Ad copy formulas that actually convert (and the four that quietly don't)

There are exactly three reasons a person clicks a paid ad: they recognized themselves in it, they recognized the problem in it, or they recognized the offer in it. Everything else is decoration. Every copy formula that survives contact with real performance data is doing one of those three things on purpose. Every formula that doesn't is dressing up nothing.

The marketing internet still recommends AIDA, PAS, FAB, the 4 Ps, and Before-After-Bridge as if it were 2014. Some of these still work. Some of them died quietly when feeds got faster, attention got shorter, and the median ad got better. Below is the working list and the dead list, with the actual lines that make each one fire.

What an ad copy formula is doing, mechanically

A formula is a shortcut for a question: which order should the sentences appear in, so the reader gets to the click without dropping out?

There are roughly 50 named formulas in the copy literature. They reduce to a handful of underlying moves:

1. Name the reader so they recognize themselves
2. Name the problem so they recognize their week
3. Name the cause so they trust your diagnosis
4. Name the proof so they trust your claim
5. Name the offer so they understand what they get
6. Name the next step so they know what to do

Different formulas reorder these moves. The ones that still work in 2026 are the ones that get to the click in three to five sentences. The ones that don't are the ones that require seven.

The five formulas that still convert

These are the formulas that show up disproportionately on top-quartile creatives across the categories we score. None of them are new. What's new is that the ones that survive are the ones that run faster.

1. PAS, ruthlessly compressed (Problem, Agitate, Solve)

The classic four-paragraph PAS is dead. The three-sentence PAS is the most reliable conversion formula on the feed right now.

The compressed shape:

"You're spending 6 hours a week reviewing ad creative. That's 25 hours a month not spent on anything that earns. Score them in 30 seconds instead."

Problem in sentence one. Agitation as a single specific number in sentence two. Solve as a verb in sentence three. No transition words, no hedges, no "imagine if" framing. The reader is in the click decision by the end of sentence two.

PAS fails when you let the agitation run long. Two sentences of agitation reads as guilt-tripping. One sentence reads as recognition.

2. Named pain → Named relief

This is closer to a frame than a formula, but it's the most reliable copy structure for problem-aware audiences in B2B and prosumer SaaS.

"If your last creative review meeting produced 14 sticky notes and no priority, you ran a review that hurt the work instead of helping it. The fix isn't a longer meeting. It's a scored review."

The structure: describe the exact moment of pain in detail specific enough that the reader knows you've sat through it, then name the relief as the absence of that exact moment.

This formula works because it bypasses the "is this for me" question. By the second clause, the reader has already answered.

3. The single-statistic open

A specific number, surfaced first, with the rest of the ad anchoring against it.

"71% of paid creative we score has the same five mistakes. Score yours and find out if you're in the other 29%."

This formula works for two reasons. One, falsifiable numbers buy credibility on a feed where vague claims get scrolled past. Two, the second sentence converts the statistic into a self-test, which is a stronger CTA than asking the reader to act.

The trap: the statistic has to do real work in the offer. A bare "73% of marketers struggle with X" with no payoff reads as filler. A statistic that the reader can put themselves on either side of converts.

4. Direct address with a narrow audience

This is the formula that punches above its weight when the audience is narrow enough to name.

"Solo marketers running three or more clients: this is the weekly workflow that saved 8 hours last month, with the exact checklist."

The reader has to opt out instead of opt in. They have to actively decide "no, that's not me" to scroll, and that micro-decision buys you a beat of attention. By the time they're considering the offer, they've already self-selected as the right reader.

Direct address fails when the audience is too broad. "Marketers" is too broad. "Performance marketers running paid social on under $30k a month" is narrow enough that the people who match it lean in.

5. Reverse benefit (the bridge, but inverted)

Instead of describing the better world your product creates, you describe the world the reader is escaping from, and let the product become the door.

"Your week shouldn't be 6 hours of creative review and 2 hours of actual creative direction. Flip the ratio."

