Le Rendez-vous E-commerce · Danilo Mura Duchesnes

Winning ads Meta : notre méthode pour trouver les variantes qui scalent

·19 min·1 clip

Advertisers are testing many new concepts but failing to create enough variants from their winning Meta ads.

1. Le Rendez-vous E-commerce covers the process of building creative variants from winning Meta ads, with host Danilo Duchenne and returning guest Elisabetta from DHS Digital. 2. Elisabetta is a DHS Digital expert who has appeared in previous episodes; Danilo hosts the show and runs the agency. 3. The episode argues that brands have improved at producing new creative concepts but systematically under-produce variants of their existing winning ads, and presents a structured method for fixing that. 4. Elisabetta distinguishes two types of variants: variations on high-performing creatives where all metrics are green (changing creator, setting, product, or concept format), and targeted variants designed to fix a specific weak point identified in a creative's engagement metrics. 5. For high-performing creatives with no weak points, variation dimensions include swapping the creator or persona, changing the filming location or setting, applying the same script structure to a different product, or replicating a UGC approach across multiple products. 6. Danilo describes a client analysis session where he mapped creatives by creator, angle, concept, persona, and hook in a table and found that while podcast-format concepts were well-varied, creator and persona variation was neglected. 7. Tools like Atria and Motion are recommended for identifying which creator, hook type, or CTA performs best, with nomenclature-based tracking enabling systematic pattern recognition. 8. The weak-point diagnostic focuses on three storytelling metrics: hook rate (DHS benchmark 20-25%), hold rate (5-7%), and CTR (minimum 1%), all compared against both the industry benchmark and the account's own average. 9. Hook click rate — the proportion of people who watched the first three seconds and then clicked — is highlighted as an additional signal of hook relevance; a high hook rate paired with a very low hook click rate suggests the hook is attracting the wrong audience. 10. The hypothesis-based variant process works as follows: identify the low metric, form one specific hypothesis about its cause (e.g., 'the hook does not state the problem immediately'), then produce a single targeted variant to test that hypothesis. 11. A concrete example: a client's top hooks all used a trending vocal audio effect; a creative with a weak hook rate was given a variant using that effect on the hook only, then queued for testing. 12. Another example: a client selling spicy sauces found that showing the sauce flowing into a dish in the first seconds consistently produced strong hooks; creatives missing that opening were identified as candidates for hook variants. 13. Hypothesis validation requires confirming the same change works across three or four different creatives before replicating it broadly; a single confirmation is treated as inconclusive. 14. Danilo gives the example of a client with a winning 10-second CTA that was tested as a hypothesis across multiple creatives: each time the CTA was added, it validated, confirming it as a replicable winning element. 15. Variant hypotheses and results are tracked monthly or bi-monthly in review sessions, with validated hypotheses logged in a spreadsheet and replicated across new and existing creatives. 16. GoMarble, a tool that integrates with Meta through Claude MCP, is recommended for analyzing video creatives directly from the Meta account — identifying common patterns, marketing angles, and winning script structures without manual export. 17. Gemini is used to analyze scripts from top-performing videos, identifying the elements and sequence that appear most frequently across winning creatives. 18. The episode is conversational and example-driven, with Danilo and Elisabetta alternating between process explanation and specific client illustrations; energy is collaborative rather than lecture-style. 19. Performance marketers managing Meta ad accounts who want a systematic process for creative variant testing will find the most value. 20. Listeners looking for beginner Meta Ads setup guidance or those without an existing portfolio of winning creatives to vary from may find the episode premature for their stage.

Listen to the show on

Apple Podcasts Spotify

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