Le Rendez-vous E-commerce · Danilo Mura Duchesnes

Automatiser votre stratégie créa avec l’IA : le guide d’une agence experte Meta Ads

·41 min·3 clips

Every ad has exactly three testable parts — hook, hold, and CTA — and one weak metric in any of them tells you precisely which hypothesis to test next.

1. Le Rendez-vous E-commerce covers DHS Digital's updated creative strategy process for Meta Ads in 2025, with host Danilo Duchenne joined by Elisabetta and David, both strategists at the agency. 2. Elisabetta has appeared in multiple previous episodes on creative strategy; David is primarily known for media buying episodes; this is their first episode together. 3. The episode's central argument is that AI tools (ChatGPT, Claude, Gemini, Atria) can be integrated at every stage of a structured creative workflow to compress research and ideation time without sacrificing quality. 4. DHS Digital's process has eight stages: product/persona/brand analysis (onboarding), competitive research, micro analysis of top performers, macro account analysis, ideation briefing, content production, internal and client validation, launch, and creative analysis. 5. Previously, creative strategy responsibilities were split informally between media buyers and the creative studio with no unified working method; the new process documents each stage with dedicated prompts and tools. 6. The onboarding phase creates a product file covering USP, seasonality, price, sales funnel, and benefits, and a persona file covering needs, aspirations, problems, convincing factors, fears, and frustrations — both used to ensure ads address the specific target audience rather than a generic one. 7. Competitive research uses Atria to analyze 5-10 competitors' MediaMix, top personas, USPs, hooks, ad copy, and landing pages, identifying formats and angles that have spent heavily (indicating performance) as well as gaps the client's brand could exploit. 8. Customer review and ad comment scraping feeds raw social proof text to ChatGPT, asking structured questions about pre-purchase problems, buying motivations, benefits, objections, and post-purchase life changes — building a voice-of-customer brief from publicly available data. 9. Reddit is used alongside review scraping to identify the exact vocabulary the target audience uses when describing their problems, ensuring ads use language the audience already recognizes. 10. Micro analysis evaluates top-performing creatives — defined as any creative that generated 10 purchases at target CPA and ROAS — against storytelling KPIs (hook rate, hold rate, CTR, hook click rate, scroll-stop rate) to identify fixable weak points even in winning ads. 11. Macro analysis uses Atria's reporting features to identify the best-performing formats, creators, axes, concepts, emotions, and funnel stages across the entire account, generating a combination map for new production. 12. The ideation brief is generated by feeding the full research stack (product, persona, competitive, macro, micro analyses) plus reference files and a master prompt to ChatGPT, which returns a Mediacrea brief with concept name, content description, 3-4 marketing axis variations, image requirements, social proof elements, and color palette. 13. Reference files — examples of desired brief outputs — are included in the ChatGPT prompt to standardize the format of responses across all media buyers in the agency. 14. The Hook-Hold-CTA framework decomposes every creative into its first three seconds (hook), content body (hold), and call to action (CTA); when a storytelling metric falls below account average, one specific hypothesis is formed for the corresponding section and tested in isolation. 15. Brands Danilo identifies as having mastered this dissection process — 900K (900.cash) and Boku — achieve consistent scale by accumulating validated elements across many hypothesis-test cycles until they have a verified assembly of known-working parts. 16. AI-generated ideation briefs are reviewed and refined by the creative strategy team before handoff to the creative studio, ensuring that volume-generating AI outputs are elevated to higher quality through human judgment. 17. ChatGPT 4.0 image generation is used to turn ideation brief descriptions into rough directional visuals that the team can show the creative studio — not as production-ready assets — because text placement and consistency remain unreliable in generated images. 18. The episode is technical and process-heavy, mixing screen-share demonstrations (visible to YouTube viewers) with verbal explanations, making it more detailed for video than audio listeners. 19. Performance marketers and creative strategists at agencies or in-house teams who want to build a repeatable, AI-augmented creative workflow will find the most value. 20. Listeners seeking beginner Meta Ads guidance or those with no existing creative production process will find the episode assumes operational maturity they may not yet have.

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Le Rendez-vous E-commerce