Free YouTube Transcribe

Video transcript

Meta Ads Creative Fatigue: Why Your CPM Keeps Rising & How to Fix It (2026)

Alex Moreau | AI for Ads & SEO-GEO · 1,275 words · 6 min read

Want to search this transcript, jump the video from any line, or download it as TXT, SRT, or VTT?

Open in the transcript tool

Full transcript

Meta is starving small ad accounts

0:00Meta is quietly starving small ad

0:01accounts, not banning them, not flagging

0:04them, starving them. The algorithm feeds

0:06the accounts that ship 20 fresh

0:08creatives a week, and it buries the ones

0:10shipping two. That's not a theory,

0:12that's how the auction works. Volume

0:13wins. And 20 a week used to mean an

0:15agency, a designer, a copywriter, and a

0:18five-figure invoice every month. So,

0:20most stores ship two, pay more per click

0:22every quarter, and blame the audience.

0:24Today, I'll show you why the auction

0:25rewards volume, why handmade ads lose

0:27even when they're good, how to run

0:29creatives like an assembly line instead

0:31of a craft project, and at the end, the

0:33boring industries that are quietly

0:35winning all of this while nobody's

0:36watching. Four chapters. How the auction

0:39actually feeds, why handmade loses, the

0:42assembly line, and the boring winners.

Chapter 1: How the auction feeds — learning phase & the fatigue tax on your CPM

0:44Let's go. Chapter one, how the auction

0:47feeds. Here's the mechanic most people

0:49never see. Every ad you launch starts in

0:51a learning phase. Meta shows it to a

0:53slice of people, watches who clicks, who

0:55buys, who scrolls past, and decides how

0:58much to charge you for the next

0:59impression. Fresh creative gets a cheap

1:01slice the system wants to learn. A

1:03creative that's been running for three

1:05weeks has been learned. The audience has

1:07seen it, the click-through slides, and

1:08the price of every impression climbs to

1:10compensate. That's fatigue, and it's not

1:12a feeling, it's a number on your CPM.

1:15Now, multiply that by the account. An

1:17account with 20 new pieces a week always

1:19has something in the cheap learning

1:20slice. An account with two is paying the

1:22fatigue tax on almost everything almost

1:25all the time. Same audience, same

1:27product, same daily budget. One of them

1:29buys clicks at a discount, the other at

1:31a premium, and the gap widens every

1:33month. And there's a number behind it.

1:35An ad set leaves the learning phase

1:37after roughly 50 conversions in a week.

1:39Small accounts with two creatives almost

1:41never get there, so they live in

1:42learning limbo forever, paying learning

1:44prices without ever getting the learned

1:46discount. 20 creatives across a few ad

1:48sets get there fast because the system

1:50has enough variety to find what works,

1:52and enough spend on what works to lock

1:54it in. Meta doesn't announce this, it

Chapter 2: Why handmade ads lose — 5 steps that wait for humans

1:56just quietly hands the cheap traffic to

1:58whoever keeps feeding it. Chapter two,

2:00why handmade loses. Your ads are

2:02handmade and that's why they lose. Look

2:04at what an ad actually is inside a

2:06normal account. Somebody has an idea,

2:08somebody makes the creative, somebody

2:10sets up the test, somebody looks at the

2:12numbers and calls a verdict, somebody

2:14scales the winner. Five steps and every

2:17single one of them waits for a human to

2:18find time. The idea waits for the Monday

2:21meeting, the creative waits for the

2:23designer to finish the other client. The

2:25test waits for someone to duplicate the

2:26ad set, the verdict waits for the weekly

2:29report, the scale waits for approval. By

2:31the time the winner gets budget, it's

2:33been 3 weeks, the creative is tired, and

2:35the thing you're scaling is already

2:36fatigued. And here's the part that

2:38hurts. The handmade ad might be great.

