What to look at on day four of your Meta campaign before you pull the plug

Day four of a Meta campaign feels like the worst possible time to be patient. You've spent real money, your phone hasn't rung, and every instinct built from working with at least one agency that burned you is screaming to shut it off. That instinct makes complete sense. It's also the wrong call at day four almost every time.

Here's what's actually happening, and how to read it correctly.

What Meta is doing in the first week

When you launch a new ad campaign, Meta's algorithm doesn't immediately know who to show it to. It's running a live test across a large pool of potential viewers, collecting signals about who engages, who clicks, who fills out the form. This is the learning phase. It typically runs for the first 50 optimization events, or roughly seven days, whichever comes first.

During that window, cost-per-lead numbers are unreliable. They swing hard. You might see a $90 lead on day two and a $14 lead on day five from the same campaign, same budget, same ad. That variance isn't the campaign failing. It's the algorithm sorting itself out. Every change you make during this window, pausing the campaign, editing the budget, swapping the creative, resets the learning phase and adds days to the timeline. If you paused on day four and relaunched on day six, you didn't save money. You extended the uncertainty and spent the same amount getting less usable data.

This is the part agencies almost never explain, because if they did, more clients would stay calm and stay live instead of calling to complain.

The signals that are actually worth looking at on day four

Not everything in the day-four dashboard is noise. A few metrics tell you something real.

Relevance and engagement rate on the ad itself. If your ad has a low click-through rate (under 1% for a cold audience), the creative or the copy isn't connecting. This is worth noting, not acting on immediately. You're looking for a pattern, not a single day's reading.

Lead form completion rate. Meta will show you how many people opened your Instant Form versus how many submitted it. If a high percentage of people are clicking the ad but abandoning the form before submitting, the form itself has friction. Maybe it's asking for too much information. Maybe it's loading slowly. This is a structural problem, not a learning-phase problem, and it's worth flagging on day four because it won't fix itself.

Cost per result relative to your job economics. Do the math specific to your trade. If you're a painter and your average job is worth $1,800, a campaign that produces leads at $40-$80 per lead is likely working even if it feels slow. If leads are coming in at $200 each, the math breaks down regardless of how much time you give the algorithm. Knowing your actual job value before you launch is what makes this calculation possible. Most contractors start a campaign without doing it, then make pause decisions based on raw spend numbers rather than whether the economics are actually off.

Whether leads are opening and going nowhere. This one matters more than people realize. A lead that fills out the form and never responds to a call or text within 48 hours isn't always a bad lead. Often it's a timing problem. The contractor was on a job, saw the notification three hours later, called back, got voicemail. The lead moved on. This isn't a Meta problem. It's a follow-up speed problem, and it shows up most clearly in the first week when you're watching the results closely.

What an actual problem looks like versus learning-phase noise

Real problems on day four have a different texture than normal variance.

Your ad is being shown exclusively to the wrong audience. If you're a roofing contractor and your leads are coming from people in a completely different city, or your form submissions read like people looking for jobs rather than roofing services, that's a targeting problem. Kurt Welch, a painter who came through ASN's sales process, described exactly this from his last agency: "I was getting people looking for jobs instead of people looking for to-do work." That's a signal worth acting on, because it doesn't resolve with time. The campaign is pulling the wrong intent.

Your budget is too thin to generate a learning signal. A $10/day budget in a competitive trade like roofing or HVAC is not enough to generate 50 optimization events in a reasonable timeframe. The campaign will technically keep running, but the algorithm won't have enough data to optimize. If you're running $10/day and seeing nothing on day four, the issue may simply be that the budget is too low to produce a result in the first place, not that the campaign is failing.

Your ad creative looks generic. If the image or video on your ad could have been made for any contractor in any city by someone who's never met you, it will perform like one. Generic creative produces generic results. The clicks you do get will be lower intent because the ad didn't speak to a specific person in a specific situation. If your ad creative is stock-photo roofing imagery with no location reference, no social proof, no specifics, day four is a fine time to flag that, even if it's not a good time to blow up the campaign.

What to actually do on day four

Document what you're seeing. Leads generated, cost per lead, form completion rate, click-through rate. Write it down or screenshot it. Don't make changes yet. You're building a baseline.

Check your lead response time. If you're getting form submissions and not following up within the first few minutes, you're losing leads that the algorithm already paid to acquire. This is a fixable operational problem, not a Meta problem.

If you're running with an agency or a managed service, ask for the actual numbers. Not a reassurance. The numbers. Cost per lead, leads generated, reach, click-through rate. If the agency can't or won't show you the live dashboard, that's a problem that has nothing to do with the learning phase. That's the same pattern that burned you the last time.

If you're running with ASN, those numbers are in your own Meta Ads Manager account, which is where all the work happens. You don't have to ask for a report. You can look directly.

When to actually pause

Give a properly funded campaign seven days before drawing conclusions. If after seven days your cost per lead makes the math impossible for your trade, the audience is visibly wrong, or the form completion rate is under 20%, those are legitimate reasons to stop and diagnose. Day four is not.

The contractors who get the most out of a first Meta campaign are almost always the ones who watched carefully, changed nothing for the first week, and made decisions based on a full week of data instead of the anxiety of day four. That patience isn't blind trust. It's just how the system actually works.

If you're thinking about running Meta ads for the first time, or trying again after getting burned, the contact page at americanservicenetwork.com is the right place to start. There's no setup fee, no contract, and the numbers from your campaign live in your own account from day one.