YT YouTube2Text

YouTube2Text → Transcripts → The NEW BEST Way To Scale Facebook Ads

The NEW BEST Way To Scale Facebook Ads — Transcript

by Professor Charley T · 5,143 words · 733 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Every scaling strategy I've ever seen on
0:01YouTube is built around one idea. Find
0:03what's working and spend more on it. In
0:05this video, I'm going to show you why
0:06that single idea is the reason your ad
0:09account keeps hitting the same ceiling
0:10and the system that'll get you past it
0:12for good. So, here's what's actually
0:14happening. You find something that
0:16works. The numbers look good, so you
0:18raise the budget. For a few days, it
0:20looks fine. Then CPA starts to creep up
0:24and rorowaz goes down. So, you get
0:27nervous. You pull back the budget and
0:29before you know it, you're right back
0:30where you started. This isn't an
0:34algorithm problem. It's a structural
0:36one. Because the moment you increase the
0:38budget, you're adding load to the
0:40machine. You're asking your ads to do
0:43more work, reach more people, find more
0:45conversions. As spend goes up, more of
0:48that money moves higher up in the funnel
0:50because that's how you find more people
0:52to convert. Temporary loss of efficiency
0:55isn't a malfunction. It's the machine
0:58doing exactly the job you asked it to
1:01do. Scale. You can actually watch this
1:04happen in your own account right now.
1:06Look over the last 30 days. Break it
1:07down by week and find the week right
1:09after a budget increase. Almost every
1:12time you're going to see a drop in
1:14performance followed by a recovery if
1:16you gave it the time to recover. You'll
1:19probably also notice an impact on CPM
1:22and frequency. That's the machine
1:25learning how to do more work. But most
1:28people see the cost go up, immediately
1:30assume something's broken, and pull back
1:32the budget right away. Most of the time,
1:35that's just not true. You didn't build
1:38leverage first. So, you broke what was
1:40winning and had to rebuild from scratch
1:43every time. That's the merrygoround of
1:46pain that was good enough to get you to
1:49where you are, but will never get you to
1:51where you want to go. And across more
1:53than a billion dollars in ad spend, my
1:55own brands, my clients, my students,
1:57this is the single most common reason
1:59accounts stop growing. It's not the
2:02algorithm, it's not the economy, it's
2:04not the creative, it's this. You end up
2:08day trading the budget up, down, up,
2:11down, working harder and harder every
2:13single month just to hold the same line.
2:17Now, I know what you're thinking.
2:20Doesn't every account eventually hit a
2:23ceiling? And the answer is no. Not in
2:26the way that you're imagining it. The
2:28ceiling that you've never been able to
2:29get past isn't a limit on the business.
2:32It's a limit created by the way the
2:34moves in the ad account are being made.
2:36So finding winners and scaling them
2:38doesn't work. And having a more and more
2:40complicated account where you fight the
2:42machine harder and harder doesn't work.
2:44That means the fix isn't a bigger
2:46budget. It isn't a better ad. We're not
2:48looking for winning ads. That's not what
2:50it's about. The fix is to make sure you
2:52always have leverage at every move to
2:55make the next one. Now, before I show
2:58you the whole system, I need to show you
3:00the one number that decides whether
3:02you're allowed to make a move at all.
3:04Because once you see it, everything else
3:06in this video clicks.
3:11Here's the number. Let's say your target
3:13CPA is $50. That's the most you can pay
3:16for a sale and still get the business
3:17results that you need. And let's say
3:20your actual trailing 7-day CPA is 45.
3:24Most advertisers look at that and say,
3:26"Great, I'm 10% better than my target.
3:28Let's raise the budget." But the
3:30important thing is not that 45 is a good
3:33number. The important thing is the gap.
3:37That $5 between where you are and where
3:40you're not allowed to go anymore. That
3:42is headroom. It's the only thing that
3:45makes scaling more survivable because
3:47the load you're about to add has to land
3:49somewhere. And notice that headroom is a
3:52gap, not a score. $45 is not good news
3:55by itself. It's not an inherently good
3:58CPA. $45 against a $46 target has almost
4:01no headroom at all. Same CPA, still
4:03profitable, completely different
4:06decision. Which is why the goal was
4:08never to hit your target. The goal is to
4:12always have headroom. If you decide to
4:16scale when you hit your target, you're
4:18already at the ceiling. And the normal
4:21temporary completely expected efficiency
4:24cost of adding volume now looks like
4:26proof that scaling doesn't work for you.
