Ad Copy
Ad copy is the most constrained writing in marketing — you have 3 seconds and sometimes 30 characters. Here's how to use them.
Two ads. Same audience. Same budget. One got 10× the results.
A fitness app was running Meta ads to acquire new users. Both ads targeted the same audience: women aged 25–40 interested in fitness and health.
Ad A:
"The #1 Fitness App for Women. 500+ workouts, meal planning, progress tracking. Download free today."
Clicks: 840. Cost per install: $4.20.
Ad B:
"I was doing everything 'right' — and the scale wouldn't budge.
Then I stopped obsessing over the number and started tracking how I felt.
[App name] helped me lose 14kg. But more than that — I stopped dreading Mondays.
Free for 7 days. No credit card."
Clicks: 6,200. Cost per install: $0.57. (illustrative scenario — specific results vary widely by product, creative, audience, and targeting)
Ad A listed features. Ad B told a story, named a feeling, and led with empathy before asking for anything. Both are truthful. One is copy; the other is advertising.
The gap between average ad copy and great ad copy is one of the highest-leverage skills in marketing. A 10× improvement in CTR (click-through rate) means you pay 10× less for each customer you acquire. That's the difference between a business that scales and one that bleeds cash.
What ad copy has to do in 3 seconds
Every paid ad — regardless of platform — competes against everything else in a person's feed, inbox, or search results at that moment. You don't get a polite audience. You get distracted people who didn't ask to see your ad.
Your copy has approximately 3 seconds to do this:
Stop → Connect → Compel. Every word in your ad should serve one of these three jobs.
The 4 ad copy formats
Format 1: Problem-first
[Name the pain immediately. No preamble.]
Spending 3 hours on reports that should take 20 minutes?
[Product] automates the boring parts. You handle the insights.
Free 14-day trial. No credit card.
Works because: it opens with something the reader is experiencing right now. If they're not having this problem, they scroll past — and that's fine. You're filtering for the right audience.
Format 2: Outcome-first
[Lead with the result, make it vivid and specific.]
Sarah closed $47,000 in new clients last month. She works 30 hours a week. She has no team.
[Product] is how she does it.
Works because: specific results create both desire and credibility. The reader thinks "if she can, maybe I can."
Format 3: Story/identification
[Make the reader recognise themselves.]
I spent two years applying to jobs with the same CV. 200+ applications. 4 interviews.
Then I changed three things about how I wrote my experience section. The next month I had 11 interviews.
Here's what changed:
Works because: people stop scrolling when they recognise their own situation. The cliffhanger ("here's what changed") creates the itch.
Format 4: Direct offer
[For people already in market — no story needed.]
[Product]: Free 14-day trial. No credit card. Cancel any time.
Used by 40,000 marketing teams. Join them.
Works because: for search ads and retargeting (people already aware of the category or brand), the decision is made — they just need the offer and a low-friction CTA.
There Are No Dumb Questions
"Which format should I use?"
It depends on the audience temperature. Cold audiences (never heard of you) need Stop + Connect — problem-first or story formats. Warm audiences (visited your site, engaged with content) can go straight to outcome or offer formats. Retargeting (people who already visited) needs a direct offer + risk removal.
"Do longer ads outperform shorter ads?"
On Facebook/Instagram: longer copy often outperforms for cold audiences because it has room to build the case. But the first line must work as a hook — most people only see the first 1–2 lines before "See more." On Google Search: you're constrained to headlines of 30 characters and descriptions of 90 characters — every word has to earn its place. On TikTok/Reels: the spoken script is the ad — visual and audio together.
Platform-specific constraints
Every ad platform has different technical limits. Know them — they shape what's possible:
| Platform | Headline limit | Body copy | Key constraint |
|---|---|---|---|
| Google Search Ads | 30 chars (×3 headlines) | 90 chars (×2 descriptions) | Must be direct — people are actively searching |
| Meta (Facebook/Instagram) | 40 chars primary | 125 chars before truncation | Hook must work in first 2 lines; image/video stops the scroll |
| LinkedIn Ads | 70 chars headline | 150 chars intro | Professional tone; targeting is by job title/industry |
| TikTok Ads | First 3 seconds of video | Spoken script | Video hook is the ad — text is secondary |
| Display/Banner | 25 chars | 90 chars | Extremely short — one benefit, one CTA, nothing else |
Specifications as of early 2025 — ad platforms update character limits periodically. Verify current specs in each platform's ads help centre before building campaigns.
