Your ROI didn't collapse because the ads stopped working

If your Meta ad campaign started strong and then fell apart around month four or five, the agency almost certainly told you one of two things: the creative needed a refresh, or the audience had fatigued. Both explanations sound plausible. Neither one addresses what actually changes in a maturing campaign.

The more common cause is that the campaign started attracting a different kind of traffic entirely. Not homeowners who need a painter or HVAC tech. Bots, incentivized clickers, and form-fillers who have no intention of booking anything. The leads look real in Ads Manager. The numbers look worse every week on the phone.

What changes in month five that didn't change in month one

When a Meta campaign first launches, the audience pool is fresh and the algorithm is still learning. In the early weeks, the system is testing placements, audiences, and creative combinations, and the leads that come in tend to skew toward people who genuinely responded to what the ad said. The cost-per-lead is often higher in this period, but the lead quality is higher too.

By month four or five, a few things have shifted. The algorithm has optimized heavily for conversion volume, which sounds good until you realize Meta's optimization is pointed at the action you defined, not the outcome you actually want. If the campaign is optimized for form completions, it gets very good at finding people who complete forms. That group is not the same as people who answer phones, describe a real project, and book appointments.

At the same time, as a campaign runs longer and spends more, it naturally reaches further into audiences that are more loosely connected to the original targeting. This is where bot traffic and low-quality click farms start appearing in volume. The cost-per-lead often drops in this phase, which agencies sometimes present as a win. The real signal is what happens on the phone. If your close rate dropped alongside the cost-per-lead, the leads got cheaper because they got worse.

Why agencies misread this as a creative problem

The agency's dashboard shows the same campaign structure it always has. The creative is still running, the audience is still the audience they set, the budget is unchanged. From their view, if results declined, something in those variables must have changed. The obvious candidate is creative fatigue, because it's the most visible and the easiest thing to test. They update the images, rewrite the headline, maybe split-test two new versions.

This changes nothing about the underlying traffic quality problem. A fresher-looking ad still reaches the same degraded audience pool. The cost-per-lead might move slightly, but the phone quality doesn't recover, because the creative was never the issue.

The other reason agencies reach for this explanation is that it keeps them busy and keeps the engagement going. Testing new creative is work they can bill against. Diagnosing a bot lead problem requires looking at lead quality data they may not be tracking at all, because most agencies only report what Ads Manager reports, and Ads Manager does not flag individual leads as low quality. It optimizes toward the conversion event and calls it done.

What to actually look at

There are three signals that point clearly toward a bot lead problem rather than a creative problem.

First, compare your contact rate by month. Take your total leads and divide by the number of leads that actually picked up the phone, responded to a text, or engaged in any real conversation. If that rate dropped from month one to month five without a change in your follow-up process, the leads changed, not you.

Second, look at your form completion time. Legitimate leads from homeowners take some time to read an ad, think about it, and fill out the form. Bot traffic and incentivized clickers move fast. Most Meta Ads Manager dashboards don't surface this metric natively, but if your agency has access to server-side event data or a CRM with timestamps, you can often see when average form completion time drops sharply. That drop usually precedes a quality drop.

Third, look at what follow-up actually surfaces. One of the practical advantages of immediate AI-assisted follow-up, the kind Remi handles for ASN clients, is that it contacts every new lead within seconds. When a real homeowner fills out a form, that contact usually produces a response, even if it takes a few hours. When a bot or a ghost lead fills out the form, there is no response, not in seconds, not ever. Running a 30-day campaign with immediate follow-up on every lead gives you clean data on actual response rates, which is a better quality signal than anything Ads Manager will show you.

What the numbers look like when the work is done right

The Safe Step campaign gives a concrete reference point. That campaign generated 247 leads at $11 per lead on $2,800 in total spend. That cost-per-lead held because the campaign was built with niche-specific targeting from the start, not broadened progressively to juice volume. The leads stayed workable because the audience definition didn't drift.

That number, $11 per lead in a home services trade, is not a special case. It's what targeting discipline and a managed campaign look like when the goal is lead quality rather than lead volume. The campaigns that collapse from 4x to 1x over five months are almost always the ones that were optimized for volume from the beginning, often because that's what makes an agency's dashboard look good in month two.

For comparison, Kurt Welch, a painter who ended up working with ASN, had previously been locked into a contract with an agency that charged $1,000 upfront and delivered leads that were wrong for his trade entirely. He described getting leads from people looking for jobs rather than people looking for work done. That's the same mismatch as a bot lead problem, different mechanism, same result: form completions that never turn into booked jobs.

What to look for before you run another campaign

If you're evaluating whether to try Meta ads again after a campaign that degraded mid-run, the question to ask any agency is not "what would you do differently with the creative?" It's "how do you measure lead quality separately from lead volume, and what happens when those two numbers diverge?"

An agency that can't answer that question concretely isn't tracking the thing that matters. An agency that leads with creative changes as the fix to a volume-quality divergence is diagnosing the symptom rather than the cause.

ASN tracks lead quality through Remi's follow-up data, which surfaces contact rates and response patterns that Ads Manager alone won't show. If you want to see how that works against your specific trade and service area, the contact page at americanservicenetwork.com is the right next step.