4:29But it isn't proof of that at all. You
4:31just had nowhere to go. This is the
4:33single most crippling self-imposed
4:36limiting belief because the conclusion
4:38people take away from it is that their
4:40account has some arbitrary hard limit.
4:42It doesn't. It just had no leverage.
4:45Scaling an account isn't about finding
4:47the right strategy to maintain a CPA
4:50while you increase the budget. What your
4:53account needs is leverage. And the way
4:56you give it that is by keeping things
4:59simple. A small number of moves, each
5:02doing a specific job so that your
5:05account always has headroom instead of
5:08constantly scaling until it breaks and
5:10rebuilding over and over again. And
5:12that's what the system I'm going to show
5:14you today was built to do. And it starts
5:17with just three moves.
5:24So if the problem is complexity, abusing
5:26leverage, and destabilizing the account,
5:28the solution has to be the exact
5:31opposite. Simple, structured. Every move
5:34has one job. This is infinite scaling.
5:39Three moves, three jobs. always run in
5:43the same sequence. Move number one is
5:46the budget move. If you have the
5:48headroom to increase the budget, just do
5:51that. If you're winning, keep winning.
5:54Don't over complicate it. The worst
5:57thing you can do when you've built a
5:59system that can handle more load is
6:02change the system. Everything you do in
6:05your account is getting you to one
6:07question. Can I spend more money
6:09tomorrow? And if the answer is yes, do
6:11it. We'll get to how to automate that
6:13process in just a little bit, but
6:14understand that everything comes down to
6:18this step. If you don't have the
6:19headroom to increase the budget, the
6:21next move is addition by subtraction.
6:24It's extremely common to look at an ad
6:26account and say, "This isn't working.
6:28Let me launch a bunch of more ads."
6:30That's the exact opposite of what you
6:33should be doing. First, you need to
6:35understand what is consuming the
6:38headroom. Any ad spending 10% or more of
6:41the budget when removed will shift how
6:44the campaign allocates budget to each
6:47ad. Any ad spending more than 40% of
6:50your budget when removed will completely
6:52change how everything works. Find the
6:55highest spending ad below that 40%
6:57threshold that isn't doing its job well
6:59enough and turn it off. Removing that
7:02liability is the simplest, safest, and
7:04highest confidence move to get you back
7:07to a yes on move number one. If no ad
7:10fits this profile, instead of removing
7:12an ad, we have to move to step three.
7:15This is where creative testing fits into
7:17the strategy. The goal of that test is
7:19simple. Earn spend and create leverage.
7:23We're not trying to find winners. We're
7:26trying to improve the team. Facebook ads
7:30is a team sport. Using a 4PI analysis to
7:34understand why you can't say yes to move
7:36number one and what the ads are doing in
7:38move number two that prevents you from
7:40removing any of them tells you what your
7:42problem is. Using a 322 ad gives you the
7:45best possible tool to solve that
7:48problem. And you will only ever have one
7:50of two problems. And the solution to
7:53either one is simple. Either you can't
7:56scale without your best ads getting so
7:58bad that you lose all of your headroom.
8:00So you need to scale the opportunity
8:02with an upperfunnel ad or you don't have
8:06any headroom at all. So you need to
8:08scale the efficiency with a lowerfunnel
8:10ad. And a win is simply defined as this.
8:13The test got spent and the campaign is
8:15now good enough to say yes to step one.
8:18We can now increase the budget. And I'll
8:20show you how to run a 4PI analysis using
8:22a real ad account in just a little bit.
8:24But first, let's break down move number
8:26one. I'm going to show you how to scale
8:28your spend and even how to automate it
8:30so you never have to touch the budget
8:32ever again.
8:37Increasing the budget is the first move
8:39of infinite scaling. And there's one
8:41golden rule that governs how we do it.
