Creative Strategy and Concept Development for Paid Media

Most accounts asking for more creative do not have a volume problem. Open the library and twenty-odd live assets collapse into three claims in different typefaces, which is exactly why the fourth one performs like the first. The first thing we do is map every running asset to the single claim it makes, and the map is almost always shorter than the client expects it to be. The claim map: Every live and recently retired asset, tagged with the claim it argues, the objection it answers, the format it uses and the result it got. Alongside that, a read of competitors' longest-running assets in the Meta Ad Library. Run time is the only public proxy for what is working, since spend is invisible, and an asset live for seven months tells you something that a screenshot of a clever one does not. The output is the list of claims your category makes, the ones nobody is making, and a view on which of those you can actually support. Where claims come from: Five to eight calls with the people who hear objections all day: support, sales, onboarding, whoever answers the community inbox. Review mining across your own listings and two competitors', reading the three-star reviews rather than the one and five star ones, because the middle is where people explain themselves. Claims that survive this are phrased in the customer's words and are frequently less flattering than the ones in the brand deck. That is the point of doing it. Concept briefs: Each concept is a page: the claim, the person who disbelieves it, the evidence that appears in frame, what happens in the first three seconds, the format it has to be produced in, and the condition under which we agree it failed. That last line is written before anything is made. Without it a concept never dies, it just gets a new edit and another month. How we read results: Hook rate, three-second views over impressions, separates a first-frame failure from an offer failure, and those two get confused constantly in review meetings. Below roughly twenty-five percent on Reels, the frame lost before the argument started and no rewrite of the copy repairs it. Hold rate and click-to-landing behaviour do the rest. We do not claim incrementality we cannot defend; where the only evidence is a platform-attributed number, we say that rather than dress it up as proof. What is not included: We do not buy media and we are not agency of record. We do not produce the assets under this engagement. Shoots and copy are separate pieces of work and you are free to take the briefs elsewhere, which some clients do and we would rather they said so early. No brand platform, no positioning architecture, no six-week discovery before anything runs. No workshop for a room that has not yet agreed what the product is for. That meeting belongs to your leadership and running it as a creative exercise wastes a good day. Who this is not for: Teams spending under about thirty thousand euros a month across paid social and search. There is not enough throughput to test concepts against each other, results stay ambiguous for a quarter, and the sensible move is to run two clear ideas properly rather than pay for a strategy function to watch noise. It is also not for anyone who needs the answer to be flattering. The claim map has embarrassed a brand or two, and the version that embarrasses is the version that changes what gets made. Cadence: On retainer, concepts arrive in cycles rather than to a monthly quota, because a cycle ends when the running work has told us something, not when the calendar turns. In practice that lands between three and six weeks. Retirement is called on frequency and click decay together, not on a rota.