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.