This formula is doing two things at once. It validates the reader's current frustration and offers an outcome stated as a familiar mental model (flipping a ratio). It works because the offer is implied, not pitched. The reader fills in "Adverdly does this" before you say it.

The trap: this formula only works if the reader is already in the frustration. It dies on cold traffic.

The four formulas that quietly stopped working

These show up in every copywriting guide and they're slow death on a modern feed. If your ads are running any of these unmodified, you're paying a hidden tax on every impression.

AIDA, in its full form (Attention, Interest, Desire, Action)

AIDA requires four distinct phases to land. In 2014 you had 12 seconds of attention to deliver them. In 2026 you have 3. The full AIDA structure now reads as a wall of text by the time the reader gets to "Action," and they're gone.

Use instead: PAS compressed. Same arc, half the sentences.

The full Before-After-Bridge

BAB used to win because the "after" gave the reader a vision to walk into. It now loses because the "before" reads as projection. "You're feeling overwhelmed by your ad creative workflow." The reader hears "you don't know me" and scrolls.

Use instead: reverse benefit. Same arc, but the "before" is stated as a fact about the work, not a feeling about the reader.

Feature, Advantage, Benefit (FAB)

FAB was a sales-training formula for a face-to-face conversation where the prospect couldn't leave. On a feed, the reader leaves between feature and advantage. By "benefit" you're talking to nobody.

Use instead: lead with the benefit, support with one feature, drop the advantage line entirely.

The 4 Ps (Promise, Picture, Proof, Push)

The 4 Ps still work on long-form sales pages where the reader is captive. On a 3-second hook, the "Picture" step (an imagined scene the reader is supposed to visualize) burns the only attention you have. By the time you get to proof, the reader is out.

Use instead: single-statistic open. The statistic does the work of "Proof" and "Promise" at the same time, and the reader supplies their own picture.

How to pick the right formula for your specific ad

The wrong formula on a good offer underperforms the right formula on an average offer. Use this short decision tree before you write:

1. Is the audience problem-aware? Yes → PAS compressed or named pain. No → single-statistic open.
2. Is the audience narrow enough to name? Yes → direct address. No → narrow the audience, then write.
3. Is the product easier to describe by what it removes than by what it adds? Yes → reverse benefit.
4. Do you have a verifiable specific number to anchor on? Yes → single-statistic open. No → don't fake one. Pick a different formula.

Run the tree in 90 seconds before you draft. The formula is upstream of the words. Picking the right one is the difference between a copy refresh that moves CTR and a copy refresh that moves nothing.

What "good" looks like in the framework

When we score a creative, Messaging and Persuasion are two of the eight dimensions. The formulas above map directly to both.

A high Messaging score means the copy is specific about who the ad is for and what it does, in language the reader recognizes as their own. A high Persuasion score means the copy contains at least one verifiable claim, one mechanism explanation, or one piece of social proof tied to a named outcome.

Most underperforming copy fails on Messaging first. The brand is talking about itself instead of the reader. The fix is almost always swapping the opener for one of the five formulas above, then letting the rest of the ad anchor to it.

A 9 on Messaging usually does three things:

• Names the reader in the first sentence (role, situation, or recognized pain)
• States the offer as a verb the reader can do
• Avoids any sentence that could appear unchanged in a competitor's ad

A 4 on Messaging usually does one thing: it talks about the product before it talks about the reader.

The shift to make on your next ad

Before you draft, pick the formula. Write its name on a separate line at the top of the brief. If you can't pick from the working five, you don't yet know who the ad is for, and any words you write are going to underperform.

Once you know the formula, the actual writing takes 10 minutes. The decision in front of the writing is what makes the ad work.

Want to see how your current copy scores on Messaging and Persuasion? Score it free on Adverdly. You'll get the score per dimension plus the priority fix at the top, which tells you exactly which sentence in your existing copy is doing the most damage. Replace that one sentence first, then ship the next ad with the right formula chosen before the first word.

The teams that move CTR consistently aren't writing better sentences. They're picking better formulas before they write the sentence.