8:44We never leverage all of the opportunity
8:46at once. Your job is to add load in
8:49small enough increments that if you got
8:50no extra conversions, you'd still be
8:52happy with the result. The campaign's
8:54job is to distribute that additional
8:56load to learn and then produce enough
8:58leverage that earns you the right to do
9:00it again. That's the idea. You scale the
9:02budget, the machine adapts, you regain
9:04leverage, and you scale again. If you
9:06never run out of leverage, you can
9:08always scale. And because smaller
9:11increases are more stable, we're less
9:13likely to break what's already working.
9:16My best practice is to never increase by
9:18more than half of your headroom. And
9:20small changes made over and over again
9:22can fundamentally change the landscape
9:24of your business while providing extreme
9:26clarity and making the rest of the work
9:29very simple. So now I'm going to show
9:31you the three different ways to scale
9:33and how to automate each of them so that
9:35you never have to touch budgets ever
9:37again. Strategy number one is called
9:39linear scaling. It's a small controlled
9:42increase on a schedule that becomes a
9:43smaller and smaller percentage change
9:46over time. This could be as simple as
9:48adding $10 a day. That's the whole move.
9:52And here's what it looks like in Ads
9:54Manager. Now, I understand $10 doesn't
9:57seem like a lot, but $10 a day for a
10:00year means you're spending over $3,600
10:03more every single day one year from now.
10:06And then go crazy. Add $20 a day and
10:09you'll be adding over $7,000 a day to
10:11your daily spend, putting you at over
10:13$10,000 total on a daily basis. By
10:16simply adding $10 a day, everything gets
10:18so much easier because now all you have
10:20to do is make sure you never have to
10:21turn that rule off. The entire business
10:23becomes as simple as accomplishing one
10:25very specific goal. Really simple. No,
10:28not easy, but very simple. Practically
10:31speaking, if you're currently spending
10:32$50 a day, add $2 a day and you'll
10:35double your budget in less than a month.
10:36up it to $3 a day. And when you're at a
10:38hundred, you've doubled it again in just
10:41over a month. When you're at 200, add $5
10:44a day. And you've doubled again in just
10:46over one more month. Now add $10 a day.
10:50And by this time next year, your budget
10:51is close to $4,000 a day. In 12 months,
10:54you have 80xed your daily spend with one
10:58simple rule. Where linear scaling is
11:00slow, safe, and the most stable way to
11:03scale. The counterpoint is fractional
11:05scaling. Linear is a marathon.
11:08Fractional is a sprint. The mechanics
11:11are the same. A small controlled
11:13increase on a schedule, but it's the
11:14same percentage change every time.
11:17Fractional scaling leans on the
11:19principle of compounding interest. And
11:21here's what that rule looks like. Now, I
11:24want to take you on a little journey. If
11:25you took a chessboard and put one grain
11:27of rice on the first square and then
11:28doubled it every single day, there's a
11:31point where you can really no longer see
11:33in your imagination what's happening.
11:35One becomes two, two becomes four, four
11:38becomes eight. And I would imagine
11:40everybody watching this video would
11:41correctly assume that by the time you
11:44get to the 64th square, you would cover
11:46the board entirely. But that's not even
11:50close to how much you'd really have. You
11:54wouldn't just have enough rice to cover
11:56the board. You would have enough rice to
11:59cover the entire surface of Earth under
12:02a layer 10 ft deep. Now, obviously,
12:07doubling your budget every day is
12:09completely unrealistic. But what about
12:111%.
12:131% a day would take 70 days to double
12:17your budget. And interestingly enough,
12:19it would [snorts] take another 70 days
12:21to double it again. There's a thing
12:24called the rule of 70. It goes back to
12:27Luca Pacioli, and forgive me if my
12:29Italian isn't any good, who documented
12:31the rule of 70 in a book published in
12:33Venice in 1494 called Suma de
12:36Arithmetica. Again, forgive my Italian.
12:39So, what happens at the end of a year,
12:411% a day, you end up spending 37.8 times
12:45as much, which is almost identical to
12:49what we just showed with linear scaling.
12:51Adding $10 a day starting at $100 leaves
12:54you with a budget of $ 37.50. Adding 1%
12:58a day starting at $100 leaves you at
13:013780.
