Meta Ads For Residential Contractors

You will need access to Ads Manager, a clear service radius, recent job photos or videos, a defined offer, and someone assigned to review results and make changes. The owner should approve the offer and service area. The person running marketing should make the weekly campaign checks. The sales or office manager should report whether leads were qualified and whether inspections were booked. If nobody owns those jobs, the account will drift back to old habits.

The Andromeda Shift Map: Change What Meta Uses to Find Buyers

The strategy is the Andromeda Shift Map: broad local delivery, hard exclusions only, and a large set of genuinely different creative concepts. Use it when your campaigns depend on stacked interests, 1-5% lookalike audiences, manually selected placements, or one ad that has been running for months. It is also the right correction when your cost per lead rises even though your service area, offer, and sales process have not changed.

Project Andromeda does not mean “target everyone everywhere.” It means separate the limits that protect your business from the limits that merely reflect an outdated Facebook playbook. Keep the hard business limits: your actual service radius, existing customers if you do not want repeat leads, and employees. Remove unnecessary audience layers such as “roofing,” “homeowners,” “storm damage,” and “home improvement” stacked together.

A 10-20 mile service radius around the areas your crew can reliably reach is a useful starting point. A city-wide audience may include neighborhoods that are too far away, while a pile of interests may exclude homeowners who never showed interest online. Within the service area, Meta needs enough room to compare different types of responders.

The most important input is creative. Creative means the photo, video, headline, primary text, and offer shown in the ad. Under Andromeda, those elements help Meta decide which people are likely to respond. Fifteen to twenty genuinely different active ad concepts give the system more useful options than twenty small changes to the same image.

For Darren’s roofing crew, different concepts could include a close-up of failed flashing, a short inspection walkthrough, a before-and-after replacement, a homeowner explaining why the crew was easy to work with, a dated local roof-check offer, and a restoration cleanup video. These are not minor headline swaps. Each one gives Meta a different subject, message, and visual signal.

The campaign should use Leads as the objective. A lead is a person who submits contact information or a form requesting service. Use an Instant Form inside Facebook or Instagram, or send people to a dedicated website form connected to the Meta Pixel and Conversions API. The Pixel is Meta’s browser tracking code. Conversions API sends event information from your server or another connected system. Running both together gives Meta more complete feedback than relying on browser tracking alone.

Execution Steps: Rebuild the Audience Before You Judge the Ads

1. Write down the old audience rules and remove the unnecessary ones. Allow 60 minutes for this review. Record every interest, demographic filter, lookalike audience, and placement restriction currently in use. Keep only the service area and hard exclusions for existing customers or employees. The checkpoint is simple: one local audience should remain broad enough to include likely homeowners without requiring an interest label. If the audience is still narrowed by several stacked interests, the step is not complete.

2. Set the local boundary around the crew’s real working area. Allow 30 minutes. Start with a 10-20 mile service radius, then check it against actual drive times and job history. Do not choose a city-wide area just because the city name is easy to select. Darren’s office manager should compare the proposed radius with the last 90 days of completed jobs and flag neighborhoods the crew does not serve. The checkpoint is that at least 90% of new leads should come from areas the sales team can actually visit.

3. Build a creative list of 15-20 different concepts. Allow one working day for planning and two to five days for gathering material. Assign the owner or production manager to approve accurate job claims, and assign the marketing person to produce the ads. Each concept needs a different main idea, not just a new headline. Mark the list complete only when it includes multiple formats, such as vertical video, job-site photos, customer explanation, inspection education, and a concrete local offer.

4. Launch with broad delivery and Advantage+ placements. Advantage+ placements are Meta’s automatic system for showing ads across available Facebook and Instagram locations. Leave them fully open instead of restricting delivery to Feed only or Reels only. Use Advantage+ campaign budget, which lets Meta move the campaign budget between ad groups based on response. Give the campaign seven full days before making routine changes unless there is a clear policy problem or a serious lead-quality issue.