2:41Beautiful photography, a clever hook, a

2:43headline someone was proud of, it still

2:45loses because it arrived alone 3 weeks

2:47late into an auction that rewards the

2:49account that shipped 20 okay ones this

2:51week. The product can be handmade, the

2:54pipeline that sells it can't afford to

2:55be. And there's a second problem with

2:57two ads that nobody talks about. Two ads

3:00can't tell you anything. If ad A beats

3:02ad B, was it the hook, the photo, the

3:04offer, or the format? You don't know, so

3:07you can't repeat it. Learning needs a

3:09grid. Five hooks, four formats, same

3:11product. Now when the price hook wins in

3:14vertical and loses in square, that's a

3:16fact you can build on next week. The

3:18agency version of this ships eight

3:20creatives a month for a five-figure

3:21retainer, which is roughly a thousand

3:23dollars per ad before you've spent a

Chapter 3: The assembly line — 20 ad creatives from one product

3:25cent on media. Nobody builds a grid at a

3:28thousand dollars a cell. Chapter three,

3:30the assembly line. Here's what the

3:32alternative looks like, step by step.

3:35It starts with your own assets. Product

3:37photos, past ads, the brand guide, the

3:40customer reviews, not stock, not a

3:43template, your store.

3:45From that, the agent generates a batch.

3:47Not one ad, a batch. Same product,

3:50different hooks. The price hook, the

3:53problem hook, the before and after, the

3:56review quote, the comparison.

3:58Same hook, different formats. Square for

4:01feed, vertical for reels, static,

4:03motion.

4:0520 pieces from one product in an

4:06afternoon. Each one a real photo of the

4:09thing you sell with the offer written in

4:11your voice.

4:13Then the batch goes straight into the

4:14ads workflow, not into a folder you'll

4:17get to, into a test ad set with a small

4:19budget launched the same day.

4:22The agent watches the numbers as they

4:23come in and after enough spent, it calls

4:26the verdict itself.

4:27Losers get cut. Winners get the losers

4:30budget. Nothing waits. Nothing needs a

4:33meeting.

4:34And next week, it does it again because

4:36the winners from this week are next

4:38month's fatigued ads.

4:40That's the whole trick. It isn't smarter

4:42ads. It's a pipeline that never stops

4:44feeding the auction, so the auction

4:47never stops feeding you.

4:48The algorithm gets fed, the account gets

4:51cheap clicks, and the human in the loop

4:53does exactly one thing. Approves the

4:55batch.

4:57Everything under that is already done.

4:59One rule makes the grid honest. Change

5:02one thing per pair.

5:03Same photo, two hooks. Same hook, two

5:07formats. Same format, two offers.

5:10That's how the verdicts become knowledge

5:12instead of luck. And it's exactly the

5:14kind of tedious discipline a human

5:16abandons by week three and an agent

Chapter 4: The boring winners — B2B's empty auctions

5:19never does.

5:19>> Last chapter, the boring winners.

5:22Everyone assumes this stuff is for drop

5:24shippers and skin care brands. Wrong.

5:26The quiet killers right now are B2B

5:28accounts in industries nobody puts on a

5:30mood board.

5:32Pipe fittings, commercial flooring,

5:34industrial fasteners, lab supplies.

5:37Two reasons. First, zero competition.

5:40While e-commerce brands fight each other

5:42in the auction with 20 creatives a week,

5:44a company selling valves shows up with

5:46five decent ones and owns the category

5:49because nobody else even showed up.

5:51Second, the buyer.

5:53B2B buyers research for weeks. They read

5:55the comparison, the spec sheet, the case

5:58study, and then they call. That's

6:00exactly the buyer a real photo, a real

6:02spec, and a plain headline converts and

6:05exactly the buyer a flashy template

6:07loses.

6:08So, the creatives look different, too.

6:10A spec sheet ad, the actual dimensions

6:12on the actual part. A case study ad, one

6:15customer, one number, one quote.

6:18A comparison ad, this valve versus the

6:21one they're using now.

6:23Boring on a mood board, deadly in a feed

6:25where the buyer is already looking for

6:27exactly that.

6:28So, the boring company runs the same

6:30assembly line, feeds the same auction,

6:32and gets the same cheap clicks, except

6:35the clicks are worth 20 times more each,

6:37and there's no one bidding against them.

6:39That's the whole game right now. Volume

6:42feeds the auction. The pipeline makes

6:44the volume, and the quietest industries

6:46have the emptiest auctions.

6:49If you want to build this yourself, the

6:51tools exist, and everything you need is

6:52in this video. If you'd rather the

6:54budget just shipped every week, the link

6:56is below.

6:57Either way, stop starving your account.

7:00See you in the next one.

Recently added transcripts

Browse the whole transcript library

This transcript was generated from the captions YouTube publishes for this video. Get the transcript of any YouTube video atfreeyoutubetranscribe.com, free, unlimited, no sign-up.