Good ad copy: a five-point check and the examples that pass it
What separates good ad copy from copy that only sounds good: a five-point check you can run on any ad, real campaigns that pass it, and line-by-line rewrites for Meta.
Marián Cabadaj
9 min read
Good ad copy makes one specific person stop, believe one claim, and do one thing. Everything else on this page is a way to check whether a line does that.
Most ad copy fails the same way. It describes the product instead of the buyer's problem, it makes three claims where one would land, and it ends on "Learn more" because nobody decided what the reader should do. The copy reads fine in a doc. In the feed it gets skipped.
This guide gives you a five-point check you can run on any ad, real campaigns that pass it, and rewrites of weak lines into strong ones. If you want the full writing process, from research to testing, we covered that separately in how to write ad copy. This post is about judging the copy once it exists.
What makes ad copy good?
Good ad copy passes five checks: it names a specific reader, it makes one claim, it proves that claim, it answers the objection the reader already has, and it asks for one action. Copy that fails any of the five usually still gets impressions. It just does not get the click or the sale.
The five-point check
Run every ad through this table before it goes live. One "no" is a rewrite, not a launch.
| Check | The question to ask | It fails when |
|---|---|---|
| Reader | Could a stranger tell who this ad is for from the first line? | The first line would fit any brand in the category |
| One claim | Can you say what the ad promises in seven words or fewer? | The ad lists features, benefits and a discount all at once |
| Proof | What in the ad makes the claim believable? | The only proof is an adjective like "amazing" or "best" |
| Objection | Which doubt stops this reader from buying, and does the ad answer it? | Price, effort or "will it work for me" goes unanswered |
| Action | Is there one next step, and does the reader know what happens after the tap? | The CTA is generic and the landing page says something else |
The check is deliberately short. Frameworks like AIDA describe the order of a whole ad. This table is for the five seconds a reader actually gives you.
What the research says about copy and creative
Nielsen's 2017 analysis of nearly 500 FMCG campaigns found creative was still the most important driver of sales lift from advertising, even as the media share of that lift grew to 36% from 15% over 11 years (Nielsen, 2017).
What this means for a DTC brand: better targeting does not rescue a weak message. When two brands buy the same audience on Meta, the words and the image decide who gets the sale.
Space is the other constraint. Meta's ads guide recommends 50 to 150 characters of primary text and a 27-character headline for Facebook Feed image ads (Meta Ads Guide, 2026).
What this means for a DTC brand: the five checks have to fit in about two sentences and a short headline. If your proof needs a paragraph, it belongs in the image or on the landing page.
Our articles are still written by HUMANS! Get human written articles in your Google feed.
Add as preferred sourceAd copy examples: which ones are worth copying?
The best ad copy examples are the ones that do one job so well you can lift the move into your own ads. The list below sorts real campaigns by the job their copy does: stop the scroll, prove the claim, or kill the objection. For each one we name the brand, the year and the move to steal.
Hook: Dollar Shave Club, "Our Blades Are F***ing Great" (2012)
Dollar Shave Club launched in March 2012 with a YouTube video where founder Michael Dubin says the whole offer in the first lines: for a dollar a month, the company sends high-quality razors to your door. The title alone did the work of a headline. Entrepreneur reported that the video cost $4,500 to make and that about 12,000 people signed up in the first 48 hours (Entrepreneur, 2012).
Why it works: the reader, the claim and the price arrive before the joke does. Humor kept people watching, but the offer was clear in the first sentence.
The move to steal: say the offer in plain words before you get clever. Our advertising hooks examples break down ten opening patterns if you need more options.
Hook: Spotify, "Thanks 2016, it's been weird" (2016)
Spotify's in-house team ran billboards in 14 markets built from real listening data, such as a line addressed to the 3,749 people in the UK who streamed "It's The End Of The World As We Know It" on the day of the Brexit vote, as The Drum reported (The Drum, 2016).
Why it works: a number nobody else could know is a hook and a proof point at once. The reader believes the brand because the line could only come from Spotify.
The move to steal: put one number from your own data in the first line: orders shipped, repeat rate, the most common question in your inbox.
Proof: Rolls-Royce, "the loudest noise comes from the electric clock" (1958)
David Ogilvy's headline for Rolls-Royce read: "At 60 miles an hour the loudest noise in this new Rolls-Royce comes from the electric clock." The rest of the ad was a numbered list of similar facts about how the car was built and tested.
Why it works: the ad never says "quiet" or "luxury". It gives one checkable detail and lets the reader draw the conclusion.
The move to steal: replace every adjective in your copy with the fact that made you want to use it. "Fast shipping" becomes "ordered by 2pm, at your door the next day", if that is true.
Objection: Avis, "We try harder" (1962)
In 1962 Doyle Dane Bernbach wrote a campaign that admitted Avis was only No. 2 in car rentals, behind Hertz, and turned that into the reason to rent from it: when you are No. 2, you try harder. The ads listed small proofs like clean ashtrays and short queues.
Why it works: every reader already knew Avis was smaller. The copy said the doubt out loud and gave it a reason that favored the buyer.
The move to steal: write down the one objection your reviews mention most, then make it the headline. Problem-solution ads use the same structure.
Objection: Domino's, "Pizza Turnaround" (2009)
Domino's launched its Pizza Turnaround campaign in late December 2009. The ads showed customers on camera calling the old crust cardboard, then showed the company's chefs remaking the recipe. In its first-quarter 2010 results, Domino's reported domestic same-store sales growth of 14.3% (Domino's Pizza, SEC filing, 2010).
