Two Interns Wrote a $2M Facebook Ad: Copywriting Is a System, Not a Talent

Here's what it's costing you to believe copywriting is a gift. Every time a winning ad gets attributed to one talented writer's instinct, you've quietly decided that your ad output is capped at whoever that person is and how they're feeling that week. If they're booked, you wait. If they leave, the winning streak leaves with them. You've built your acquisition on a personality, which is about the most fragile thing you can build on.

I don't buy it, and the evidence doesn't either. The best-known version of this story is two interns, no real copy background between them, who were handed a system and went on to write an ad that did over $2M. The lesson everyone takes is "wow, hidden talent." The actual lesson is the opposite. They didn't have talent. They had a process, and the process did the heavy lifting.

That's what I want to lay out here. Not a pep talk about creativity. The actual inputs we assemble so that a junior creative strategist, in their first month, can produce a real contender instead of a polished guess. Because copy isn't summoned. It's built from parts you can prepare in advance.

What "copy is a system" actually means

When a strong writer sits down, it looks like inspiration. It isn't. They're pulling from a stack of raw material they gathered before they wrote a single line: what the customer actually said, what the product actually does, a structure to pour it into, and examples of the form done well.

Take those four things away and even a gifted writer produces mush. Hand those four things to someone average and they produce something that converts. The difference between a winning ad and a losing one is mostly in the prep, not the prose.

So the system is just those inputs, made explicit and reusable. Here's each one, and how we build it.

1. The research document: their words, not yours

This is the foundation, and it's the part most people skip because it's slow and unglamorous.

Before anyone writes, we build a research doc that is nothing but the customer's own language. Real reviews, good and bad. Support tickets. The angry one-star rant and the gushing five-star one. Reddit threads. The exact phrases people use when they describe the problem to a friend.

Why their words and not ours? Because copywriting is a series of head nods. Every sentence should make the reader silently think "yes, that's me." The fastest way to earn that yes is to describe their problem back to them in the precise words they'd use themselves, which you can only do if you've collected those words first. Write from your own head and you'll describe the product the way the founder sees it, not the way the buyer feels it. Those are rarely the same.

A good research doc for a single product might be a few thousand words of pure customer voice, sorted into the problems they mention, the outcomes they want, and the objections they raise. That document alone does maybe 90% of the work. If you know what your people want and why they want it, the writing is almost an afterthought.

2. The feature-to-benefit list: every claim, translated

The second input is a simple two-column list that we build for every product, and ideally every feature of every product.

Left column: the feature. Right column: what that feature does for the human buying it. The bridge between them is two words, "so you can."

It sounds almost too basic, but it's the thing juniors get wrong constantly. They list features and assume the benefit is obvious. It never is. "Double-insulated wall" is a feature nobody cares about. "Double-insulated wall, so you can leave your coffee on the desk for three hours and it's still hot when you get back to it" is a benefit. Same fact. Completely different effect.

So we go through the whole product, feature by feature, and write the "so you can" for each. Then we do the same for the next product, and the next. Over time that list becomes a bank you draw from. When a strategist needs to write an ad, half the persuasive lines are already sitting there, translated and waiting. They're not inventing benefits under deadline pressure. They're choosing from ones we worked out calmly in advance.

3. The formulas: a skeleton to pour the words into

The third input is structure, and we keep it deliberately simple. New strategists learn two formulas, no more.

The first is PAS: pain, agitation, solution. You name the customer's frustration, you press on it until it stings, then you offer the way out. It starts from the raw problem, so it pairs beautifully with a research doc full of complaints.

The second is AIDA: attention, interest, desire, action. Hook them, build interest, grow the want, then tell them exactly what to do. It's more open-ended and works when you're leading with a benefit or a curiosity rather than a pain.

Why only two? Because they sit at opposite ends. One starts from hurt, one starts from intrigue. Learn to run both and you've covered most of how a person can be walked from scrolling to buying. A junior with two reliable skeletons writes far better than one staring at a blank page deciding how to begin. The structure removes the scariest decision, which is where to start.

One rule I drill alongside the formulas: create desire before you justify anything. The most common junior mistake is jumping to price or technical specs before the reader actually wants the thing. Emotion first, logic second. Justify the purchase only once they've already decided, in their gut, that they want it.

4. The swipe file: proof the form can be done

The last input is a swipe file, which is just a saved collection of great copy to draw inspiration from. Ads, emails, sales pages, the work of writers who clearly knew what they were doing.

It isn't there to copy from. It's there for two reasons. One, when you're stuck on structure, you find a piece that solved a similar problem and you study how it flows. Two, it quietly teaches the form by osmosis. Read enough strong copy and your own instinct for rhythm and pacing sharpens without you trying.

We keep a shared one so a new hire on day one inherits years of collected examples instead of starting from nothing. That's a meaningful head start. They're standing on the shoulders of every good ad we've ever filed away.

The brief we actually hand a new strategist

Here's where it comes together. When a new creative strategist gets a product, they don't get told "write me an ad." They get a one-page brief that forces them through the system in order:

• Audience. Who exactly is this for? Pull three real customer quotes from the research doc that prove this person exists and talks like this.
• Problem. What's the single sharpest frustration this product solves, in the customer's own words? One sentence, lifted not invented.
• Angle. Which desire or pain are we leading with this time? You cannot lead with all of them, so pick one.
• Proof. Which two or three feature-to-benefit lines support this angle? Take them from the bank.
• Formula. PAS or AIDA, and why this one fits this angle.

Fill that out honestly and the ad is basically written before you write it. Every hard decision has already been made with evidence behind it. The actual drafting becomes assembly, which is exactly the point. We've turned the scary, talent-shaped part of the job into a checklist.

And once it's drafted, there's one test that catches most of the weak ones. Read it back line by line and ask, at the end of each sentence, would the reader nod yes or flinch no. One genuine "no, that's not me" and you've lost them, no matter how clever the next line is. Copy is a chain of small agreements. Break the chain once and the rest doesn't get read.

So the question worth sitting with isn't "do we have a talented writer." It's "have we actually built the system, or are we still hoping someone shows up with the gift." One of those you can control. The other you're just waiting on.