Independence Network·14 July 2026·8 min read

How Long Meta Ads Take to Work: Why Week One Is the Floor (2026)

Most Meta campaigns underperform in week one, then compound. Here's the real 2-3 week timeline, why quitting early costs you, and the 71× proof.

TL;DR

New Meta ad campaigns almost always underperform in their first week. That's the floor, not the verdict. Meta starts nearly blind (the learning phase), needs conversion data and a few creative tests before it locks onto your buyers, and the appointments it books stack forward into the following weeks — so results compound. A med spa we work with near Nice told us it 'wasn't enough' on day five. By week six that same campaign had produced 193 leads at €3.21 each, 88 booked clients, and a 71× return on ad spend — and the owner was hunting for a bigger location. Judge ads on a 2-3 week arc, not day one. Turning them off early resets the learning and guarantees you only ever pay the highest price.

Five days into a new campaign, most owners have already decided it's failing.

The leads came in slower than they hoped. The calendar isn't full yet. So they do the one thing that guarantees they lose money: they panic, switch the ads off, and go back to word of mouth.

Here's what they never get to see. Week one is the worst your ads will ever perform. Not the average — the floor. And the owner who quits on day five is the one who never reaches week three, where it actually starts working.

Why do Meta ads take time to work?

When you launch a campaign, Meta knows almost nothing about who your buyers are. It's guessing. That guessing period has a name — the learning phase — and it's the reason week one looks rough.

During the learning phase, Meta shows your ad to a wide, cold spread of people and watches who responds. It needs data. Roughly 50 conversions per ad set before it settles down. Until it hits that, your cost per lead is high and your volume is low, because the algorithm is paying to learn.

This is not a sign the ads are broken. It's the price of admission. Every campaign pays it. The businesses that win are the ones who understand they're funding the learning, not the failure.

And there's a second thing owners miss: the appointments compound. A lead that comes in on day three often doesn't book until day ten. The calendar you're staring at on day five is missing everything that's already in motion. You're judging a movie by the first ninety seconds.

Week one looked like a mistake. Week six, she needed a bigger location.

A med spa we work with near Nice is the cleanest example of this we've got.

We launched her campaign and the first appointment landed 1 hour and 27 minutes after the ads went live. Fast start. But five days in, she told us flat out: "It's not enough. I don't think this is really working."

We understood the feeling. Her calendar wasn't full yet. But we could see what she couldn't — appointments were stacking up across the next two weeks, and when a launch starts like that, we already know where it ends.

So we told her to give it room.

By week six, the same campaign she almost killed had produced 193 leads at €3.21 each, 88 booked clients, and a 71× return on ad spend off roughly €620 in spend. She'd never worked that much in her life. On our next call she wasn't asking whether ads worked anymore. She said: "I think I need a bigger location. Maybe hire someone."

Her problem had flipped. Getting clients wasn't the issue. Handling all of them was. She's out scouting a bigger space right now — because she didn't turn the ads off on day five.

If you're staring at a fresh campaign wondering if it's dead, book a free audit and we'll tell you whether you're looking at a broken setup or a normal week one.

The learning phase is real — and quitting resets it

Here's the trap that costs owners the most. They launch, week one looks weak, they panic and switch the ads off. A few weeks later they try again. Week one looks weak. They switch it off again.

Every time you do that, Meta's learning phase starts over. You keep paying the highest, blindest, most expensive version of your campaign and you never once reach the part where it compounds. It's like flooring the gas, slamming the brakes, and wondering why you never get anywhere.

Patience isn't a personality trait here. It's a setting. Leave the campaign alone long enough to learn and it gets cheaper and better on its own.

Creative testing needs reps you can't rush

The first image or video you run is almost never your best one. You don't know that yet — and neither does Meta.

Real optimization means testing a few angles: a video, a plain photo, a customer story, a different hook. Running three or more formats gives the algorithm more to work with and can cut your cost per result by around 32%. But you can't test three creatives in three days. It takes reps, and reps take time.

This is why the campaign at week three beats the campaign at week one every time. Not because the market changed. Because you and Meta both learned something.