13:02Kind of wild, right? Both are totally
13:04acceptable and both are completely
13:07doable, but that's also the least
13:09aggressive version of fractional
13:12scaling. If you were to move it to 2%
13:14instead of one, you don't double every
13:1670 days, you double every 35. And by the
13:20end of the year, your daily budget would
13:23be over $137,000.
13:27Fractional scaling at anything over 1%
13:29is completely unsustainable for most
13:31accounts. But you reach the upper
13:33threshold of your leverage much faster.
13:36And that is the goal. Get to the ceiling
13:39as quickly as possible and then change
13:41the strategy. Either you drop back down
13:44to linear scaling or you move to
13:46strategy three which is called marginal
13:50scaling. Linear and fractional scaling
13:52are both built around one principle. You
13:54keep the rule on until you have to turn
13:56it off. Marginal scaling is different.
13:58It's conditional. It's always on, but it
14:00only fires when the headroom allows it.
14:03Going back to our example target CPA of
14:06$50. If we want to increase the budget
14:08by 5%, we need to be beating that $50
14:11target by 10%. From the rule of never
14:14scaling by more than half, that means
14:17anytime we're at $45 or cheaper on a
14:19trailing 7-day average, we're allowed to
14:22scale the budget by 5%.
14:25And here's what that rule looks like
14:26when you build it. One important
14:29stipulation with marginal scaling, we do
14:32not let it fire every day. We run it two
14:35or three times a week. I like Monday,
14:38Wednesday, Friday. This way, we don't
14:39get the runaway compounding we just
14:41talked about with fractional scaling.
14:43And you always protect your leverage.
14:46Linear scaling makes the entire business
14:49about making sure the rule never gets
14:51turned off. Fractional scaling is
14:53basically seeing how long it runs before
14:54you have to turn it off because nobody
14:56could ever let that run forever.
14:58Marginal scaling is built around a
15:00simpler idea. Whenever I have leverage,
15:02just take the opportunity. it becomes
15:04passive. You essentially can never scale
15:07beyond your headroom. Generally
15:08speaking, this is where most accounts
15:10end up. It grows faster than linear
15:13scaling, but not as dangerously as
15:15fractional. However, because it's based
15:17on a trailing 7-day average, and it
15:19fires a few times a week, we need that
15:217-day average to be fairly stable, and
15:24that requires a meaningful volume of
15:26transactions to smooth out the peaks and
15:29valleys so performance is less
15:30hyperbolic from day to day. Marginal
15:33scaling rewards you when you're doing
15:35well. And when it doesn't fire, it's
15:36telling you that you need to go to move
15:38number two or move number three to put
15:41the account back in a position where
15:43it's eligible to fire again. And that
15:45makes the business really simple. And
15:46that's why I love these three ways of
15:49scaling. They're all completely
15:50automated. They just never push you past
15:53the point of leverage. And they give you
15:55one job. Keep the system in a position
15:57where the rule is allowed to run.
16:03So now that you know how to scale, what
16:06happens when you can't? And that brings
16:08us to move number two and move number
16:10three. Both of them are built around the
16:134PI. 4PI is the four performance
16:16indicators that tells us how the machine
16:18is using our ads. Here's what the
16:20dashboard looks like. And I'm going to
16:22walk you through each column. And I'm
16:24actually going to show you my 4PI plus
16:26two because it's what I use for
16:27basically every e-commerce business
16:29right now. I've been teaching in the
16:30MetaMBA for years. Everybody in the
16:32academy has it and so I'm going to show
16:33it to you also. Spend tells you how much
16:35the machine is investing in that ad.
16:37Spend is a very good indicator of what
16:38is contributing to the bottom line and
16:40what the machine actually wants to spend
16:42on. Frequency shows us where in the
16:45funnel that ad is. I don't care about
16:47frequency over weeks or months. We need
16:49to look at the buy breakdown. The higher
16:53the frequency number, the lower that ad
16:55sits in the funnel. A frequency of 1.05
16:5805 means roughly 5% of people saw it
17:00twice. A frequency of 1.95 means 95% of
17:03people are seeing that ad twice. That's
17:05heavily investing in retargeting. Now, I
17:07want you to reject the idea that
17:09middlefunnel exists. That's old thinking
17:11and honestly, it doesn't make any sense.
