A business owner once asked us why his ad, with an objectively better offer than his competitor's (same product, 15% cheaper), was performing worse than that same competitor's ad. The answer wasn't in the offer, it was in how it was written. A good price told badly convinces less than a normal price told well, and that's a hard thing to accept when you're convinced your product "sells itself".
Ad copy isn't literature and it isn't a place to show off vocabulary. It's a tool with one very specific job: getting someone who didn't know you three seconds ago to understand what you're offering, why it matters to them specifically, and what they need to do right now to get it. When an ad doesn't convert, it's almost always failing at one of those three points, not at grammar or how clever it sounds.
The headline: the one thing that decides whether anyone keeps reading
The headline has one job, and only one: stop the scroll. Not sell, not explain, not fully convince, just get the person to pause a second longer than they were going to. The best-performing headlines usually do one of three things: name the person's problem directly ("Is your inbox filling up with spam instead of real customers?"), promise a concrete, measurable result ("Cut your electricity bill by 30% without changing your setup") or spark a specific curiosity tied to their situation ("The mistake 80% of small businesses make when choosing an accountant"). What almost never works is the generic, brand-centred headline: "Welcome to [Company name]" stops nobody, because it says nothing about the person reading it.
The body: from general to specific, never the other way round
Once you've got attention, the body of the ad has to answer, in order, three questions the reader is asking without realising it: "is this for me?", "why should I believe you?" and "what exactly do I get?". The most common mistake is starting by talking about the company (who we are, how many years we've been around, how many awards we've won) when at that point the reader hasn't yet decided whether the topic interests them. That kind of information, which does matter, belongs after you've connected with the person's problem or desire, not before.
A real rewrite example: a language academy had an ad that opened with: "We are an academy with 20 years of experience and native-speaking teachers". We changed it to: "You've been saying for years that this is the year you finally learn English. Here's a way to do it that actually fits your schedule." Cost per click dropped 22% in the first two weeks, without changing anything else about the campaign. The difference wasn't the quality of the service, it was who the text spoke to first.
Features versus benefits: the most expensive mix-up
A feature is a fact about your product ("our mattress has seven differentiated firmness zones"). A benefit is what that feature means for the person's life ("you sleep without tossing and turning because every part of your body gets the support it needs"). Weak ads stop at the feature and expect the reader to make the translation themselves. Ads that convert make that translation for them. You don't need to remove features entirely (sometimes they're the proof that backs up the benefit's credibility), but the benefit should always come first, with the feature following as support.
Social proof: why "trust us" isn't enough
Saying "we're trustworthy" convinces nobody, because anyone can say it. What does convince is a concrete figure ("over 2,000 small businesses already manage their accounting with us"), a real customer's direct quote, or a verifiable data point (a rating with a visible number of reviews). The more specific and less generic the data, the more credibility it adds. "Our customers are happy" is worth nothing; "94% of our customers renew the service the following year" is.
The call to action: why "learn more" is the weakest option there is
The ad's final button or line has to tell the person exactly what will happen if they click, and at what level of commitment. "Learn more" is vague and low-commitment, which sounds safe but in practice attracts clicks from undecided people who then don't convert. Phrases like "Book your free 15-minute call", "Download the PDF guide" or "See prices" set a clear expectation, and that filters better who clicks: fewer total clicks, but from more qualified people, which almost always ends up cheaper per sale in the end.
Text length: the myth that "shorter is always better"
There's no universal ideal length. There is a more useful rule: the text should be as long as it takes to answer the person's real doubts, and not a line longer. An impulse-buy, low-priced product (a 20-euro t-shirt) needs little text because the decision is fast and low-risk. A high-priced service with a considered decision behind it (a kitchen renovation, management software) usually performs better with more text, because the reader needs more arguments before taking the step of getting in touch. Shortening a complex product's copy doesn't make it "more effective", it just makes it incomplete.
A practical exercise: how to take apart one of your own ads in five minutes
Before writing a new ad, it's useful to run this quick check on one you already have live. First, cover everything except the headline and ask yourself whether, from that line alone, you'd understand what it's about and feel it was speaking to you. Second, read the body copy and mark the exact line where something about the company first appears (name, years of experience, awards); if that line shows up before the third sentence, the ad is probably starting in the wrong place. Third, look for the first concrete figure or data point in the text; if there isn't one, the ad is leaning entirely on generic claims any competitor could copy without changing a word.
