
Head of Content @northbeam & author of The Media Buyer π°οΈ ads, business, marketing
New York, NY
Joined August 2013
- Tweets2.6K
- Following790
- Followers2.7K
- Likes1K
Pinned Tweet
I publish @northbeam performance marketing benchmarks every week in The Media Buyer.
Join 124,000 other marketers π
themediabuyer.northbeam.io
Yes creative volume is key. But I would say the point is wide reaching variety of styles, not just infinite iterations on the same winning ad
1.4% of the companies running Meta Ads are responsible for 36% of all the ad creative on Meta.
That's industrial-strength creative testing at scale.
How do you even achieve this? (short answer: @northbeam)
Of course you have extreme creative output. But that's only half of it.
To run a good creative experiement, you need a better understanding of ad performance than the ad platforms alone can give you.
@BryanBumgardner explains:
bryan π§ retweeted
1.4% of the companies running Meta Ads are responsible for 36% of all the ad creative on Meta.
That's industrial-strength creative testing at scale.
How do you even achieve this? (short answer: @northbeam)
Of course you have extreme creative output. But that's only half of it.
To run a good creative experiement, you need a better understanding of ad performance than the ad platforms alone can give you.
@BryanBumgardner explains:
I expect this to cross 30% before Dec 31 2026
AI will soon originate >15% of all organic search traffic.
After Black Friday, this number will likely be higher.
Are we already in an agentic reality?
Some @northbeam data, presented by @BryanBumgardner
bryan π§ retweeted
High AOV brands are outperforming this year. What's going on?
@BryanBumgardner explains, chart attached.
bryan π§ retweeted
Northbeam Γ© um saas de multi-touch attribution e media mix modeling.
Γ um modelo de dados high level que grandes marcas DTC usam pra operar.
Eles tem acesso a conta das grandes marcas e o Bryan faz esse Index semana a semana que mostra conversΓ£o, roas, %budgt alocado nas principaius plataformas.
Insights Chaves:
> Meta continua sendo top spender channel (+55%)
> Youtube ads nΓ£o representa nem 5% do budget de mΓdia, puta oceano azul. E se Γ© oceano azul lΓ‘, aqui Γ© mais ainda...
> Tiktok tem quase que o dobro do %budget que o Youtube Ads e ainda representa sΓ³ 4.7% do total.
More this week from @northbeam: ROAS begins to drop as the Prime Deal Days rush settles down.
Still, the core channels actually saw ROAS improve week over week since so many advertisers burn in the red during promos. This is a regression to the mean.
We are approaching critical mass. In the first week of November, all of this will tank as spend skyrockets and conversion behavior freezes up.
Soon, everybody will be waiting for a BFCM discount. Expect these performance numbers to plummet over the next 4 weeks.
bryan π§ retweeted
The biggest blocker for scale is your measurement.
Bad data leads to bad analysis which leads to bad strategy.
It all stems from the data. That's why your ads have plateaued - because you're still looking at in-platform data instead of @northbeam.
@bryanbumgardner lays it out:
bryan π§ retweeted
Why do marketers rely so heavily on last-click attribution?
Yes, it's the easiest way to track marketing's impact on sales, but last-click is woefully limited.
Yet every day we see six-figure ad budget decisions made exclusively on last-click. It's maddening.
Last-click ignores MOST of your customer's journey.
Making decisions exclusively on last-click is undervaluing your own hard work.
@BryanBumgardner explains:
bryan π§ retweeted
How are you measuring the performance of your "performance" marketing?
If you're still relying on data provided by your ad platforms...
You're likely not seeing the whole picture.
In-platform ROAS is limited.
Ad platforms can only report on what they see inside their own apps.
Your customers are flying around between your website, your email, and ads on multiple social media platforms.
If you aren't using independent ad performance data to measure your ad performance, you're probably lying to yourself.
@BryanBumgardner explains why @northbeam is the fix. π°οΈ
Most of the ad creative on Meta is made by a select handful of ad accounts.
There are beasts out there running 1,000+ ads at one time.
Our highest performing brands on @northbeam (in terms of ROAS and CAC) are launching more than 180 ads A WEEK.
Its not just enterprise or legacy brands either. Most are actually younger high-growth brands.
Industrial-strength scale requires independent ad performance data.
Go beyond in-platform reporting.
9
12
940
Theory: nobody has felt "normal" since COVID, especially in public settings.
It's time to heal our socialization skills.
Slowly but surely, Americans are starting to remember how much fun it is to hang out.
We were locked in our apartments for so long, working remote, being terrified, that we're all recovering from a collective trauma.
@northbeam has started focusing on in-person events as our marketing strategy. And it's giving me life.
In college I was super extroverted.
But ever since COVID, there's been this nagging weirdness that pulls at me every time I'm about to head out the door.
Once I get there I'm thrilled. I have a great time.
But that hurdle can be hard for most people to jump.
Get outside. Hang out with people.
It's the past, present and future of human existence.
AI can never replace this.
Pic is @dtcwithzach, Gigi Avendano and myself chatting pre-alcohol supplements
4
16
1,201
The difference between mid and S-tier operators is their data.
Specifically, what they use as their "source of truth" on the performance of their ad experiments.
For years I only relied on in-platform data, meaning ad performance data you see inside Meta, YouTube, Google, etc.
99% of folks don't really see an issue with this.
And it's true, if you sell low AOV products on extremely short consideration cycles, you can probably scale pretty far using the data these ad platforms give you.
But once you introduce any type of complexity - new channels, crazy LP testing, pure TOF campaigns like CTV - suddenly that in-platform ROAS looks really unreliable.
I see this over and over in brands booking demos with @northbeam.
You scale a TOF campaign on TikTok, for example, and your Amazon revenue takes off. But your TikTok ROAS looks terrible, even though you know those ads are working.
This is the "walled garden" problem.
S-tier operators have graduated beyond in-platform data. They are using tools that can help them connect that TOF spend with revenue.
Proving revenue as a marketer is the holy grail.
Get data that justifies your marketing spend and you'll take over the world.
3
271
Stop relying on last-click. It makes no sense.
Yes, it's a reliable way to attribute revenue to ads. But it only catches the absolute "last mile" of the customer journey.
Why are we pretending that's some type of proxy for marketing performance?
4
9
792
bryan π§ retweeted
Replying to @andrewjfaris
Yes we use northbeam product attribution quite a bit to see which ads are driving what type of revenue
We see really interesting behavior. Like, during sweepstakes we see our ads drive a MUCH wider distribution across different product categories.
Wake up and smell the multi-touch attribution, my dude.
The fastest-growing brands in ecommerce history love @northbeam
This is the measurement stack I see across 99% of them:
1
1
362
Does anybody use any desktop iPhone emulators?
I want social media apps on my desktop, but I want the mobile app interface.
Does this exist? Or do I just use some kind of mirroring?
2
2
728
People who use @northbeam are just built different.
Northbeam is trusted by the fastest-scaling ecomm brands in history: @comfrt, @ridgewallet, @grunsdaily - the list goes on.
What else do I need to say?
I'll try anyway:
4
324









