The End of Facebook Ads as We Know It

$196

billion

That's how much Meta made from ads last

year, up 22% from the year before, and

every year we hear the same thing.

Facebook ads are getting more expensive.

CPMs climb, cost climb, people swear

it's not like it used to be, and sure,

some advertisers do quit, but that's not

the interesting part. The interesting

part is most brands keep spending

anyway. Brands like Ridge, who classic

and comfort are more successful than

ever. So, here's the question I can't

stop thinking about. If Facebook ads are

getting so expensive, why do so many

companies keep spending money on them?

If advertisers truly believed Facebook

ads no longer worked, we'd expect

budgets to get cut, but that's not what

we're seeing. Meta's ad revenue is

climbing at a record pace. So, whatever

advertisers are saying publicly, their

budgets are saying something else.

Budgets don't lie, and all of this

points to a story you've probably heard

before.

When confidence is high, people spend.

When confidence breaks, they stop. So,

let's keep it simple. This is a chart

that tracks confidence through spending

over time. To make the comparison fair,

we're putting each market on the same 0

to 100 scale. First, housing. If we

watch the line, it climbs slowly, and

then accelerates, and then it hits a

peak, and then it breaks.

In 2008, houses didn't suddenly stop

being useful. People still needed

somewhere to live. What changed was the

belief in what happens next. People

stopped believing that prices would keep

rising. Banks stopped believing loans

would get repaid, and builders stopped

believing new homes would sell.

Confidence disappeared. And in some

markets, home prices fell by more than

50%.

Now, let's look at NFTs.

It's a slow rise and then a spike. It's

basically the same shape, although far

more aggressive, and we see that

collapse. People were paying six figures

or more for pictures of monkeys, and a

year later, many of those same assets

were down more than 90%.

The technology didn't change. Confidence

did.

And once confidence leaves a market,

demand tends to leave with it.

The most important thing isn't housing

or NFTs, it's the pattern. It's the

shape. It's the behavior.

When confidence truly breaks, you can

see it. Now, let's look at Facebook. If

confidence truly disappeared, we'd

expect to see something similar. Maybe

not a total collapse, but at least a

meaningful decline. But when you chart

it, this doesn't look like a market

that's losing confidence. It pretty much

only ever goes up and to the right. That

is not what a collapsing market looks

like. Sure, there's a dip around 2021,

but compared to housing, compared to

NFTs, it's barely a blip, which creates

a problem. Because if you're running

ads, it sure doesn't feel like up and to

the right. So, if confidence really

never disappeared, why does it feel like

you're paying more for less?

What's going on here? I think the chart

is hiding the answer. Because one graph

makes Facebook ads feel like a

consistent story, but it hasn't been

consistent at all.

When I started in 2012, it was a

completely different world. And then

COVID hit, and today is something else

entirely. So, instead of one big story,

I want to try to break this down into

the eras where the rules actually

changed. I think there are four. And

when we break this into eras, this chart

starts to tell a very, very different

story. So, let's unpack each one of

these stories.

In 2012, Facebook made just under $4.3

billion in ad revenue, and CPMs could be

as low as $0.55.

You couldn't optimize for purchases.

That wasn't an option. There was no

Facebook pixel, no purchase event. At

the time, I was a supervisor at Omnicom,

and sometimes I would manage over a

million dollars a day in engagement

campaigns. And that sounds ridiculous

now, but that was the game, and the game

was cheap because supply was basically

infinite, and demand was low because

there just wasn't much confidence you

could reliably turn ad spend into

revenue. High supply, but low

confidence. And that led to low spend.

But all of that was about to change.

On October 14th, 2015, Facebook released

the Facebook pixel, and that changed

everything. For the first time, you

could buy customers with confidence. The

conversation shifted overnight. You

stopped asking, "How many people did we

reach?" and started asking, "How much

money did this campaign make?" Facebook

could turn spend into sales, and the

floodgates opened. The market was on a

historic bull run, consumers were

spending, rates were low, and Shopify

made it stupid easy to launch a brand.

