A client once asked me to rewrite forty ad headlines because performance had flattened. I read through their landing page first. The offer was confusing, the value proposition was buried under three paragraphs of company history, and the call to action competed for attention with a chat widget popping up every eight seconds. The headlines were fine. Nothing else was.
This happens more than people want to admit. Teams treat copy as the lever they can pull without touching anything structural, because rewriting ten headlines feels faster and cheaper than fixing a landing page, a checkout flow, or an offer that doesn't actually match what the audience wants. Copy gets blamed because copy is easy to blame. It's also usually not the actual problem, and there's a structural reason the misdiagnosis repeats: headlines are the one asset the marketing team can change without a ticket, a stakeholder, or an engineering sprint. When the only tool you can swing without permission is a hammer, every performance dip becomes a nail, and the account fills up with headline tests while the actual leak sits untouched two clicks downstream.
Run the funnel math before the rewrite
Here's the diagnostic I run before accepting any copy brief, and it takes ten minutes. Lay out the funnel's three numbers: click-through rate, landing conversion rate, and close rate, next to honest category benchmarks. Then ask which number is furthest below its benchmark, because that gap is where the money is. A campaign with a healthy CTR and a two percent landing conversion does not have a words-in-the-ad problem; the words did their job, someone clicked, and then nine out of ten interested visitors left. I've seen brilliantly written Meta ads drive strong click-through into pages converting at that two percent, and the instinct on the account was always to test new ad variations rather than ask why the interested people were leaving. Copy testing in that situation is rearranging the storefront window while the cash register is broken.
Good ad copy does one job: it sets an accurate expectation and gets someone to take the next step. It cannot fix a weak offer, and it cannot compensate for a landing experience that contradicts the promise made in the ad. In fact the better the copy, the more expensive the downstream failure, because great headlines buy more clicks into the same broken page. The page after the click is where flat accounts usually get un-flattened, at a fraction of the cost of a creative overhaul.
When it is the copy: specificity wins
When I do sit down to write or direct copy, the brief I care about most isn't tone or brand voice guidelines, even though those matter. It's specificity. "Save money on your energy bill" converts worse than "average customer saves $340 a year, verified across 40,000 accounts," every time I've tested the two side by side. Vague benefit language reads as marketing. Specific, checkable claims read as fact, and audiences increasingly reward the second one with their attention and their trust. The mechanism is simple: twenty years of advertising taught every consumer that adjectives are free and numbers are auditable. A claim with a number and a source is making itself falsifiable in public, and readers correctly interpret that willingness as confidence. The vague version isn't just weaker writing. It's a tell.
Specificity has a second job now that most copy briefs haven't caught up with: with broad targeting, the copy is the audience filter. The specific claim qualifies the reader; "for controllers closing multi-entity books" repels everyone who isn't one, which is precisely the point. Generic copy under broad delivery isn't just bland, it's untargeted, and the media bill pays for the difference.
The other thing I'd tell any team writing performance creative right now: write for the platform's actual behavior, not for how you wish people consumed content. Nobody reads a Facebook ad the way they read a magazine spread. They scroll past it in half a second unless the first three words earn the next three. That means front-loading the specific, provable claim instead of building up to it, even if it feels less elegant as writing. The elegance the format rewards is compression, not build.
Content written for search and AI answer engines follows a similar rule with different mechanics. The content that actually gets used, whether that's ranking well or getting cited inside an AI-generated answer, tends to be the content that states a clear, specific, checkable fact early and structures the rest around supporting it. Buried leads and slow builds work in essays. They don't work in advertising, and increasingly they don't work in AI-facing content either. One writing discipline now serves three masters, and the discipline is the same: the checkable claim, stated first.
If your creative has flattened and the first instinct on your team is to rewrite the headlines, take one detour first. Look at what happens after the click, and run the three-number funnel check before anyone opens a doc. That's usually where the real story is, and it's the first place we look on copywriting and content engagements at Echelon before touching a single word of ad copy.