Which Meta performance goal is actually working for local home service lead gen right now

If your last agency ran Meta ads for your contracting business and got you a pile of leads that went nowhere, there's a decent chance the campaign was optimizing for the wrong thing from day one. Not the wrong audience. Not the wrong creative. The wrong performance goal, set at the campaign level, before a single dollar was spent.

This matters more now than it did two years ago, because Meta's ad delivery system changed significantly with Andromeda, an internal ranking model update that shifted how Meta decides which users to show your ads to. The short version: Meta used to rely more heavily on your targeting inputs to find your audience. Now it leans harder on the performance goal you set, using it as a signal to figure out who to find. That shift changed which goal setups actually produce qualified leads for local service businesses, and it made a common lazy default (optimizing for link clicks or landing page views) substantially worse.

What the performance goal actually controls

Most contractors who've run ads, or paid someone to run them, think of the campaign goal as a formality. You pick "leads" because you want leads, click next, and move on. The reality is that the goal setting tells Meta's delivery system what kind of person to seek out across its entire user base.

Set your goal to link clicks, and Meta optimizes for people who click things. That population is wide, cheap to reach, and largely useless for a local service business. Your cost per click looks great. Your cost per actual phone call or booked job looks like a different spreadsheet entirely.

Set your goal to leads via Instant Form, and Meta optimizes for people who have historically submitted lead forms. That's a more qualified pool, but Instant Forms have a completion problem: they're frictionless enough that people fill them out and forget they did. You've probably seen this. You call within a few hours, the person has no memory of leaving their information, and they're annoyed you called.

Set your goal to conversions, tied to a real event on a landing page or a confirmed calendar booking, and Meta optimizes for people most likely to complete that specific action. This is harder to set up. It requires a landing page, proper pixel installation, and a defined conversion event. It's also the configuration that produces the most qualified lead pool, because the threshold someone has to cross to trigger your conversion event is higher than clicking or form-filling.

What Andromeda changed, specifically

Before Andromeda, Meta's targeting inputs did more of the heavy lifting. You could set a tight geographic radius, layer in homeowner behaviors or income brackets, and get reasonably close to your actual buyer without a deep conversion signal history.

Andromeda shifted the system toward what Meta calls "broad learning," meaning Meta's model now wants to find its own patterns rather than be constrained by advertiser inputs. Tight audience targeting started underperforming. Broad audiences with a strong performance signal started outperforming. The performance signal Meta relies on most is the conversion event you've defined.

For a local service business, this creates a specific problem: your audience is geographically constrained. A roofing company in Calgary or a junk removal operation in Phoenix has a hard radius. You can't just open the audience to all of North America and let Meta optimize freely. What you can do is give Meta the clearest possible signal about what a real conversion looks like, so that within your geographic constraints, it's finding the people most likely to actually book.

That means a confirmed landing page form submission beats an Instant Form submission. A calendar booking confirmation beats a landing page form submission. The more downstream and intentional the action, the better the signal quality, and the better Meta's model performs within the limited pool it's working from.

The lead quality gap in real numbers

The Safe Step campaign, a rubber resurfacing client ASN managed, ran with a proper conversion setup: landing page, defined conversion event, optimized for genuine form completions. The result was 247 leads at $11 cost per lead on $2,800 in total spend. That CPL held because the optimization target was qualified from the start. The system was finding people who completed a real form, not people who clicked and bounced.

Compare that to a clicks-optimized campaign on the same budget. The click volume would be higher. The CPL, measured at the lead stage, would look worse, because half or more of those clicks never become any kind of lead at all. And the leads that do come through are a less filtered group than what a conversion-optimized campaign attracts.

This gap compounds over time. Meta's algorithm learns from the conversion events it records. If your campaign records 50 real form submissions in the first few weeks, the system has data to work from. If it records 500 clicks, most of which were accidental or uninterested, the system is learning from noise.

What to look for if you're evaluating an agency

If you're considering paying another agency to run Meta ads after getting burned before, the performance goal setup is one of the fastest ways to separate serious operators from people who will take your money and run a lazy campaign.

Ask them directly: what conversion event are you optimizing for, and where does that event fire? If the answer is "link clicks" or "traffic," that's a signal. If the answer is "Instant Form submissions" with no mention of what happens after, ask what their average contact rate is on those leads. If they can't answer that, they're not tracking the thing that matters.

The agencies that produce real results have a landing page, a pixel firing on a meaningful action, and ideally a follow-up system that contacts leads fast enough to catch them while they're still thinking about the job. For what it's worth, ASN runs all three: conversion-optimized campaigns, landing pages included in the standard service, and Remi, an AI follow-up tool that responds to new leads within seconds and books them onto your calendar before they forget they filled out anything.

What to do next

If you want to see how a properly configured campaign actually runs for a trade like yours, with real numbers and a setup you can look at before you commit to anything, that's what the first conversation with ASN covers. No setup fee, no contract. The contact page is the right place to start.