I tested over 200 ads for one client last month.
A big chunk of them worked at low testing budgets. But only a handful held a strong CPA once we pushed past $10k/day.
Those high scale ads are the ones that carry an account. They're what we're always hunting for.
But here's what nobody talks about. The ads that support your top spenders matter just as much.
Having a handful of ads that perform well at lower budgets works to your advantage.
→ They pick up purchases here and there
→ They hit smaller pockets your big ads don't reach
→ They quietly contribute to your blended numbers
Not every ad needs to scale to $10k/day to be worth running. Your account is carried by a few big winners and held together by a bunch of smaller ones working underneath them.
You need both.
that girl you scrolled past on instagram last week. 15,040 followers, mirror selfies, captions about being single. guys in her DMs at 1 AM calling her baby. a paid subscription page with a growing fanbase.
she was generated in AI in under 10 minutes. nobody in her audience knows that. and the person running her accounts clears more per month from a laptop than most people take home from a desk job.
almost nobody outside this space has seen how the full operation actually works. here it is.
the girl: one AI tool (nano banana pro) generated the base face. a second (seedream) pulled 20 pose variations from that image. a third (cdream) locked her consistency across every platform so she looks like the same person in every photo. one afternoon produced a 60-image content vault that powers her instagram, tiktok, and threads presence simultaneously. she has a name, a city in her bio, a personality, a backstory documented in a reference sheet the whole team works from.
the audience: her instagram was built from zero with a 3-week boost at £14/day targeting US and UK men only. at 3,000 followers the ads stopped and organic took over. in parallel, a tiktok operation pushes 25 carousel posts per day across 5 accounts, all featuring her, timed between 9 PM and 5 AM to land during the US late-night scroll. every platform routes through a single landing page to her subscription.
the revenue: subscribers pay $4-5/month. that fee is roughly 10% of total income. the other 90% comes from the conversation layer. when someone subscribes, a personalized welcome fires after a deliberate 2-minute delay (instant messages read automated and kill trust on arrival). the team captures his real name in the first few messages. from there, content is sold through sequenced conversation that escalates naturally, not a single blast to the whole list. one engaged subscriber regularly spends hundreds in a single session through this approach.
who's paying for this: men aged 19 to 65. every income level. every profession. scrolling alone at night when the house is quiet. this isn't a niche demographic hiding in a corner of the internet. male loneliness is now documented as a full-scale epidemic and the spending pattern behind it is one of the fastest-growing revenue streams in online business. the market underneath this model hasn't shown a single sign of slowing down.
one AI persona. built in an afternoon. maintained by a small team. and the infrastructure behind her is significantly more sophisticated than it appears from the outside.
i documented the entire system, every tool, every workflow, every SOP and revenue mechanic, and dropped it into a free telegram group. nothing behind a paywall.
RT this + comment BLUEPRINT and i'll DM you the link (must be following)
this girl made $10,000 without leaving her room, and her name is on nothing
she opened youtube and found a kids video with 40 million views, no name on the channel, no face anywhere, just a cartoon about counting and colors
she copied the link, pasted it into pixart flow, pressed generate. five minutes later she had a video that looked exactly like it
she posted it and went to sleep, woke up to 4,000 views
a 3 year old has no skip button in their head, they just watch until the screen goes dark, then cry until someone turns it back on. youtube counts every second of that, rewards it, and pushes the video to the next kid
$66 a month in tools, no face, no name on the channel
month 3, $10,000
she still doesn't have her name on anything, and that was always the plan
5 static variants of the same idea is not diversification.
1 static + 1 UGC video + 1 founder talking head + 1 testimonial video + 1 product demo, all of the same idea, is real diversification.
Why: format mismatch is what reaches new audiences. Meta segments delivery by format preference. Some users will only convert on video, others only on static, others only on UGC.
Look at your last 30-day creative output. Count formats. If video is under 30%, you have a format-monoculture problem. Fix that before launching another static variant.
Most people will tell you there is one perfect Meta ads structure for ecom.
There isn't.
We use 4 different structures across our clients depending on AOV, niche, budget, and creative volume.
A $40 AOV supplement brand spending $500/day does not need the same setup as a $200 AOV apparel brand spending $5K/day.
Single CBO. Bid cap. Cost cap. ASC sales campaign as a CBO.
The brands stuck at flat ROAS are usually running a structure that doesn't fit them. The structure isn't wrong in general. It is wrong for THEIR account.
Stop copying structures. Start matching them to the brand.