What this means for a DTC brand: admitting a known flaw is a proof tactic when you can show the fix. Without the fix, it is only a confession.
The move to steal: quote the harshest real review you have answered, then show what changed. Our guide to social proof ads covers which reviews to use and how to keep them believable.
Restraint: Patagonia, "Don't Buy This Jacket" (2011)
On Black Friday 2011 Patagonia ran a full-page ad in The New York Times with the headline "Don't Buy This Jacket", above its own best-selling fleece, as the company explains in its story of the ad (Patagonia, 2011).
Why it works: on the loudest sales day of the year, the one ad telling you not to buy got read. The copy also proved the brand's repair-and-reuse stance with a real cost, its own sales.
The move to steal: this only works when the restraint is true to the brand. A discount brand cannot borrow it. A brand with a real repair program, a real guarantee or a real "you do not need two" story can.
How do you turn weak ad copy into good ad copy?
Rewrite one line at a time, in this order: the headline, then the first line of primary text, then the CTA. Each line has a separate job, so each gets its own fix. The rewrites below are illustrative, written for an imaginary skincare brand to show the move, not taken from a live ad.
| Line | Weak version | Stronger version | What changed |
|---|---|---|---|
| Headline | The best serum for your skin | Fades dark spots in 8 weeks | One claim with a time frame, under 27 characters |
| Primary text | Our serum uses premium ingredients to give you amazing results. | Dark spots after summer? 2% tranexamic acid, applied twice a day. Most customers see a change by week 8. | Names the reader, gives the active ingredient, sets the expectation |
| Proof | Loved by thousands | 4.7 stars from 2,140 reviews | A countable number instead of an adjective |
| Objection | Suitable for everyone | Fragrance-free. Safe next to retinol. | Answers the two questions in the review inbox |
| CTA | Learn more | Get the 30-day size | Says what the reader gets after the tap |
Every number in the stronger column has to be true for your product before it ships. A made-up review count is worse than "Loved by thousands", because it is checkable.
Rewrite the headline first
The headline is the line most people read and the one Meta gives the least room. Cut it to one claim. If it needs a comma, it has two. The hook generator gives you twenty starting points when you are stuck on a blank line.
Then the first line of primary text
Only the first line or two show before "See more". Lead with the reader's situation, not your brand name. Keep the brand name for the image or the headline.
Then the CTA
Match the button to the landing page. If the button says "Get the 30-day size", the page has to open on the 30-day size, not on the homepage.
How do you know your ad copy is good, not just clever?
You know from the numbers, not from the room. Good ad copy raises click-through rate without lowering conversion rate. A clever line that doubles clicks from people who never buy is a worse ad, because you pay for every click.
Read CTR next to conversion rate
Put the clicks, impressions and purchases for two versions into the calculator below. Look at CTR and click-to-conversion rate together.
Your numbers
Add this to see what share of your clicks turned into something, which is the number that decides whether a high CTR was worth having.
Your click-through rate
CTR
—
Impressions per click
—
Click to conversion
—
The formula
CTR = (clicks / impressions) x 100
CTR Calculator: the full tool, with the benchmarks and the FAQ
If version B wins on CTR and loses on conversion, the copy promised something the page did not deliver. Fix the page or the promise, not the targeting.
Change one line per test
Test the headline against the headline, with the image and the primary text held the same. Change three lines at once and you will not know which one moved the number. Our ad creative testing guide covers how to set up and read those tests.
How Selzee finds the copy that already works
The examples above are famous because they worked once, for one brand. The faster route is to look at what is working in your own market right now, and that is the part Selzee does.
Selzee reads what your competitors run in the Meta, TikTok, Google and Pinterest ad libraries, next to what your customers say in reviews and on Reddit, and recreates the winning ads for your brand, with your products, logo, fonts and colors. Every ad comes with the angle and the insight it is built on, so you can run the five-point check on the reasoning and not only on the words. You edit a headline by asking, approve the ones you like, and ship them in one click.
For a single line, the free ad copy generator writes headline and primary text options from your product and your audience.
Good ad copy: FAQ
What is good ad copy?
Good ad copy is copy that makes a specific reader stop, believe one claim and take one action. In practice it names the reader, makes one promise, proves it with a fact, answers the main objection and ends with a clear call to action.
What are the best ad copy examples to learn from?
The most useful ad copy examples are the ones that do one job well: Dollar Shave Club (2012) for a hook that states the offer, Rolls-Royce (1958) for proof by one fact, Avis (1962) and Domino's (2009) for answering an objection head on, and Patagonia (2011) for restraint that fits the brand.
How long should Facebook ad copy be?
Meta recommends 50 to 150 characters of primary text and a 27-character headline for Facebook Feed image ads. Longer copy can work for considered purchases, but the first line still has to carry the hook, because the rest sits behind "See more".
How do I test whether my ad copy is good?
Run two versions that differ in one line, with the same image and audience, and compare click-through rate and conversion rate together. The version that wins on both is the better copy. A win on clicks alone usually means the copy overpromised.
Should ad copy be funny?
Humor helps when the offer is clear before the joke. Dollar Shave Club's launch video was funny, but the price and the product were in the first sentence. If the reader laughs and still does not know what you sell, the joke cost you the sale.
The short version
Good ad copy names one reader, makes one claim, proves it with a fact, answers the doubt and asks for one action. Steal the move from a proven ad, rewrite one line at a time and let CTR and conversion rate decide.
If you want the winning angles in your market found and turned into ads for you, see how Selzee works on your own brand.