Google Search Ads: the constraint is a gift
With 30-character headlines and 90-character descriptions, every word counts. This forces a clarity that longer formats often lack. The best search ad copy:
- Matches the search query — if someone searches "email marketing tool," your headline says "Email Marketing Tool"
- States the differentiator — what makes you better/different from the other ads on the page
- Contains the CTA — "Try Free 14 Days" or "Get a Quote" in the description
- Answers the searcher's intent — informational queries need different copy than transactional ones
Using AI for ad copy: Prompt: "Write 10 Facebook ad variations for [product/audience/goal]. Use these formats: 2 problem-first, 2 outcome-first, 2 story/identification, 2 direct offer. For each ad, write the primary text (under 125 characters before cut-off), headline (under 40 characters), and CTA button text. Vary the emotional angle — some should use fear of missing out, some desire for gain, some relief from pain."
Generate 10, eliminate the obvious losers, test the remaining 3–4 with a real audience.
Ad Copy by Format
25 XPThe hook: stopping the scroll on social
For social ads (Meta, TikTok, Instagram), the creative (image or video) stops the scroll. But the first line of copy either keeps them or loses them.
The 5 hooks that work on social:
- The direct call-out: "Remote workers: this is for you." — Names the audience, filters immediately.
- The provocative question: "What if your biggest productivity problem isn't laziness?" — Creates cognitive dissonance.
- The specific number: "I added 8 hours to my week. Here's exactly how." — Specificity creates credibility.
- The relatable moment: "3pm on a Tuesday. I've 'worked' for 6 hours. I've accomplished almost nothing." — Recognition triggers a dopamine hit.
- The bold claim: "Most productivity apps are making your focus worse." — Counterintuitive statements stop people mid-scroll.
Writing the CTA for ads:
| Weak CTAs | Strong CTAs |
|---|---|
| "Click here" | "Try it free for 14 days" |
| "Learn more" | "See how it works in 60 seconds" |
| "Buy now" | "Get your first month for $1" |
| "Sign up" | "Join 40,000 remote workers who fixed their focus" |
The CTA should reduce risk ("no credit card"), create specificity ("14 days"), or add social proof ("join 40,000"). Often all three.
Testing: the only way to know what works
In ad copy, your instincts will fail you. The ad you think will win often loses. The one you're less excited about often wins by 2×.
This is why testing is non-negotiable:
Never run one ad. Always run at least two — preferably three or four. The cost of testing is low; the upside of finding a winner is massive.
What to test (in order of impact):
- The hook / first line (biggest lever)
- The format (problem vs. outcome vs. story)
- The CTA (button text and offer)
- The creative (image vs. video, lifestyle vs. product)
- The headline
Test one variable at a time if you want clean data. Test multiple simultaneously if you need speed and are willing to make educated guesses about which variable drove the result.
Write a Test Set
25 XPFull Ad Campaign Draft
50 XPBack to the two fitness ads
Ad A listed features because whoever wrote it was thinking about the product. Ad B named a feeling because whoever wrote it was thinking about the person. "The scale wouldn't budge" is not a product description — it's a mirror held up to the reader's lived experience, and the moment someone recognises themselves in an ad, they stop scrolling. The 10× gap in results came from three compounding differences: Ad B was specific ("14kg" rather than "results"), it addressed the emotional pain point beneath the surface problem (dreading Mondays, not just the number on the scale), and its CTA removed the only remaining objection ("free for 7 days, no credit card"). Each of those elements alone would have lifted performance. Together, they turned a mediocre cost-per-install into one that made the campaign scalable.
Key takeaways
- Ad copy has 3 seconds: Stop the scroll, connect with the audience, compel a click. Every word serves one of these three jobs.
- Four formats: Problem-first (cold audiences), outcome-first (desire-driven), story/identification (recognition triggers), direct offer (warm audiences).
- Platform constraints are gifts — they force clarity. Know the character limits before you write.
- The first line is everything on social — most people see only the hook before deciding to "see more."
- Always test at least 2–3 variations. Your instincts about which ad will win are wrong more often than you think. Let the data decide.
Knowledge Check
1.A cold-audience Facebook ad for a meal planning app opens with: 'Our app has 500+ recipes, automatic grocery lists, and nutritional tracking. Try it free.' Why is this likely to underperform compared to an ad that opens with a relatable struggle?
2.A Google Search ad headline reads: 'Best Digital Marketing Services for All Your Business Needs.' What are the two most important problems with this headline?
3.A marketer creates one ad and runs it for a month. It performs below expectations. What should they do?
4.Which CTA is strongest for a retargeting ad targeting people who visited a pricing page but didn't sign up?