17:14What you have is a baseline, the overall
17:16performance of your campaign. If your
17:18campaign is sitting at 1.23 on a daily
17:20basis, any ad with a daily frequency
17:22lower than 1.23 is higher in the funnel.
17:25Any ad with a daily frequency higher
17:26than 1.23 23 is lower in the funnel
17:29relative to your baseline. That's it.
17:32CPM tells us the price of the inventory.
17:35Really, this is quality of experience.
17:38If an ad is appealing to complete
17:40strangers, the CPM tends to be lower.
17:41That's why ads that earn a ton of spend
17:43and always have a low frequency also
17:45tend to have a low CPM. The machine
17:47wants to show it to new people, and
17:48you're getting rewarded because it gives
17:50people a positive experience. Ads that
17:52can't spend as much and have a higher
17:54frequency are retargeting ads. there are
17:56fewer people to spend on. The frequency
17:58is higher and the user experience
18:01suffers. It's a lowerfunnel ad. If
18:03somebody ignores an ad three times a
18:05week, that is objectively a worse
18:06experience than complete strangers
18:08liking, commenting, and watching.
18:10Inherently, the CPM will be higher. Then
18:13we have efficiency. If an ad has a
18:15frequency below average and a CPM below
18:18average, we expect the cost per result
18:20to be a little bit worse than average
18:22because it's filling the funnel. Any ad
18:24with a frequency above average and a CPM
18:26above average needs to have a CPA that's
18:28better than average. Otherwise, it's a
18:30bad retargeting ad. Lowerfunnel ads
18:33should be more efficient on the sale.
18:35You're hammering warm traffic more
18:37often. You're sacrificing reaching new
18:39people to make sales from the warm
18:41audience you already have. If you can't
18:44close them nearly as well as complete
18:45strangers, it's a complete waste of
18:47money to even try. Now, there are two
18:51additional metrics that are really
18:52important. The first is GPT, gross
18:55profit per transaction. That's AOV minus
18:58CPA. And here's how to set that up. This
19:01tells us the profit from each sale. A
19:03lot of people focus on return on ad
19:05spend, but that doesn't mean anything. A
19:07$1 CPA on a $10 AOV is a 10x rorowaz,
19:10but the GPT is only $9. A $50 CPA on a
19:14$100 AOV has a rorowaz of two, but you
19:16made $50 on every transaction. You don't
19:18care how profitable the last ad looked
19:21in some fractional ratio based on last
19:23touch. You care about how much profit
19:24you make on every sale. If you're trying
19:26to increase the volume of profit, it's
19:27simply the volume of transactions
19:29multiplied by the gross profit per
19:31transaction. Make the volume bigger or
19:33the GPT bigger and the volume of profit
19:35goes up. That's why we're in business in
19:37the first place. More profit. The second
19:40additional metric is AOV. And I look at
19:42AOV as customer quality. Are we selling
19:44something at $90 or something at 60?
19:47Maybe an ad has a great CPA because the
19:50AOV is really low and it's just a easy
19:52sale product that has no real value to
19:54us. And if we're in an ad set where the
19:56AOVs are dramatically different, say one
19:58is 65 and another is 100, we're selling
20:01dramatically different products to
20:03dramatically different people. That
20:04means we're not teaching the machine
20:06what success actually looks like. We're
20:08just saying sell anything. But you don't
20:10want to just sell anything. You want to
20:12make the most money. This puts into
20:15context what we need to do when we can't
20:16increase the budget. If move number one
20:18is a no, move number two is to remove
20:20the ad that it's the biggest liability,
20:22saying yes to move number one. The
20:24simplest way to do this is to sort your
20:26ads by spend. Find any ad that spent at
20:28least 10% but no more than 40% of the
20:31budget where the CPA is above average
20:33and the GPT is below average. That ad is
20:36soaking up a ton of money so that you
20:38can pay more money to buy a worst
20:40customer. Every time money goes into
20:42that ad, you lose. The opportunity cost
20:45is crushing your profitability and it
20:46makes zero sense to keep it running. If
20:48we can remove that ad, everything gets
20:50better overnight. Hopefully, that buys
20:52us enough leverage to say yes to move
20:54number one. And if after 3 days, the
20:56machine still hasn't shifted enough to
20:58get us there. We go back to step two. Is
21:00there something else to remove? If yes,
21:03do it again. And for the same reason
21:04that marginal scaling fires only two or
21:06three times a week, we make move number
21:08two two or three times a week. We have
21:11to let the machine adapt. You cannot
21:14compound choice on top of choice on top
21:16of choice. That leaves us somewhere we
21:18cannot possibly plan for. If there's
21:21nothing left to remove and performance
21:23still isn't good enough, that is when we
21:25look at the 4PI and ask one question. Do
21:28we have an upperfunnel problem or a
21:30lowerfunnel problem? That puts the
21:32context on what we need to do for move
21:35number three, which is to test new
21:38creative.