Soon, you had a whole generation of

brands, Allbirds, for example, proving

that you could launch on Shopify, scale

on Facebook, and hit nine figures. They

did it in only three years. And at the

same time an entire industry was born.

Facebook specialists,

media buyers, agencies, creative teams

built around one platform. People didn't

just get good at marketing, they got

good at Facebook. In late 2017, Facebook

held a private event for the top

advertisers on the platform called Built

to Break. The D2C brands in that room

were Purple Mattresses, Movement

Watches, MeUndies, Skullcandy, Dollar

Shave Club, and yes, here's a picture of

me plus a little sticker I kept as a

souvenir. Facebook unveiled what they

called the Power 5, a generational leap

towards an automated machine-led ad

platform. And for those of us in the

room, it felt like they were handing us

the keys to the kingdom. They gave us

almost a 2-year head start on where the

platform was going. Now, by the end of

this era, CPMs had gone from $0.55 to

nearly $10 by 2020. Ad revenue had gone

from 4.3 billion to nearly 85. Ads had

gotten 18 times more expensive, and

brands were willing to spend about 20

times more. Why?

Confidence.

By 2020, billion-dollar businesses had

been built on Facebook. This was the era

where teenagers could spin up a Shopify

store, sell fidget spinners, and make

more in a week than their parents made

in a year. And it proved something

important.

Even at 18 times more expensive, the ads

were still worth it because the value

was undeniable.

And that brings us to era number three.

If you look back at our graph, this is

the first time that up into the right

line actually dips. Now, when you try to

explain this, the easy answer is to say,

"iOS 14 broke everything." And yes, iOS

14 mattered,

but it doesn't fully explain what we're

seeing here because there were still

brands absolutely crushing it in this

window. So, I went digging for something

else, anything that had the exact same

shape at the exact same time.

And I think I found it. This is the

stock market.

Same curve, same drop, same timing.

These shapes are almost identical,

which tells you this wasn't a platform

confidence problem. It was a

macroeconomic

confidence problem. Yes, this era was

brutal, but it wasn't brutal for

everyone. Brands like Rage and True

Classic were still scaling and the gap

widened. The advertisers who adapted

kept moving and the ones who didn't,

especially the ones still trying to run

the old manual playbook, got wiped out

in a way we hadn't seen since Facebook

became a real performance channel. The

D2C graveyard filled up with former

darlings. Casper, BH Cosmetics, even

Allbirds, the poster child of the gold

rush, started to fall apart. Now, let's

do a quick scoreboard check because this

entire era is only about 2 years. Ad

revenue went from roughly 84 billion to

about 113,

but in 2022 it dipped down about 1.5

billion dollars and even with that, CPMs

kept climbing.

Now, we were sitting at little over $12.

So, the market didn't die, it changed.

And that brings us to era number four.

It's easy to say that era number four

was defined by AI, but I think it's

actually defined by democratization

because after the stock market recovered

and confidence came back, Meta doubled

down on a completely different strategy.

Instead of building better tools for the

biggest spenders, they built smarter

tools for everyone else.

Tools like Advantage Plus and updates to

the algorithm like Andromeda, the system

doing more and more of the heavy

lifting. And the message got pretty

clear.

Stop trying to outsmart the algorithm.

Work with it. And this matters because

Meta already had the biggest advertisers

in the world. But interestingly enough,

there's not a lot of upside in

convincing brands like Groons or Savage

X Fenty or even Amazon to double their

budget.

So they went after everyone else. Today,

there are more than 10 million

advertisers on the platform. And over

80% of them spend less than $100

a day. Meta grew by getting millions of

small businesses to spend a little bit

more.

Restaurants and roofers, dentists,

creators, local gyms, mom-and-pop

print-on-demand stores. If you could

double, triple, even quadruple the ad

spend there, that's massive. The big

brands that survived era three weren't

surviving on arbitrage.

They were already good at the

fundamentals. So these new tools weren't

really built for them.

They were built to bring the next

million advertisers online. So what did

that do to the price of ads? Let's look

at the scoreboard.