What week one really costs you if you read it wrong

| What you see in week one | What's actually happening | |---|---| | "Only a few leads" | Meta is in the learning phase, paying to find your buyers | | "The calendar isn't full" | Appointments booked now show up over the next 2-3 weeks | | "This is too expensive" | Cost per result is at its highest point it will ever be | | "It's not working, turn it off" | Turning it off resets the learning and restarts the high cost |

Read that column on the right before you touch the off switch. Almost every "the ads don't work" story is really a "we quit during the learning phase" story.

The start is the lowest it will ever be

That's the whole thing to hold onto. When you turn on a new campaign, you can only go up from there — if you let it run. The algorithm warms up. The creatives sharpen. The booked appointments stack forward. Six weeks in, you're playing a completely different game than you were on day five.

Most people think ads are a switch. On, clients, off. They're not. They're an engine that compounds — and compounding needs the one thing owners hate giving it: a little time.

Your 30-Second Audit

Three honest questions:

  1. Has your campaign actually run for a full 2-3 weeks, or are you judging it on the first few days?
  2. Have you tested at least 3 different creatives, or is one ad carrying everything?
  3. Are you counting the appointments already booked forward, or just the ones that showed up today?

If any answer was shaky, book a free audit. We'll pull your numbers and tell you whether you're in a normal learning phase or a genuinely broken setup — even if you never work with us.

The best result we've ever gotten a client started with her telling us it wasn't working. Don't quit on week one.

Frequently asked questions

How long do Meta ads take to work?

Plan on 2 to 3 weeks before a new Meta campaign shows its real performance, and about 6 weeks before it's fully optimized. The first few days are the worst it will ever look, because Meta is in its learning phase — it starts nearly blind and needs roughly 50 conversions per ad set to stabilize. During that window cost per lead is high and volume is low. As the algorithm gathers data and you test a few creatives, cost drops and booked appointments stack up. Week one is the floor, not the forecast.

Why are my Facebook ads not working after a week?

Because a week isn't enough time for Meta to learn. When a campaign launches, the algorithm has almost no data on who your buyers are, so it spends the first days guessing — expensive clicks, thin results. That's normal and temporary. It doesn't mean the ads are broken. Most owners who say 'it's not working' on day five are looking at the learning phase and mistaking it for the ceiling. The appointments booked in week one also keep showing up in weeks two and three, so early numbers understate the real return.

Should I turn off my Meta ads if they're not working in the first week?

No. Turning them off in week one is the single most expensive mistake local businesses make. Every time you kill and relaunch a campaign, Meta's learning phase resets and you pay the high, blind, week-one price all over again. You never reach the part where it compounds. Unless the setup is genuinely broken — no tracking, no offer, wrong audience — give a new campaign at least 2 to 3 weeks before you judge it, and change one thing at a time instead of pulling the plug.

Do Meta ads get cheaper over time?

Usually, yes — if you leave the campaign running and feed it good data. As Meta learns which people actually book, it stops wasting spend on the wrong clicks, so your cost per booked appointment falls over the first few weeks. Refreshing creative before it fatigues keeps that curve going. The businesses stuck at a high cost per lead are almost always the ones restarting campaigns every couple of weeks and never letting the algorithm settle.

What is the Meta ads learning phase?

The learning phase is the period right after you launch or significantly edit a campaign, when Meta is testing who to show your ad to and gathering conversion data. It typically needs around 50 conversions per ad set to exit. During this phase results are unstable and cost per result is at its highest. Once it exits, delivery smooths out and costs drop. Big edits, budget swings, or turning ads off and on restart the learning phase — which is why patience beats tinkering.

LF
Léo Ferreira · Founder, Independence Network

Aerospace engineer turned marketing entrepreneur. We run paid ad campaigns (Meta, Google, LinkedIn) for local businesses across 15+ industries. Best client result: 71× ROAS, $3.21 CPL, first appointment booked 1h27 after ads went live (Holistic Bien Être, Nice).

LinkedIn →

Independence Network

Want us to fix these for you?

We audit, fix, and manage your Meta Ads campaigns end-to-end. Book a free 20-minute call and we’ll look at your numbers together.

Book a free audit