21:42So, now that we know we can't increase
21:44the budget because we don't have the
21:46leverage and there are no ads we can
21:48remove to create more leverage, the next
21:50move is to introduce new ads. And that's
21:53step number three, creative testing.
21:55Creative testing is not about finding
21:57winners, it's about making the team
22:00better. Remember, Facebook ads is a team
22:04sport.
22:06How do you make your team better? What
22:07is the biggest gap you have? With the
22:114PI analysis, you should already know
22:13whether you have an upperfunnel problem
22:15or a lowerfunnel problem. If your ads
22:18aren't efficient enough, your frequency
22:20is low, and your CPA is high, introduce
22:23lowerfunnel ads. If you have great
22:25efficiency, but your frequency and CPMs
22:28are high, and every time you scale the
22:29budget, everything gets worse, you have
22:32an upperfunnel problem. When we go to
22:34test creative, the single best way to do
22:36it is simple. We put the ads in the same
22:39campaign, we use broad targeting, and we
22:42use 322 ads. Two of the questions I get
22:44asked all the time are, "Do I put it in
22:46the same ad set or a new one?" And, "Do
22:49I use CBO or ABO?" And the answer to the
22:52second question is easy. We never use
22:54always CBO. And here's why that
22:56matters. If the only way your account
22:58works is by you micromanaging it, that
23:01means the only way it scales is by you
23:03doing more work, eventually you're just
23:05going to work harder and harder against
23:07yourself more and more, building a more
23:09and more complicated machine that will
23:11never get you to where you want to go.
23:12You're watching this video because you
23:14already know that doesn't work. For the
23:18first question, same ad set or new ads
23:20set, it's also actually a really simple
23:22decision. Can you afford to get a second
23:25ad set out of learning? If the answer is
23:27no, the new ads go in the existing ad
23:30set alongside what you're already
23:31running. If the answer is yes, put those
23:33new ads in a new ad set. I also highly
23:36recommend that you don't launch just one
23:37322 ad. Launch two attempts at the same
23:40concept, trying to solve the same
23:42problem, just slightly different
23:44variations. If you only test one ad,
23:47you're essentially asking, "Is this ad
23:49good? Yes or no." You learn nothing from
23:52that. It's a binary outcome. If you test
23:55two, you can understand whether the
23:56concept works, and if so, which version
23:59works better. And if neither works,
24:02don't test the concept again. It's not
24:03worth pounding into a loss. Let me take
24:06a moment to define what a 322 ad is and
24:08what it isn't. And I want to be clear, I
24:10invented the 322 ad over half a decade
24:12ago. So, it's been genuinely great to
24:14see basically everybody using it now.
24:16Even Alex Becker is telling everyone
24:18it's the only thing they should be
24:19doing. So, thanks to Alex Becker for all
24:21the shout outs. And I'll leave a link to
24:23a video down below where he was just
24:26giving me flowers. So, thank you, Alex.
24:28Now, a 322 ad is simple in construction.
24:31Three creatives, two primary text, two
24:33headlines. Starting with the headlines,
24:35they need to be substantially different.
24:38They need to focus on a different
24:39reason. Somebody should care. They give
24:42context to the creative, but they cannot
24:44just be rewardings of the same point.
