Twelve things
worth knowing
this morning.

“App download brands have the best ads. They have to get someone to fill out a quiz, download an app, and then pay inside the app. I've swiped from them in ecom and every time those ads became top spenders.”

The hardest conversion path produces the best creative. An app-install advertiser has to carry a stranger through three separate commitments before a dollar moves, so the hook is doing work an ecom hook never has to do. Swipe from a funnel harder than your own. Bizop is the same three-step shape — click, set something up, then pay — and nobody in our rooms is looking at that category.

“Average creative strategist: I'll rip this Ryze ad, it's been live for 127 days, must be a winner. Top 1%: a deep psychological understanding of the ICP, and a tight feedback loop.”

Run time proves the ad makes money for its owner — with their offer, their lander, their payout and their account history behind it. None of that transfers. Ripping is what gets used in place of the customer read, not alongside it. Sits directly against what the rooms say below; see the split.

“These are not the real ads btw.” — on a wall of Ad Library screenshots. Asked whether the point was to keep competitors out or to get past review: “It helps with FB ad review.”

In some categories the creative sitting in the library is not the creative being served. That puts a ceiling on how much any library-based tool can be trusted: run time and page history are real, the asset is not. Worth holding next to our own spy radar, which reads the same source.

“I'm having too much fun with these Seedance 2.0 hooks. You can literally make a yapper like this for ANY product.”

The clip is a kitchen talking-head that was never filmed. The cost of a UGC hook has gone to roughly zero, which means hook volume is no longer gated by shooting — and the bottleneck moves one step back. What a creative pipeline should now be producing is angles, not assets.

“Five different briefs with 30 ads each, across 20 ad accounts. One brief is $500. If it spends $150 and doesn't have 3 sales it goes off. If it does, it gets duped 3×.”

The 150 creatives a day is not the method — the cut is. That volume is only survivable because the kill rule is arithmetic and fires before anyone's judgement gets involved. Copy the rule, not the headcount.

“Other people's CTRs from their ads and lander combos have no meaning to you.”

A borrowed benchmark always arrives with a lander you don't run and a payout you don't get. The only comparable number is the last version of your own funnel — which is an argument for keeping the history, not for collecting other people's screenshots.

“One launch could be 10 angles, 50 images — five per angle — and three ad copies per angle. Spend $1,000–2,000 and break even. Then mass scale what came out of it.”

Break-even is the pass mark at test stage. Profit is a scaling-stage question, and asking it of a test is how a good angle gets killed at $300 before the account has any idea who to show it to.

“Shot on an aging iPhone 11 by a 50-year-old who is not a photographer. Slight motion blur, slightly overexposed, head cropped at the top edge, autofocus soft on the face.”

That is a working image prompt, and it reads as a list of defects. Every instinct that makes a photograph good — sharp, centred, lit — is the instinct that makes it read as an ad and lose the scroll. Goes straight into the bizop image brief.

“Manual is still fine for spending like $10–30k a day.” — said while describing domain rotation, lander hosting and split-test launches all running through scripts.

Someone has put a number on the ceiling of hand-run buying. Above it the binding constraint stops being budget or creative and becomes how much of the launch you have handed to software. That is a scoping number for our own tooling, not a boast.

“Very unlikely to work without a VSL or long prelander — could be this publisher hosts the VSL themselves, and that's why the advertiser never shared it.”

In bizop, a funnel arriving with no prelander is a flag on the offer, not on the traffic. And when the page is the thing nobody will send you, the page is the asset — which tells you where the work actually is before a dollar goes in.