Google Ads and Meta Ads do different jobs, and most contractor accounts fail by assigning them the same one. Google captures demand: someone’s AC died, they searched, you showed up. Meta creates demand: nobody was searching, but your offer, your crew, and your reviews planted the decision for later, or surfaced the project a homeowner kept postponing. Run them by those roles and they compound. Run them interchangeably and you buy expensive confusion on two platforms at once.
This is the playbook: what each platform is for, how to structure them for the trades, how to split budget, and the measurement standard that keeps all of it honest.
Before any of it: the measurement decision
Every structure decision downstream depends on what the platforms are optimizing toward. If your tracking reports form fills, both platforms will get very good at buying form fills, and no campaign structure fixes that. The prerequisite is tracking that feeds real outcomes back: leads measured correctly, calls tracked with outcomes, closed-job revenue routed home from your CRM. That build is its own complete guide; this playbook assumes it, and everything below works dramatically better with it.
Google Ads: capture the demand that already exists
Search is the core. For the trades, Search campaigns aimed at service intent are the highest-converting paid inventory on the internet. Structure them around your economics, not around keywords alphabetically:
- Split emergency from scheduled. “Burst pipe” and “water heater replacement cost” are different customers, different close rates, different values, and they deserve different bids, ads, and landing pages.
- Split service from install. A $300 repair and a $12,000 replacement should never share a budget, or the cheap conversions will eat it. This is the single most common structural mistake in contractor accounts.
- Run a brand campaign. Cheap insurance that your own name, and the customers your other marketing created, do not leak to a competitor bidding on you.
- Mind the negatives. DIY searches, parts, careers, and rental intent burn real money in every unmanaged trades account. The negative keyword list is unglamorous and pays for itself monthly.
After-hours is a strategy, not an accident. Emergency intent does not keep business hours. Call-focused ads with tracked numbers during nights and weekends, if your team actually answers, is some of the best money in the account.
Local Services Ads belong alongside Search, not instead of it: pay-per-lead, top-of-page, driven by reviews and responsiveness, with far less control. Run both where the economics support it.
Meta Ads: create the demand for later
Meta will not out-convert Search on emergency intent, and it should not try. What it does for contractors:
- Fills the shoulder seasons. Maintenance plans, tune-up offers, financing-led replacement campaigns when nobody is searching yet.
- Builds the name that makes Search cheaper. Homeowners click the roofer they have seen before. Brand familiarity shows up as higher click-through and close rates everywhere else.
- Retargets the almost-ready. Estimate viewed, form abandoned, service page visited: modest budgets here consistently punch above their weight.
Creative rules the results. Real crews, real trucks, real before-and-afters, and real reviews outperform polished stock imagery in the trades, reliably. The ad that looks like your actual company is the ad that gets trusted.
Lead forms versus landing pages is a real trade-off. In-platform lead forms cost less per lead and produce lower intent; landing pages cost more and qualify harder. Choose based on your follow-up speed: fast, disciplined follow-up can profitably work cheap leads, slow follow-up cannot.
Not sure which half of this your account gets wrong? The free audit covers both platforms and the tracking underneath them. Request a Free Audit. Five business days. No cost. No commitment.
How to split the budget
Sequence, not formula:
- Fund demand capture first. Search (plus LSA) until you are reliably visible for your core service intent in your service area. This is the money that returns fastest and proves itself most clearly.
- Add creation once capture is saturated. When pushing more budget into Search stops adding booked jobs, the next dollar belongs to Meta, seasonal offers and retargeting first.
- Rebalance seasonally. Capture-heavy in peak demand, creation-heavy in the run-up. The account that plants demand in April owns July.
For what the total should be, work backwards from capacity and job economics; we wrote the budget math up separately.
The failure patterns to avoid
- One campaign for everything. The averages hide the economics; segmentation is the whole game.
- Judging both platforms by the same metric. Meta will always lose a cost-per-lead contest with Search and often win the quarter anyway. Judge capture on cost per booked job; judge creation on what it does to the whole system’s volume and pricing.
- Set-and-forget, or its twin, constant meddling. Automated bidding needs stability to learn and honest data to learn from. Change one significant thing at a time, then let it settle.
- Chasing cheap leads across both platforms. The cheapest lead in home services is usually the least valuable one. Without revenue in the account, cheap is all the algorithm can chase.
The measurement standard
Same standard for both platforms: booked jobs and closed revenue per campaign, from your CRM, not platform-reported conversions. Platform numbers inflate and leak in opposite directions; the CRM is the referee, and on the Meta side its verdicts feed back through the Conversions API. If the reporting cannot show which campaigns produce revenue, structure changes are guesses with better vocabulary; the fix is a Performance Dashboard, not just GA4.
Frequently asked questions
We can only fund one platform. Which one?
Google. Demand capture converts existing intent, returns fastest, and proves the math for everything after. Add Meta when capture is funded and saturated, not before.
Should both platforms run year-round?
Capture, yes: emergencies do not take quarters off, and neither should your visibility. Creation flexes: heavier before and during your season, lighter but rarely zero in the trough, because the pipeline you plant in the quiet months is the one you harvest later.
What about Performance Max?
Fed with real conversion values, it can genuinely extend reach; fed with form fills, it sprays budget across inventory you cannot inspect. It is an amplifier of your measurement quality in both directions. Fix the data before you hand the platform that much autonomy.
How long before a restructured account performs?
Weeks, not days: automated bidding relearns after significant structural change, and the relearning is the cost of getting the structure right. Restructure once, properly, with the measurement in place, rather than quarterly by vibes.
The platforms are not mysterious; they are literal. They buy more of whatever you prove you want. Structure tells them where to hunt, and measurement tells them what a trophy looks like. If you want both graded on your account, request the free audit: a 20-minute discovery call, then five business days to a quantified answer.