Hey,

Last month I wrote that creative volume is the number one constraint for most brands between $1M and $5M, and that your quality bar might be the bottleneck.

The reply I got more than once:

“Fine, I’m convinced. But HOW does anyone actually produce that much without the quality collapsing?”

Fair question.

It’s the single most common thing I get asked on audits, and the honest answer surprised me too the first time I watched it up close.

So today I’ll show you the machine.

A brand I work with ships 20 to 30 video ads plus around 10 statics every single week, and their account keeps getting stronger instead of drowning in mediocre tests.

They can do that because volume was never the goal.

The system is the goal.

Volume is what falls out of it.

Why volume decides winners now

Quick context on why this matters so much.

At meaningful spend, a winning ad dies in weeks.

If you ship three ads a month with a typical hit rate of one winner in ten attempts, you find a winner roughly once a quarter, while your current one fatigues in three or four weeks.

That math leaves your account on life support between lucky breaks, and every plateau gets blamed on the algorithm.

The brands pulling away right now simply run more shots on goal through a repeatable process.

Here’s what that process actually looks like.

Rule 1: Script the hook, free the body

Every video starts from a scripted hook and a scripted CTA.

The middle stays loose.

Creators tell their own story in their own words, and the brand controls only the first three seconds and the last five.

That combination sounds small. It changes everything.

The hook is where most of an ad’s performance is decided, so you want it engineered and repeatable.

The body is where authenticity lives, so you want it human and unpolished.

Script both and you get robotic content. Script neither and you get lottery tickets.

The brief a creator receives is one page: three scripted hook options, the CTA line, and a bullet list of story beats to cover however feels natural.

Recording three hooks adds ten minutes to a creator’s shoot and hands the brand three distinct ads from a single session.

Their best-performing hook family right now is time-based testimonials.

“I’ve been using this for 90 days, here’s what happened.”

Time markers spark curiosity and carry built-in credibility, because nobody claims day 90 on day 2.

Rule 2: Winners get repackaged, never retired

When a hook wins, most brands run it until it dies and then mourn it.

This team treats a winning hook as raw material.

One winner becomes a split-screen version, a green-screen version, a B-roll cut with voiceover, and a trend-format edit.

Same message, four different faces.

The platforms increasingly punish creative similarity, so identical repostings fatigue fast, while genuine format variety keeps a proven message alive for months.

Their editor turns a winner into the next three formats within a day or two, and every one of those is a high-probability bet, because the concept already proved itself.

A single winning concept can quietly power 10 to 15 distinct ads before it’s exhausted.

That’s where half the weekly volume comes from.

You feel like you need 30 new ideas a week.

You actually need two or three, plus a repackaging habit.

Rule 3: The 70-20-10 split

Every week’s output follows the same risk budget:

• 70% variations of proven winners. New creators, new formats, new hooks on validated concepts.
• 20% adjacent experiments. A new angle on a known persona, or a known angle on a new persona.
• 10% wild swings. Things that will probably fail and occasionally redefine the account.

The split protects them from the two classic failure modes: accounts that only exploit (everything works until everything fatigues at once) and accounts that only explore (constant novelty, zero compounding learning).

If budget is tight, shrink the batch and keep the ratio.

Five ads a week on 70-20-10 beats twenty ads of pure gambling.

All testing runs in separate testing campaigns with their own budget.

Winners graduate into the scaling campaigns (while also getting scaled up in the testing campaign).

Nothing gets tested inside the campaigns that pay the bills.

Rule 4: Creators who post first, get paid second

Their creator pipeline has one filter that outperforms every casting call: they prioritize creators who already use the product and post about it organically.

If the content works on a creator’s own page, it earns paid amplification.

If someone only performs enthusiasm for money, the audience smells it in one scroll.

Affiliate codes follow the same logic.

Selective, earned through performance, and never sprayed at everyone with a pulse.

They also match creators to buyer personas deliberately.

The person in the ad should look like the person the ad is for. Obvious, and skipped by almost everyone.

Rule 5: Label everything, so learning compounds

Every asset carries a hook ID and a body ID.

When something wins, they know WHICH hook won, on which persona, in which format, instead of a vague “that video did well.”

If you can’t say what made your winner win, you didn’t test. You guessed.

The weekly rhythm around this is one 30-minute creative review.

What won, what died, which winner gets repackaged next, which wild swing gets a shot.

Every decision feeds the hook library, so the system gets smarter every week instead of restarting from zero.

What this actually requires

Less than you think.

One person owning the hook library and briefs.

A bench of five to ten creators on flexible deals.

An editor for the repackaging work.

The brand added people as spend grew, but the SYSTEM ran exactly the same back when output was ten assets a week.

If you’re spending $50k to $100k a month, the miniature version below is enough to change your account inside a quarter.

Your move this week

You don’t need 30 ads a week at $2M. You need the machine in miniature:

1. Take your single best ad from the last 90 days and repackage it into three new formats.
2. Write five time-based testimonial hooks for your hero product.
3. Set your testing budget to the 70-20-10 split and move testing into its own campaign.

Do that for four weeks and you’ll have more validated creative than the last six months produced.

Talk soon,
Andrej
​

PS: Every founder discovering why the “lucky” accounts stay lucky →​