CPMs are higher, sure, around the

mid-teens. But annual ad spend has

exploded. In 2026, it's projected to be

over $200 billion.

So if Facebook ads feel more expensive

than ever, but more businesses than ever

keep spending, there's got to be

something else going on.

To review, demand has grown faster than

supply and confidence has gone up. So,

yes, prices have gotten more expensive,

but something still isn't adding up.

Most of you aren't seeing $15 CPMs.

You're seeing $25,

$35,

$50, sometimes more. And I see it across

my businesses and across all the

advertisers inside Disruptor Academy and

the Meta MBA program. So, why is it so

much more expensive for us?

If you open your ad account, you'll see

CPMs can be wildly different from ad to

ad. And if you sort your ads by spend,

you'll notice something interesting. The

ads getting the most spend often have

the lowest CPMs, but they're rarely the

most efficient on cost per result, which

tells us something important. CPM is

just a price of attention. It doesn't

actually tell you what that attention

was worth. So, here's the real question.

If you paid twice as much to reach

people, but they were four times more

likely to buy, did advertising actually

get more expensive? Because cheap CPMs

don't keep a business alive. CPM was the

KPI back before the gold rush.

Businesses survive on profitable cash

flow. So, if CPM only tells us the price

of attention, the metric everyone uses

to judge whether ads are profitable is

ROAS, right? Return on ad spend. It's

simple. Attributed revenue divided by ad

spend. Spend a dollar, make back four,

that's a 4x ROAS. Last year, Meta said

that the average advertiser earned well

above $3 back for every dollar spent

across over a hundred billion dollars in

ad spend.

If the average advertiser is getting a

three to four x return, why does it feel

like Facebook ads stopped working for so

many people? And that's the problem

again. It's the pattern.

The averages don't match the experience.

And I think it's the same mistake we've

been making across this whole video,

this entire investigation. We've been

using the wrong metrics for the wrong

era and averaging things that should

never be averaged.

To put it differently, ROAS was the KPI

of brands like Allbirds.

Shortly after the 2021 IPO, the stock

price peaked at over $648

a share.

But in the last 52 weeks, it hit an

all-time low of $2.15. That's basically

a 300x swing in the wrong direction. So,

yeah, you can chase 3x ROAS and still

destroy the value of the business.

And I don't know about you, but that's

not what I want. So, what's wrong with

these numbers? The problem isn't that

ROAS is inherently a bad metric. It's

that we're averaging businesses that

have almost nothing in common. Going

back to that hundred-billion-dollar case

study from Facebook, where they said the

average advertiser was getting a ROAS of

over three, imagine averaging the ROAS

of a dentist, a restaurant, a roofing

company, Nike, Amazon, and a

twenty-million-dollar supplement brand.

You're going to get a number, but it

won't tell you what good looks like

because none of those businesses make

money the same way. Different margins,

different repeat purchase rates,

different AOVs, different LTVs. They're

playing completely different games. So,

the real question is what game should

you be playing?

When the machine is this powerful,

arbitrage on metrics like CPM and ROAS

just don't make sense anymore. The game

has changed. So, the question has to

change, too. It's not what's a good KPI.

I think it's is there an ideal business

model for Facebook ads? Because look at

the companies that we've been talking

about, Allbirds and Ridge, True Classic

and Comfort, even Amazon. On the

surface, they couldn't be more

different. Different products, different

customers, different ads, but I don't

think Facebook cares about any of that.

I think the algorithm rewards something

deeper.

Not a category, not a niche, not a

product. I think the algorithm is

rewarding a business model.

And the companies winning today aren't

playing against Facebook.

They're playing the whole game with it.

So, I started looking at businesses that

kept winning. Ones that didn't just win

in the gold rush, and not just after iOS

14. What were some examples across

multiple eras, and three companies kept

standing out. Ridge, Grooons, and

Comfort. Totally different industries,

completely different products, and

different customers, but they all kept

growing. So, I wanted to know what are

they doing that everyone else isn't?