24:47And the primary text is what people read
24:49after they see the creative in the
24:50headline. It's for overcoming
24:51objections, giving context, and building
24:53trust and authority depending on where
24:55you are in the funnel. Also, make sure
24:57to keep the primary text in line with
24:59the angles you're pushing in the
25:00headlines. It needs to be a cohesive
25:02unit. One additional note, if your
25:04primary text is long enough to trigger
25:06the see more button, understand that you
25:08now have an above the fold and below the
25:11fold opportunity, you should have a
25:13complete thought above the fold that
25:15earns the click on the see more and an
25:17immediate call to action right below it.
25:19And then finish the conversation below
25:21that. You tease, ask for the sale, and
25:23then give other reasons to click if
25:25somebody doesn't say yes right away.
25:27Now, the creative, which is probably the
25:29most misunderstood part of a 322 ad,
25:32there are three simple rules. First, all
25:34three creatives should be the same
25:35format, the same aspect ratio, and
25:37either all video or all images. Second,
25:40they should all represent the same
25:41marketing angle, the same concept, three
25:43images about features and benefits, or
25:45three videos that are UTC testimonials.
25:48They should all be basically the same.
25:50And that goes for both 322 ads that
25:52you're launching. Third, and this is
25:53what almost everybody misunderstands,
25:56the 322 is the ad. It has 12
25:59permutations in it, but you're not
26:00testing 12 different ads. You're not
26:02trying to find the winning post ID.
26:05You're trying to answer one question.
26:07Does this concept solve my problem? A
26:10322 ad is not asking which permutation
26:12is best. It's asking, does UGC
26:15testimonial solve my upper funnel issue?
26:17We're giving the machine two different
26:18322 ads to make that determination. And
26:22if you end up with one incredible post
26:24that soaks up all the budget and gets
26:26all the comments, it will be immediately
26:28obvious in your inbox. Go there, grab
26:30the post ID, and you can leverage that
26:32as a standalone ad later if you need to.
26:34Once you publish, there are three
26:37possible outcomes. One, the adset gets
26:39spent, the new ads soak up the budget,
26:41and performance gets better. In that
26:43case, going back to the rules of
26:44infinite scaling, you have created more
26:46leverage so you can increase the budget.
26:48And that is the only thing you do. You
26:50don't turn the test off. You don't move
26:51post IDs. You don't mess with it. You
26:53already won the game. Just keep winning.
26:56Two, it soaks up a bunch of budget and
26:58it doesn't move the needle enough to buy
27:00the leverage that gets you to say yes on
27:02increasing the budget. In that case,
27:04regardless of what any CPA or ROAZ
27:07metric says, it's a failure. Turn it
27:09off. Launch a new one. And three, it
27:11doesn't get spent. That is itself a
27:14signal. Now, one caveat here. If you're
27:16testing a lowerfunnel ad, it might not
27:18get spend because the other ads you're
27:19running are so good, the machine doesn't
27:21have to use it, and the algorithm
27:23doesn't know that the outcome it's
27:24giving you isn't good enough. In that
27:26case, I'm okay with putting a spend
27:28minimum on the adset, generally starting
27:30around 20%. With that minimum in place,
27:33when the ad earns spend, we ask
27:34ourselves the simple question again.
27:36Does it make the campaign better or
27:38worse? Does it give you enough leverage
27:41to increase the budget? If yes, scale
27:44the budget. If no, the test is a loss.
27:47Creative testing is as simple as that.
27:49The only reason you are testing new
27:50creative is to say yes to increasing the
27:53budget. Infinite scaling is a flywheel
27:55of making sure you always have a move to
27:57make. Whether you're increasing the
27:58budget, turning off an ad that's getting
28:00in the way of increasing the budget, or
28:02launching a creative test and a specific
28:04scientific way to buy more leverage so
28:06that you can increase the budget, that's
28:07it. Everything else you do in the ad
28:10account is a liability and a complete
28:12waste of your time. Now, if you want to
28:15know more about creative testing or how
28:16I set the campaigns up to do all of
28:18this, that's what this next video is all
28:20about.
28:24[music]
28:30>> [music]

About this transcript

This page contains the full transcript of The NEW BEST Way To Scale Facebook Ads by Professor Charley T, generated from the public captions YouTube serves with the video. The transcript has 5,143 words across 733 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.