Why Meta Ads Work One Week and Stop the Next: 8 Stages Nobody Checks

Why does a Meta ad work for one week and stop the next? A lot of the time, it didn’t stop. Something after the click stopped keeping the promise the ad made, and the ad took the blame.

I see the same pattern over and over. A coach launches an ad. The first week is good: leads come in, a few calls land on the calendar, maybe a sale or two. The next week the same ad brings about the same clicks, and almost nothing comes out the other end.

So the coach does what everyone says to do. New creative. New hook. Or more budget, because it worked once. Sometimes that’s the right move. But many times the ad was fine, and the money stopped somewhere nobody was looking.

I don’t run ads, and I say so up front. What I do is look at what happens after the click: the page, the first reply, the call, the checkout, the payment. I follow the money from the ad to the bank account and find the stage where it stops.

The short answer

1. An ad is a promise. Every stage after the click has to keep it. When one stage doesn’t, sales go up and down even when the ad is good.
2. A good week and a bad week often come from the same ad. What moved is usually further down: how fast leads get a reply, how far out calls get booked, who follows up.
3. There are eight stages to check, in order. Ad, Page, Lead, Response, Booked call, Show-up, Close, Cash.
4. Look for the stage that changed, not the one that looks lowest. A rate that has always been low is a different problem from a rate that fell last week.
5. Fix one stage before you touch the budget. More traffic into the same gap means paying more to lose people in the same spot.

1. An ad is a promise, and the funnel keeps it or breaks it

Every ad says something. “A free class on how to fill your calendar.” “Book a call and I’ll look at your offer.” The person who clicks believes it, at least for a few seconds. Then each step after the click either keeps that promise or breaks it a little.

The page has to say what the ad said. The opt-in has to lead somewhere that makes sense. The reply has to come while the person still cares. The call has to happen, and it has to be about what they came for. The payment has to go through.

That’s eight small promises, all hanging on the first one. When one of them breaks, the ad gets the blame, because the ad is the part with a dashboard full of numbers. The rest of the funnel usually has no dashboard at all.

The webinar I almost blamed on the offer

My first webinar was around 2013, I think. I ran it on GoToWebinar from a friend’s house, because his internet was better than mine. I opened with a poll so people could vote and feel part of it.

Then I forgot to close the poll. The presentation stayed black almost the whole time.

I ran it again the next day. It wasn’t the same. I sold nothing.

If I had looked only at the result, I would have blamed the offer, or the people who signed up. Neither was the problem. One stage broke, the live presentation itself, and nobody was checking it while it happened.

The problem was real. It just wasn’t where the result pointed.

2. Why a good week turns into a bad one

There are a few common reasons a funnel holds up for a week and then sags. None of them live inside the ad.

The ready buyers go first

In the first days, an ad reaches the people who were already close to buying. They need very little from the funnel. They click, they book, they pay.

When they run out, the next people need more: a faster reply, a reminder, a second touch. A funnel that never had those looks fine until the ready buyers are gone. Then it looks like the ad died.

The budget outgrows the person answering leads

More spend means more leads. If one person answers every lead by hand, reply time goes from minutes to hours. The ad keeps doing its job. The response stage stops doing its job, and nobody sees it, because reply time isn’t on any report.

The calendar fills further out

When calls book for the same day or the next, people tend to show up. When the calendar fills and the next open slot is a week away, show-up drops. The ad didn’t change. The wait did.

Sometimes it really is the ad

I want to be fair here. Ads do wear out. If the cost per click climbs and fewer people click, the first stage changed, and that’s a job for whoever runs the ads.

The point isn’t that the ad is never the problem. The point is that the ad is the only stage most people check.

3. The eight stages, one by one

For each stage: what I’d measure, what it usually means when it falls, and the question I’d ask before touching anything.

Ad

What to measure: cost per click, click rate, and which ad brings buyers, not just clicks.

If it falls: fewer people stop for the promise, or the ad is reaching a colder crowd.

The question I’d ask: Did cost per click or click rate change in the same week sales dropped, or did they hold steady?

Page

What to measure: the share of visitors who take the next step: opt in, apply, book, or buy.

If it falls: the page isn’t keeping the ad’s promise, or something on it broke. A slow load, a form that fails on phones, a button that goes nowhere.

The question I’d ask: If I read the ad and then the top of the page, do they promise the same thing in the same words?

Lead

What to measure: what happens in the first five minutes after someone opts in. Does the confirmation email arrive? Does the thank-you page say what to do next?

If it falls: leads exist on paper but never take the next step, because the next step isn’t clear or never shows up.

The question I’d ask: If I opt in right now, on my phone, what do I see in the next five minutes?

Response

What to measure: the time from opt-in to the first real reply, and the share of leads that get one at all.

If it falls: leads go cold while they wait. In the funnels I’ve worked on, this is one of the first stages to give out when the budget goes up.

The question I’d ask: Last week, how many leads waited more than an hour, and how many never heard back?

Booked call

What to measure: booked calls per lead, and how many days out they’re booked.