5. Connect both tracking systems and record the starting numbers. Allow two to four hours, depending on the website and lead system. Confirm that the Pixel and Conversions API are both sending lead events, then check Event Match Quality, Meta’s score for how well submitted customer information matches Meta accounts. Record the starting score, leads, qualified leads, cost per lead, and booked inspections. The person responsible for the account should save a screenshot and date it before launch.

6. Review the first 30 days in weekly blocks. The office manager should label each lead qualified or unqualified within one business day. The marketing owner should review delivery every Monday. Do not judge the new audience from one expensive day. Compare seven-day blocks, then compare the quality of leads produced by each creative concept. A concept that produces fewer leads but more booked inspections may be more valuable than a cheap source of unqualified requests.

Watch Outs: What to Stop Doing Before It Costs More

• Don’t stack interests because the audience looks more “qualified.” A roofing interest does not prove a homeowner needs a roof, and removing that interest does not make a good local homeowner irrelevant. The restriction may prevent Meta from finding people who respond to your ad.

• Don’t build the campaign around a 1-5% lookalike audience. A lookalike audience is a group Meta creates from people who resemble a source list. Project Andromeda gives more weight to the ad’s message and response patterns, so an old lookalike can add a false sense of precision.

• Don’t run one ad for months because it still gets occasional leads. A tired concept gives Meta fewer useful choices and can make performance slide slowly. Set a review date before launch and replace concepts when their seven-day cost per qualified lead rises or lead quality falls.

• Don’t restrict placements manually. Feed only or Reels only settings throw away delivery opportunities before Meta has tested them. Keep Advantage+ placements open, then judge placements by qualified lead results rather than personal preference.

• Don’t confuse a cheap click with a good audience. Clicks show interest in the ad, not buying intent. A roofing ad may attract people looking for do-it-yourself advice, insurance information, or a free inspection they never plan to schedule.

• Don’t use “Facebook ads do not work” as the diagnosis. First check whether the business boosted posts, sent traffic to a homepage, or used an old narrow audience. Those choices often explain the failure before the platform does.

• Don’t let the owner and marketer assume the other person is checking quality. The marketing person owns delivery and reporting. The office manager owns lead labels and response records. The owner owns the offer, service area, and final decision to change the campaign. Without those assignments, bad leads get blamed on targeting and targeting gets changed without evidence.

Success Metrics: Prove the Audience Shift With Business Results

Check delivery every Monday and lead quality every business day. During the first 30 days, track the following in one simple sheet:

| Measure | Checkpoint | |---|---| | Service-area rate | At least 90% of leads inside the actual service area | | Cost per lead | Compare against the local market range, not one isolated day | | Qualified-lead rate | Set a baseline in week one, then improve it week over week | | Booked-inspection rate | Review weekly with the office manager | | Active concepts | Keep 15-20 genuinely different concepts available | | Event Match Quality | Check weekly after Pixel and Conversions API launch | | Seven-day trend | Change decisions only after a full seven-day block |

For 2026 planning, cost per lead commonly falls around $8 - $20 in smaller markets, $20 - $45 in mid-size markets, and $35 - $65 in dense metro markets. These are ranges, not promises. A $12 lead that never answers the phone is worse than a $35 lead that books an inspection, so record both cost per lead and cost per booked inspection.

The first success checkpoint is not a perfect cost number. It is proof that the campaign can find local responders without the old interest stack. By day seven, confirm that delivery is broad, the ads are receiving enough impressions to compare, and tracking is working. By day 14, identify which concepts produce qualified leads. By day 30, compare booked inspections with the previous campaign and decide which concepts deserve more budget.

The practical lesson is straightforward: Project Andromeda did not remove the need for judgment. It moved your judgment to the part of the campaign customers actually see. Give Meta a real local boundary, then give it enough honest, different creative to learn from. That is how you stop paying to preserve an audience playbook that is no longer doing the work.

End of chapter one. 6 more chapters in the full book.