Applied to a real case: an accounting firm's lead-generation ad read: "We are experts in tax and payroll management for small businesses, with a friendly, professional team." Running the check, there was no distinct headline (it opened straight with "we are"), the first mention of the company appeared in the first sentence, and there wasn't a single figure anywhere. The rewrite, keeping the same service and the same price: "Do you know if you're overpaying on your quarterly taxes? 60% of the small businesses we reviewed last year were." Cost per lead dropped 31% in the first month with the new version, with nothing else about the campaign changed.
How to adapt the same copy to different formats without rewriting from scratch
Once a copy angle proves it works, it makes sense to reuse its logic across different formats instead of inventing a new message every time. A winning feed ad headline can become a short video's opening spoken line; the social proof argument that worked in the body copy can become a product carousel's caption. Keeping the same core idea while adapting the format not only saves time, it also reinforces the message through consistent repetition instead of diluting it with a different message every time the format changes.
Words and phrases worth watching closely
There are certain words that appear in so many mediocre ads that they've lost almost all their persuasive power through sheer overexposure: "unique", "revolutionary", "industry leader", "guaranteed quality". It's not that they're forbidden, but using them without concrete proof behind them (a data point, an example, a figure) turns them into noise the reader has learned to tune out. The alternative isn't avoiding them entirely, it's always pairing them with the proof that backs them up: not "guaranteed quality", but "if you're not satisfied within the first 30 days, we'll refund you, no questions asked".
Likewise, it's worth watching the use of unanchored superlatives ("the best", "the fastest") in regulated sectors like health, finance or education, where beyond losing credibility with the reader, that kind of unsubstantiated claim can also cause problems with the platforms' own advertising policies, which increasingly scrutinise this type of promise.
Comparison: long copy versus short copy for the same product
To illustrate how the right length depends on context rather than a fixed rule, it helps to compare two real ads for the same home insurance product. The short version, aimed at an audience that already knew the brand and just needed the final nudge ("Home insurance from €8/month. Sign up in 3 minutes."), worked well in remarketing to recent website visitors, with a low cost per sign-up. The long version, fully laying out exactly what the policy covers, what questions typically come up, and an honest comparison of what it doesn't cover, worked much better with a cold audience that had never heard of the brand, because those people needed to understand the product before trusting it. Both versions were correct; the mistake would have been using the short one with a cold audience or the long one with the remarketing audience.
Frequently asked questions
Should I always use the same copy structure across all my ads?
No, it's worth testing variations (different headlines, different argument order) because what works depends on the audience and the platform. What should stay fixed is the underlying logic: connect first, then argue, then ask for the action.
Do emojis and exclamation marks help or hurt credibility?
It depends on the sector and brand tone. In more informal categories (young fashion, leisure) they usually help catch attention in the feed. In sectors where trust and seriousness carry more weight (financial services, health, B2B), overusing them usually hurts credibility.
Is it better to write the copy myself, since I know the business, or hire a copywriter?
The ideal is a combination: whoever knows the business provides the real data, the common customer objections and the right tone, and a copywriter provides the structure and the ability to turn that into text that actually works. Copy written solely by whoever knows the business often assumes information the reader doesn't have.
How many copy versions should I test at once?
Between three and five variants per ad set is usually a good starting point: enough to detect real differences without splitting the budget so thin that none of them get enough data to learn from.
Should the copy be different for Google Ads than for Meta Ads?
Yes, quite a lot. On Google Ads the user is already actively searching for something, so the copy can be more direct and offer-focused. On Meta Ads the user wasn't searching for anything, they're scrolling their feed, so the copy first needs to capture attention and build interest before talking about the offer.
How do I know if my copy is the problem or if it's the ad's targeting?
If the ad gets a lot of impressions but few clicks, it's usually a copy or creative problem (it isn't stopping the scroll or generating interest). If it gets a good click rate but few conversions afterwards, the problem is usually more on the landing page or the targeting (it's reaching people who click out of curiosity, without real buying intent).