the call tracking guide.

Are Local Services Ads better than Google Ads Search?

Different products. Often both when eligible — compare cost per booked job after disputes. See the LSA vs Search compare for UK detail.

What CPA should a plumber target?

Use cost per booked job by intent layer. Brand and emergency must not share one vanity CPA.

Why is my ROAS low on home services Google Ads?

Often invisible phone conversions, soft primaries, blended structure, or aggregator junk — not “Google stopped working.” The US case moved 2.9× → 6.4× mainly by fixing measurement and structure.

Do I still need Local SEO if I run ads?

Yes. Map pack owns much of near-me demand. Coordinate geos and offers so paid and organic are not blind rivals.

How long before home services PPC works?

Expect 2–4 weeks for tracking/structure stabilization, then 90–120 days for compounding like the published US window.

Should I pause ads when vans are full?

Throttle emergency or non-brand themes; keep brand. Burning leads you cannot staff damages close rate and LSA/profile performance.

How do multi-city home services accounts structure campaigns?

By metro and intent. Shift budget to cities that book. Do not hide losers in a national average.

What about AI Max for trades accounts?

Optional after clean calls and negatives — AI Max insight. Not a substitute for duration floors.

What should I show the owner weekly?

Booked jobs and cost per job by metro and emergency vs scheduled; answer rate; LSA disputes — not blended form CPA.

What to do next

1. Score Lead Path Coverage and Health Score.
2. Wire call duration primaries + OCI job scheduled; demote soft forms.
3. Split brand / emergency / scheduled by metro.
4. Coordinate LSA tags if live; fix emergency lander CTAs.
5. Enable Smart Bidding only on clean signal; add PMax only after readiness.
6. Read the US home services case.

Get a free Google Ads audit — we will map invisible phone conversions, emergency vs scheduled structure, and whether your ROAS is booked revenue or form theatre.

Home services Google Ads is a dispatch and measurement system that happens to buy Search auctions. When calls, CRM stages, and intent layers are honest, Smart Bidding becomes useful. When they are not, every recommendation card just buys the wrong job faster.