Video summary

Comment Générer des Ads Gagnantes à la Chaîne | Stratégie Complète

Main summary

Key takeaways

Business

Business goal (e-commerce growth via ads/creatives)

  • The speaker argues that scaling e-commerce is driven primarily by advertising creatives (the “real 80/20”).
  • Even with a strong offer and funnel optimization, bad creative angles prevent profit scaling.
  • Target outcome: build a creative pipeline that reliably produces ads that become winners and “multiply results tenfold.”
  • Performance claim: creatives/creative projects can generate €500k+ sales (with examples including “over a million” generated from one improved creative concept).

Core problems when brands stagnate (diagnosis)

If you can’t scale or results are unstable, creatives are often the cause:

  • Quality problem: ads get clicks/looks good but can’t scale (low scaling capacity).
  • Quantity problem: not enough testing throughput / repeatedly testing the same angles.
  • Most common mistake: stagnation because teams both:
    1. test too little
    2. test similar concepts

This leads to diminishing returns.

Playbook: Creative “system” to scale (process requirements)

The speaker emphasizes an automated, repeatable system rather than ad-hoc creative output.

KPIs / metrics to manage creative performance

They track both micro signals and funnel/platform engagement:

  • Hold rate (watch retention)
  • CTR (click-through rate)
  • Watch time
  • Interactions (engagement)
  • Sales performance (final conversion signal)
  • Quality score proxy: especially engagement/sales, used to feed delivery decisions
  • Cart/basket outcomes and their relationship to AOV (and when to cut tests quickly)

Platform-specific gating signals:

  • TikTok: watch time, interactions, sales
    • The algorithm boosts or reduces after initial testing.
  • Meta (Facebook): watch time, interactions, sales, quality
    • Also affected by bounce/courpage (implied “view-through traffic quality”).

Creative throughput targets (quantity)

  • Benchmark: test at least 50 new ads/week (recommended minimum for systems).
  • Operating benchmark: test ~300 new creative ideas/week.
  • Testing volume target in the system: 5–10 new batches per week per product/market/language
    • ≈ 25–50 new creatives per week.

“Quantity + Quality” execution approach

  • Test a lot of new concepts to avoid saturation and stagnation.
  • Exploit winners: don’t let an “ATS/NATS winner” run without iterating variants.
    • Create better versions by testing:
      • different things / variants / wheels (creative refreshes)

Frameworks / structures mentioned (what to build)

1) Creative funnel logic (why ads matter first)

  • Ads are the first entry point of the e-commerce funnel.
  • Creatives must:
    • capture attention
    • force a decision (“OK I want to buy”)
    • send “green signals” to platforms to improve delivery and reduce costs

2) Neurolinguistic “decision sequence” (emotional peak)

Creatives trigger a psychological sequence leading to purchase:

  • Capture attention (hooks)
  • Maintain attention (dynamism/involvement, cinematic transitions)
  • Build an emotional peak stronger than the “pain of buying”
  • Prompt action (clear CTA + ideally urgency)

3) “Awareness pyramid / stages of consciousness” (market capture)

Segments differ by how aware the prospect is:

  • Unaware of having the problem
    • Hardest: requires education + multi-stage persuasion.
  • Aware of the problem
    • Emphasize consequences; introduce the solution path.
  • Aware of the solution
    • Differentiate vs alternatives; discredit competitors.
  • Aware of the product
    • Reinforce desire, explain the “unique mechanism,” add urgency.

Implication: many “dropshipping-like” sellers only capture the small bottom share (price/offer-driven), while stronger creative/VSL execution expands capture across the full pyramid.

4) VSL (VSL = “Long Sales Video”) structure (psychological blocks)

Ordered framework:

  • Hook (capture tension)
  • Induction / maintain tension
  • Problem + root cause / educate
  • Impact on life emotionally (empathy; often via authority figure)
  • Mechanism explained (how it works; loop/teach)
  • Discredit wrong solutions
  • Imminent prediction (worse future if no action)
  • Present the solution/product
  • Benefits (especially emotional visualization)
  • Testimonials (realistic + aligned to the specific condition/problem)
  • Offer (big offer + guarantee + urgency + “act now”)

5) 5-part automated creative production system

A pipeline with five parts:

  1. Research (product + market + competitors; identify winning scripts)
  2. Scripting (apply winning psychological structures; write in batches; multiple hooks per concept)
  3. Sourcing/production (UGC/in-house/internet sources; stock/youtube; use high-quality footage)
  4. Testing (isolate variables; test creative only; e.g., 5 ads in 2 ad sets; broaden + improve audience)
  5. Analysis (micro + macro; optimize winning creatives; determine whether issues are quality vs quantity)

Creative/production rules (actionable “how to build”)

  • First seconds must be captivating—otherwise money won’t come.
  • Sequence pacing: transitions/different footage every ~2–3 seconds
    • Dynamic, but not so fast that content becomes “unreadable” subconsciously.
  • Native ad feel: match the platform’s look and feel
    • e.g., filmed like an iPhone/video-native style rather than obvious stock-style ads.
  • No visual/auditory errors (QC).
  • Disney technique (visual psychology):
    • Use rounded shapes / friendly visuals (positive associations)
    • Avoid sharp/aggressive design cues
    • Use pleasing colors, lighting, and “satisfying” visuals
  • Emotion guidance:
    • Positive early, or positive then shift positive quickly
    • Avoid ending on negative “harm” tone; aim for understanding + action (helpful framing)
  • AI voice caution: AI can help, but the voice must sound natural.

VSL automation guidance (how to get scripts fast)

  • Recommended approach: analyze top-performing VSLs/ads, download scripts, then reproduce/adapt them.
  • If adapting from the US to France:
    • remove/adjust elements that don’t fit local norms
    • e.g., change or remove promo language like “70% guarantee,” “two items left,” etc.

Testing protocol (what to isolate and when to cut)

  • Don’t test multiple variables at once.
    • Ensure ad copy/headline/thumbnail/landing page/offer are already proven.
    • Then test only the creative.
  • Decision timing:
    • wait about 2 days to decide
    • if spend is near/bad relative to AOV with weak cart/basket signals → cut quickly
    • otherwise cut once it exceeds the maximum rule threshold
  • Goal: 25–50 creatives per product/market/language per week, with automation rules to reduce losses.

Example case (creative iteration that scaled)

  • They improved an existing winner by creating a V2:
    • example changes: no music, no text, different footage
  • Improved performance on a ~60-second ad.
  • Result: over a million generated from that ad concept.
  • Post-success risk:
    • audiences became accustomed (“concept tired out”)
    • required switching ad style after saturation

Presenter/source(s)

  • Mateo (speaker; mentioned at the end: “It was Mateo.”)

Original video