Because I don't think it's the ads. I

think it's the business they're

advertising.

Let's start with Ridge. They took one of

the oldest products in the world, a

wallet, and made it modern.

A simple product and a clear promise.

Today, Ridge spends well over $200,000

a day acquiring customers. And that's

the interesting part, because the

advantage isn't the wallet, it's the

business behind it.

Premium pricing, healthy margins, a

product you can understand in 3 seconds,

and an experience that delivers on what

the ad promised. And that matters

because of how Facebook's machine

actually works.

Facebook isn't trying to find you the

cheapest click.

It's trying to find the next person

who's likely to engage, to click, to

buy, and to feel good about it. Because

when people have good outcomes after

they click on ads, they trust the

platform more. They come back, they keep

scrolling, and they keep buying. So,

Ridge isn't winning because they found

some secret targeting trick. They're

winning because they give the algorithm

exactly what it wants.

Happy customers. And when you

consistently create good experiences and

happy customers that are happy to come

back and spend their money again, the

machine gets better and better at

finding the next one. And to be fair,

that's Meta's entire business model.

Grüner launched in 2023. And on the

surface, it couldn't be more different

than Ridge. It's greens. It's vitamins.

It's nutrition.

And they're reportedly on pace for

hundreds [snorts]

of millions a year in revenue.

But the product isn't the interesting

part. The business model is. Because

Grüner's isn't trying to win on the

first purchase. They're trying to win on

the customer. Dan Kennedy has a famous

line, "Whoever can spend the most to

acquire a customer wins the game." Now,

that sounds aggressive, but it's true.

If your average customer stays

subscribed for 6 months or 12, you can

afford to spend more to acquire them

than a business that has to make all of

its profit on day one. And that's why

CPMs can be so misleading here. A $50

CPM doesn't tell you whether someone

stays for 1 month or 12. And ROAS

doesn't tell you whether you're

acquiring a customer that was actually

profitable. Because the real question

isn't, "Did I get a return on this first

order?" The real question is, "Can I

keep buying more customers profitably?"

And when you have a sticky subscription

and you feed those really quality

signals back to Facebook ads, it can see

the rebills, the repeats, the customers

who stick, which makes the machine more

confident. And confidence is what

unlocks scale. Not because their ads are

better, because their economics are.

Comfort launched after iOS 14. It was

born into the hardest era of Facebook

advertising we've ever seen. No gold

rush, no cheap arbitrage. They don't

have a subscription, and yet just a few

years later they're on track for a

billion-dollar annual run rate. That

shouldn't happen. And like Grooons, I

don't think it's because they have the

best ads.

I think it's because they turn their

customers into their marketing

department. Every happy Comfort customer

has the opportunity to create content,

share the product, drive sales, and get

rewarded for it. UGC isn't a bonus for

them. It's built into their business

model. So every sale has the potential

to create the next sale. And that sends

an incredibly powerful signal.

Because Facebook doesn't just see

purchases. It sees what happens around

the purchase. The shares, the comments,

the creators, the repeat attention. The

machine isn't just learning who buys.

It's learning who influences other

people to buy. So Ridge wins through

profitability, Grooons wins through

lifetime value, and Comfort wins through

distribution. That's three completely

different businesses, three completely

different niches, three completely

different strategies, but they all have

one thing in common. They're not trying

to beat Facebook. They've built

businesses that Facebook naturally wants

to scale.

For this whole video, we've been talking

about confidence, CPMs, ROAS, business

models. We've looked at companies like

Allbirds, True Classic, and Comfort.

But, in looking at it, I think we've

missed the most important character in

the entire story, the machine.