Rock bottom is a beautiful place. Once you are there, you realize life can't get any worse. It's just you & God. The only way out is up. You have nothing to lose but everything to gain. Is that not the most beautiful part of life? Opportunity? Your time at rock bottom never goes to waste either. Once you make it out of rock bottom - you carry a certain swag. You can live life without fear, knowing you can lose it all & still make it back.
It’s poetry.
I made a shit tonne of money from 17-19, I lost it all by 21/22 (due to a combination of crypto, bad biz partners, bad timing & all-round bad decisions). I always remember one time where my card declined at Starbucks & my ex-girlfriend had to scoop the bill instead. Nowadays, a few years later (at 25), that story always makes me laugh as I pull up to those same drive thrus in an f8 spyder.
I remember explaining all this to someone a few months ago & they asked why I didn't just delay my dreams. As he said that, it occurred to me in that split second that even though it felt like the whole world was against me - not once did I ever think of quitting. That word didn't even exist in my subconscious.
I have a lot of time for people who have visited rock bottom in life, I have no time for quitters.
Life always gets better & when people say EVERYTHING HAPPENS FOR A REASON - they mean it - there have never been truer words spoken. I would not be where I am today if I didn't go through that shit a few years ago. If you can't see the reason yet, it’s because your journey is still being written. But one day it will click & you too will admire the romance of life.
May God protect you & the universe bless you.
If you want to retire your mum this year with Ecom, study native ads.
The brands that crack it easily scale to 100k+ per month.
But most people don't know how...
THIS GIRL MAKES $29,000 A MONTH FROM ADVERTISING
She doesn’t exist
The 0:21 mark is a video that Kling generated in 4 minutes
Over 100,000 followers
Three sponsorship deals. Her own merchandise. Hundreds of messages from fans every day
She has never filmed a single video
Never held a camera
Never posted a story
All the videos are generated
And all her accounts are managed by one guy from his bedroom
$3 and seven minutes bought me 700,000 views.
I took a girl I had already filmed.
Same face.
Same room.
Same exact clip.
Only her hair was different.
The algorithm treated it like an entirely new person.
So did half a million strangers scrolling.
They did not recognize her.
They did not notice the room had not changed.
They just saw new hair and kept watching.
Last month the same result cost me 100 hours of shooting and editing.
New outfits.
New locations.
New scripts.
This time I spent three dollars on a single change and seven minutes running it.
The output looked fresh enough to reset the entire distribution.
Same content.
New identity in the eyes of the feed.
That is the part that stays with you.
Not how many views it got.
How little it actually took to manufacture something the algorithm had never seen before.
5 native ad headlines for hair loss brands:
1. The shower habit accelerating thinning in women over 40 (it's not your shampoo)
2. A dermatologist revealed the real reason hair gets thinner after 45. It's not genetics.
3. She went from a ponytail she was ashamed of to hair her stylist asked about. In 11 weeks.
4. The reason DHT blockers stop working after 3 months has nothing to do with tolerance
5. How women over 50 are quietly reversing 10 years of thinning without prescriptions or procedures
All five use the same principles:
Challenge a belief. Remove the main objection. Sound like an article, not an ad.
1 sketch that will automate your whole FB ads game
Bookmark this 📍
Most brands overcomplicate media buying.
A simpler setup usually wins:
→ Run one broad CBO campaign per country and objective.
→ Let Meta handle targeting instead of layering endless interests.
→ Keep all proven ads inside one dedicated winner ad set.
→ Launch multiple DCTs (Dynamic Creative Tests) to test fresh angles.
Use DCTs to test:
• Static creative concepts
• Video hooks, scripts, and CTAs
• Headlines and primary text
When a creative wins, move it into the winner ad set using the Post ID.
Why?
Because engagement stays attached to the ad instead of resetting every time you duplicate it.
The result:
• Budget naturally flows toward your strongest creatives
• New concepts compete against proven winners
• Meta gets cleaner optimization signals
The less time you spend managing campaigns, the more time you can spend creating ads that actually drive revenue.📈
The good thing of spending 8 figures a month on meta is that when your personal IG’s and other accounts get hacked/disabled you’ll rep get em back in a fingersnap. That AI exploit still pretty crazy though but all good we back at it on IG.
I made $680k in 30 days dropshipping a $35 Shilajit supplement.
I copy pasted the funnel in 30 minutes, generated ads with AI, and launched.
Total setup = 2 hours using AI + Afterlib.
Like & comment "SHILAJIT" and I'll send you the exact funnel and breakdown.
I make over $50k/month by generating AI content of girls
All I do is generate content on higgsfield, post this content on IG then make more content to sell on fanvue…
And the best part is, it’s a copy & paste strategy that will work consistently once you learn how to do it
Comment below if you want me to send you my FREE Guide on this (must be following so I can DM)