If it falls: the step from interest to commitment isn’t clear, or the calendar is full and people won’t wait.

The question I’d ask: What’s the next open slot today, and was it the same two weeks ago?

Show-up

What to measure: the share of booked calls, or event sign-ups, that actually show.

If it falls: reminders aren’t going out, the wait is too long, or the person forgot why they booked.

The question I’d ask: What does someone receive between booking and the call, and does any of it remind them what they asked for?

Close

What to measure: sales per call held, and what happens to the people who said “not now”.

If it falls: the call drifted away from what the ad promised, or there’s no follow-up for the maybes.

The question I’d ask: Of the people who said “not now” last month, how many heard from anyone again?

Cash

What to measure: the money that reaches the bank, next to the sales that were reported. Failed payments, refunds, payment plans that stop.

If it falls: sales look fine in one tool and money is missing in another.

The question I’d ask: Does the number of sales in the ad platform match the number of payments in the processor? If not, which one is right?

That last stage matters more than it sounds. A lot of coaches make decisions from the number the ad platform reports. That number is an estimate. The payment processor holds the number that pays the bills.

4. The eight stages on one page

Here is the same list as a table.

5. A worked example: same ad, different week

Illustrative figures. Not a client.

A coach sells a $3,000 program through a book-a-call funnel. Week one runs at $200 a day. After a good week, the budget goes to $300 a day for week two. One person answers every lead by hand.

From the ad’s side, not much happened. Cost per click went from $1.40 to $1.45. The page turned 20% of clicks into leads both weeks. Cost per lead went from $7.00 to $7.24.

Everything after that moved. With 290 leads and the same one person answering them, only half got a reply within the hour. Fewer of them booked. The calendar filled further out, so fewer showed up. Two sales instead of four.

The cash row shows one more thing. In week one, three buyers paid in full and one started a payment plan with $1,000 down. In week two, one paid in full and one started a plan. So $12,000 in sales became $10,000 in the bank, and $6,000 became $4,000.

Cost per sale went from $350 to $1,050. The coach sees “spent more, sold less” and decides the ad stopped working. A new ad would have sent the same clicks into the same slow reply.

The stage that broke was Response. The first fix isn’t a new creative. It’s getting every lead an answer within minutes, from a person or a system, before the budget goes up again.

6. How to read a drop without guessing

Compare rates, not totals

Totals go up with spend. Rates show whether a stage got worse. “We got more leads” can hide “a smaller share of them booked.”

Put two weeks side by side

Lay the good week next to the bad week, stage by stage, like the table above. The stage where the rate moved the most is where I’d look first.

Find the first stage that moved

Stages feed each other. If Response fell, Booked call and Show-up will fall after it. The first stage that moved is usually the cause. The ones after it are only following.

Check that the numbers agree

Before trusting any of it, check the same count in two tools. Say the ad platform reports 18 sales and the payment processor shows 12 (example figures). Any decision made from the 18 rests on six sales that never happened.

7. What I’d check before changing the ad or the budget

This is the short list I go through, in this order:

1. Opt in to the funnel yourself, on a phone, and time each step.
2. Pull reply times for the last two weeks.
3. Look at the next open slot on the calendar, today and two weeks ago.
4. Read every reminder a booked lead gets.
5. Ask who follows up with the people who said “not now”.
6. Match the sales in the ad platform to the payments in the processor.
7. Only then, look at the ad.

If all six of the first checks look the same as the good week, the ad is the likely change, and new creative makes sense. If one of them moved, fix that one first. Then the next dollar of budget has somewhere to go.

FAQ

Isn’t ad fatigue real?

Yes. Ads wear out, and when click cost rises and click rate falls, the first stage changed. But fatigue shows up in the ad’s own numbers. If those held steady while sales dropped, I’d look further down first.

How long should I wait before judging a week?

Long enough for the slowest stage to report. If calls book five days out, a sale from Monday’s leads may not happen until the next week. When you can, follow one week’s leads all the way to cash before you judge that week.

What if there’s no sales team, just me?

Then Response and Show-up are the stages to watch closest, because one person’s calendar sets the limit. Before the budget goes up, ask whether every lead can still get an answer within minutes. That answer matters more than the next creative.

Which stage breaks most often?

I won’t give a percentage I can’t back up. In the funnels I’ve built and looked at, the breaks I see most are slow replies, long waits for a call, and no follow-up after “not now”. It’s almost never the headline.

Do I need special software to track this?

No. A spreadsheet with eight rows and one column per week gets you most of the way. The hard part is getting honest numbers for each stage, especially Response and Cash, which most tools don’t show by default.

Should I raise the budget after a good week?

After a good week, I’d first check whether the funnel can take more of the same. Can every lead still get a fast reply? Is there room on the calendar this week? If yes, more budget makes sense. If not, more budget buys more of the same gap.

What to do next

If you want to read these numbers yourself, that’s what my workshop is for. Data Scaling Strategies shows you how to read each stage after the click, so you know where the money stops before you raise the budget.

Martin Meneses

I’ve built sales funnels since 2010. More than 100 of them, in 12+ niches.