Because while Facebook advertising has

evolved through four distinct eras, the

algorithm has been evolving right

alongside it. And I don't think that's

an accident. In fact, I think the last

15 years have been Facebook teaching us

the same lesson over and over again. We

just didn't realize we were the ones

being trained. In the beginning, we

picked literally everything. Audiences

and bids, budget allocation, even when

the ads ran and in which sequence they

were shown. Every few years, Facebook

shipped an update that at the time

sounded insane. Things like, "Let us

optimize for purchases instead of

traffic. Let us move budgets to what's

working automatically. Let us use AI to

create untold numbers of creative

variations." Every time, marketers hated

it. Every time, it felt like Facebook

was taking the steering wheel, and every

time, a few years later,

it became the standard. Every generation

of the algorithm got better at one very

specific thing,

making decisions humans used to make,

but more effectively and at scale. Now,

sure, CPMs are up and CAC can feel

brutal,

but it's also impossible to deny that

spend is up, and the winners are bigger

than ever.

There are more brands today doing nine

figures where Facebook is the primary

growth channel than we've ever seen

before. So, if you zoom out, Facebook

hasn't been releasing random features

for 15 years. It's been steadily taking

choices away from advertisers and making

better choices on their behalf. And the

advertisers who fought that have almost

always paid dearly. You can almost split

Facebook into five generations of the

algorithm. Generation one was rules.

You told Facebook who to target and what

to bid and success

mostly depended on what you did outside

the platform. Generation two was

optimization.

Facebook stopped asking, "Who do you

want to reach?" and started asking,

"What do you want them to do?" Clicks

turned into purchases and attributable

revenue. Generation three was machine

learning.

Value optimizations, better event

signals, server-side tracking, the power

five. Facebook stopped asking you to

build perfect audiences and started

asking you to feed it better data.

Generation four was Andromeda.

Instead of matching ads to audiences, it

started matching people to ads, truly

evolving from an auction based on bids

to a distribution platform based on

merit, where the rules of the game feel

far more like the organic side of the

platform than they ever have before. And

now, I think we're on the edge of the

next generation. Large language models

and AI systems that can rank, adapt, and

generate variations faster and more

effectively than humans ever could. The

machine isn't just deciding who should

see your ads anymore. It's getting

better at understanding what your

business is, why people buy, and which

combinations of creative, offer, and

audience and overall sequencing of the

message are most likely to produce a

good outcome. It's a very different

machine than the one we had 15 years

ago. Honestly, it's a very different

machine than the one we had 15 months

ago. Every generation pushes us to be a

little less like media buyers and a

little more like business builders.

Because the only thing the machine can't

fix is a bad business.

Over the last decade of helping people

run ads and managing over a billion

dollars myself, one of the questions

I've been asked more than any other

is what's a good ROAS?

And it doesn't take long to realize how

slippery of a question that really is.

Because a 2x ROAS isn't automatically

worse than a 5x ROAS. And a $50 CPM

isn't inherently bad, just like a $5 CPM

isn't inherently good.

The numbers only make sense inside the

system they're operating within.

And if you ask the best operators what's

actually working on the platform right

now, they'll all agree on a few core

principles.

First, you measure creative quality by

how the target customer engages with the

ad. Second, they feed the algorithm as

much data as possible. And third, they

treat Facebook like the front end of a

full funnel.

And maybe we've come full circle.

Back in the early days, you obsessed

over who to target and what to bid. Now,

those questions mostly get answered by

what ads do you run and what signals do

you give the machine.

Which means success depends, once again,

on what you do outside the platform.

There are a lot of ways to build a

business that can win with ads.

In this video, we highlighted three

very, very different business models.

One that wins on profitability, one that

wins through lifetime value, and another

that wins almost exclusively through

distribution.

Three completely different systems, all

winning the same game.

And I think that's the real lesson.

For 15 years, Facebook taught us how to

optimize campaigns.

When it was really pushing us to

optimize the business.

Because the algorithm rewards businesses

that consistently create value for

customers, for themselves, and

ultimately for the platform. I don't

know what year five will look like.

Maybe we hit another recession. Maybe AI

replaces the media buyer entirely. Maybe

a new platform takes the throne. But I

do know one thing is clear.

The cost of advertising didn't go up.

The value of the right business model

did.

I hope you enjoyed the video. Subscribe

if you haven't, and take care.

Woo! Woo!

>> [music]

[music]