Paid Social Ad Copy Written to the Truncation Point

Meta shows roughly the first 125 characters of primary text in a mobile feed before it truncates, and on Reels the caption collapses nearer 55. The claim goes in that space or it does not get read. In most accounts we take over, a brand name and a warm-up clause are sitting there instead, and the sentence that would have earned the click starts after the fold. What we write: Primary text, headlines and descriptions for Meta. Captions and short scripts for TikTok in-feed and Spark Ads. Single-image and document ad copy for LinkedIn, including thought-leader posts that run from a named person's profile. Reddit copy written by somebody who reads the subreddit. Each concept ships as a small set of texts arguing the same claim different ways, not as a spread of rewordings. Platform mechanics we write to: Advantage+ creative accepts five primary texts, five headlines and five descriptions. We usually supply three and two. The delivery system converges on whatever earns the cheapest early engagement, which is not always what converts, and filling every slot buys a faster convergence on a worse combination while making the concept unreadable in reporting. On TikTok, Spark Ads run through the creator's own handle, so brand copy written over the top breaks the native read that made the format worth using in the first place. On LinkedIn, intro text truncates near 150 characters on desktop and 100 on mobile, and thought-leader copy has to survive being read by the named person's own colleagues. Instrumentation: Every variant carries an identifier in the ad name and in utm_content. Without that, the analysis you will want in March cannot be run at all, and we have inherited too many accounts where six months of testing is unreadable because the naming drifted in week three. We report against hook rate and click-to-landing behaviour rather than platform-attributed conversions alone. On calling winners: A headline comparison needs somewhere near 100 conversions per arm before the difference means anything. Below that we will tell you the test is inconclusive and either let it run or kill both. We would rather say that than hand over a winner that is noise, because next quarter's creative gets built on top of it and the error compounds quietly. What is not included: No media buying, no account management, no bid work. No landing page copy under this service, though if the page contradicts the ad we say so, once, in writing, and it changes the recommendation. No regulated claims in health or financial services drafted before your compliance function is in the loop. We want their thresholds before writing, not their objections after. Who this is not for: Anyone whose real problem is the offer. If the product is priced against a competitor shipping free next-day and nothing addresses the difference, better copy raises click-through and lowers conversion rate, and the account gets worse in a way that takes a quarter to become visible. We will say this on the first call. Some briefs end there, and that is the correct outcome for both of us. It is also not for accounts that need a high volume of near-identical variants to feed a broad testing programme. That is a legitimate way to run an account and there are teams built for it. This desk is not one. How it runs: We read the last ninety days of creative in your ad account along with the comment sections underneath it, take two calls with whoever handles support tickets, and draft against the objections that appear in both. First drafts inside a week. You get the argument behind each line, so your own team can write the fourth round without us.