Video summary

De 0€ à +1M€/Mois en E-commerce, Épisode 4: Comment Faire La Recherche Marketing

Main summary

Key takeaways

Business

Business-focused summary: How to do Marketing Research for E-commerce (scale from €0 to €1M+/month)

Marketing research is treated as the core lever that determines whether ads convert—and how precisely you can write hooks/creative/tunnel messaging. The “goal” isn’t just collecting data; it’s using research to make ad-level decisions.

The core “answer these questions” framework (copy + ads)

Research should answer 3 questions that directly drive every word in your ads:

1) When does the customer experience the problem?

  • Find micro-moments: specific times/events when pain or desire peaks.
  • Turn those moments into the basis for hooks and b-roll.
  • Example: Back pain peaks at ~5pm after a full day sitting, creating a specific “crossbar” sensation.

2) Why does it affect them that strongly? (underlying emotion)

  • Don’t search for the physical cause—identify the emotional trigger.
  • Use a “For what?” chain:
    • “Why is it a problem?” → “Why does that matter?”
  • Example: Wrinkles → shock at a 38-year-old mirror/photo → fear of aging → fear of husband looking elsewhere / fear of death.

3) Who are you really attracted to? (avatar foundation via TAM similarity)

  • Don’t treat personas as surface-level demographics.
  • Define a psychological/behavioral pattern shared across your TAM.
  • Example (mushroom coffee): Different avatars (exec woman, biohacker man, sensitive-energy mother) share a foundation:
    • clean/sustainable energy without “spit”
    • and feeling intelligent in the choice
  • That shared pattern becomes the value proposition foundation.

Playbook: Data sources (internal vs external)

The research process is split into:

  • Internal data (most reliable, from your own customers)
  • External data (to discover new angles for prospects)

1) Internal research (highest reliability)

These data reflect people who already interacted with the brand.

Advertising data

  • Identify what’s working by creative + message patterns (including gender/age patterns linked to converters).
  • Review comments under ads to extract:
    • doubts, objections, and the exact language people use
  • Tools/workflow mentioned:
    • Trustflow exports comments to CSV
    • Analyze with Lia

Email + SMS data

Look for:

  • best click-rate campaigns (angles)
  • subject lines that drive open rate
  • messages that resonate with the same segments

Process recommendation: build a CRO team for email + landing pages that reports insights synchronously to feed ad testing.

Insight claim: a strong email can become a future advertising hook.

Post-purchase forms

Collect insights in two moments:

  • Immediately after purchase (“thank you” moment) while emotion is still high:
    • first impression of the brand
    • how they found you
    • what nearly prevented purchase
  • ~30–45 days after purchase (or after customers had time to use the product):
    • why they bought / the problem they wanted to solve
    • what convinced them vs alternatives
    • what their ideal version would look like

Outcome: supports tunnel optimization and product Version 2.

Customer reviews

  • Positive reviews: identify what proves “what works” (for ads/emails).
  • Negative reviews: extract objections and anticipated friction to address.
  • Example process claim: an automated system pulls best reviews weekly and converts them into ads and emails.

Customer service insights

  • Run a monthly prompt to the team: “What objections come up most often?”

  • Example: customers thought they bought a “copy” instead of the brand—so the brand created ads like:

    • “If you want the official, quality product, it comes with a guarantee; buy from us.”
  • The speaker claims these became top-performing ads (subtitle timing appears inconsistent in the original).

Operational principle

Even if you only gather internal data, feed it back into both:

  • ad strategy
  • product optimization to improve LTV (explicitly mentioned)

2) External research (find new audience pockets and gaps)

External data is less precise than internal, but critical for discovering:

  • new audience pockets
  • unexploited angles

Analyze competitor-winning patterns (no copying—extract patterns)

Extract:

  • “hook winners”
  • persona types
  • comment themes
  • VSL/funder formats and creative formats

Competitor set can include:

  • direct competitors (same category)
  • indirect competitors (adjacent categories with the same buyer mind)

Use an “angle x format” opportunity grid

  • Winning angle + winning format → strong opportunity
  • Winning format + absent/unused angle → new market entry opportunity
  • Winning angle + winning format + underutilized persona/character → “binge” (breakout)

Specific external channels mentioned

  • Winning creatives on competitor ads (creative formats + persona comments)
  • Organic TikTok
    • identify problems, promises, tone/emotion, formats
    • claim: explosive organic content reveals messages that resonate and can be replicated in ads
  • Amazon reviews
    • extract verbatim phrases used by customers (“goldmine for copywriting”)
    • reviews reveal what was liked/disliked and why they bought
  • Forums (Reddit, etc.)
    • uncover deeper realities: experiences, deepest pains, failures
    • use language and framing to generate new angles

“Rabbit hole” / deep immersion tactic (advanced research)

The speaker argues most e-commerce businesses do superficial research (e.g., a few reviews + random Reddit reads). The “better” method is to immerse yourself in the customer’s world:

  • Live as if you have the problem.
  • Follow the subculture content they consume.

Example (fitness niche):

  • Don’t only study “how to build muscle” on YouTube
  • Also study training philosophy + nutrition approaches
  • Learn decision variables like:
    • full body vs specific weights
    • keto vs flexible diet
    • training frequency (3 vs 7 days/week)
    • protein quantities relative to body weight

Outcome: create more nuanced, surgically targeted ads, tailored even to sub-avatars (e.g., re-educate keto community on long-term effects; explain why training 7 days/week can slow results).

Practical process outputs (turn research into execution)

  • Capture findings in a note / Google Sheet / Google Doc.
  • Feed summaries into Lia to validate/critique hypotheses (“am I right or wrong?”).
  • Use the validated angles in:
    • ads
    • funnels/tunnel messaging
    • email sequences
    • product iteration (V2)

Final directive from the speaker:

  • “massive research” → “massive results”
  • skipping deep research risks losing money.

KPIs / metrics mentioned

No explicit revenue targets, CAC/LTV/churn numbers, or financial timelines were provided. However, the process references:

  • Email KPIs: open rate (subject line), click-through rate
  • Ad optimization inputs: demographic patterns + conversion-linked creative/messages
  • Product metric direction: improve LTV
  • Research timing:
    • immediately after purchase
    • then ~30–45 days after purchase (or after meaningful product usage)

Concrete examples / case studies cited

  • Liquid Des (Mike Cesario):
    • scanned comments on punk-metal band videos
    • found a recurring social embarrassment moment (non-drinkers feeling awkward holding water at bars/festivals)
    • branding water matched the micro-moment (punks with a “punk beer” aesthetic)
  • Back pain micro-moment example:
    • pain peaks around 5pm after office sitting
  • Wrinkles emotional trigger example:
    • mirror/photo recognition → fear and relationship consequences
  • Customer service-driven ad fix:
    • ads added guarantees and direct-buy instructions to address “copy vs official brand” confusion
  • Fitness rabbit hole example:
    • tailoring angles by sub-community distinctions (keto vs flexible diet; training frequency; nutrition details)

Presenters / sources

  • Presenter: Matthéo (speaker referenced as “Matthéo” near the end)
  • Named external entities/tools (as referenced):
    • Trustflow
    • Lia
    • Apple (used as a benchmark reference for “Version 2”)
    • Liquid Des (Mike Cesario)
    • Platforms/channels: YouTube, Reddit, Amazon, TikTok
    • Workflow/tools mentioned: Google Forms/Typeform, iCloud exports
    • Category reference(s) as given: OMPC / Ozempic

Original video