Comment j'ai fait plus de 200K$ en 35 jours grâce à Trendtrack + Claude (copie moi)

Business strategy (product discovery → brand-building → automation)

• Positioning: Find early, scalable products that can become a brand + ecosystem—not just “trend grabs.”
• Core execution thesis: Identify new product opportunities by spotting recently launched shops with heavy ad activity (a proxy for traction/validation). Then use AI to turn research into a ranked shortlist and a go-to-market testing plan.

Framework / playbook embedded in the method

1) TrendTrack product sourcing filters (manual “patterns”)

Shop origin filter

• Exclude India + Pakistan
  • Reason: lower traffic quality and more frequent copying of US offers.
  • Low CPM/CPC doesn’t justify the approach.

Recency filter

• Creation date: last 30 days
• Optional adjustment: 60 days to modify timelines.

Ad-volume filter (traction proxy)

• Ads: minimum 50 ads in the last 24 hours
• Rationale: spy tools can lag and show “zero traffic” at first.
  • Meta live ads indicate the store is actively testing and scaling.

2) Product selection criteria (what to keep vs skip)

Prefer products that show

• Sustained / rising ad activity
• Newness in concept/angle
  • Avoid “the same product” that’s been around for 2–3 years.
• Clear marketing angles
  • e.g., fear/anxiety narratives, seasonal needs, specific demographics
• Brand/AV (avatar) expansion potential
  • Prefer non-commodity positioning.

Avoid / be cautious with

• Products that are too established or dominated by big brands (especially in cosmetics).
• Supplements in the US
  • High CPMs for ~6 months per speaker.
• Some categories described as harder/complex operationally (notably supplements).

Category guidance

• For recurring-style purchases: cosmetics instead of dietary supplements (better/more manageable).
• US “food supplements” not recommended; for supplements-like needs, consider launching in Europe/UK.

3) AI automation using Claude + TrendTrack API (MCP/Cloud)

Workflow change

• Replace 2–3 hours of manual market + competitor + angle analysis with ~10 minutes automated analysis (then review).

Output of the automation

• A product search table
• Top 10 products ranked with scoring
• For each product:
  • avatars/subavatars
  • marketing angles
  • competitors
  • target markets
• Generates a structured “what to test” instruction set for ads + landing/site build.

4) Scoring logic (margin/ROI-centric)

Margin target used in scoring

• Prefer products with at least “X4 margin” logic:
  • “products I want to launch are products that have at least a margin of X4”
  • phrased as: sales CF of 2 × 4
  • intent: high margin
• Color-coded scoring:
  • X3 acceptable to launch
  • preference is higher margin

Additional “quality checks” included

• COGS (from AliExpress / cross-checked)
• Active ads and ad formats
• Main angle + perceived value
• Seasonality (or sells all year)
• Shipping/weight fit
  • lighter = cheaper shipping → better unit economics
• LTV relevance (explicitly referenced)

Key metrics & KPIs mentioned

• Revenue + profitability case result

  • Found a product in February via this method:
    • >$200,000 revenue in ~5 weeks
    • >20% margin

• Research speed / operational KPI

  • Manual: 2–3 hours
  • Automated method: 10–15 minutes
  • Prompt creation took weeks, but ongoing usage is fast.

• Ad-based traction metric

  • ≥ 50 ads in last 24 hours
  • Store/shop created within last 30 days

• Marketing test budget guidance (CBO)

  • $50/day starting out
  • $100/day with larger budgets

• Targeting details used for testing (avatar age bands, examples)

  • 25–38
  • 30–45
  • 50–65 (tested as a separate sub-avatar)
  • 18–24 (sub-avatar example)

• Time to handle prompt

  • ~10 minutes, sometimes requiring “continue” clicks a couple times.

Concrete examples / use cases (how outputs are applied)

Category and angle examples seen during browsing

• Pets/animals products (“trendy,” heavy scaling)
• Products targeting elderly using fear/anxiety angles
• Copper-related offering with fear/benefit positioning
• Weight loss products scaling with heavy ad presence
• Cosmetics sub-niches:
  • dark circles / under-eye
  • peptide-based creams
  • leggings/shorts variations

Market expansion / avatar strategy examples from AI output

US scenario

• Primary avatar inferred: men 25–38 (professional/lifestyle)
• Sub-avatar example: students 18–24
  • hook: “low school pain / attention crisis”
• Recommended actions:
  • Test 2 sub-avatar creatives (video/testimonial/trucker style)
  • Use $50/day CBO budget
• Creative structure example:
  • “9:16 Hook, Hour 10 of my shift”
• Targeting detail:
  • interests + Facebook Job Categories

France scenario (seasonality/event-based)

• Example men targeting: 30–45 for a pre-summer product
• Test men 50–65 due to a “real issue” (elder men concerns)
• Event-based angle:
  • urgent wedding season (May–June)
  • build a story like: “my son’s wedding in 3 weeks, I need to be ready”

Actionable recommendations

• Use the filter combo to find “validated testing” stores:
  • Exclude India/Pakistan
  • Last 30 days
  • ≥ 50 ads / 24h
• Skip categories with unfavorable economics:
  • Avoid US supplements (high CPMs ~6 months) → consider Europe/UK
• Build around ecosystem potential:
  • Choose products where you can develop a niche store and repeatable brand messaging (not just a one-off sale).
• Operationalize the process:
  • Connect Claude ↔ TrendTrack via MCP + API key to generate:
    • top-10 shortlist
    • competitor set
    • avatar + sub-avatar plan
    • recommended creatives and testing actions per market
• After AI shortlist:
  • Manually verify shop/product logic:
    • ensure the shop isn’t “repackaging” ads across unrelated products (speaker notes scams/relauched shops with different FB pages).

High-level tech / implementation steps (business enablement)

1. Get MCP URL from doc.trentrack.io → Integrations
2. Configure the connector in Claude “Cloud” settings → Connector → Custom connector
3. Create and copy a TrendTrack API key
4. Paste API key into the provided prompt
5. Run the prompt:
   • it analyzes shops over the selected date window (example shown: “analyze shops over 30 days”)
   • may require Continue a few times

Presenters / sources

• Givan (main presenter)
  • Mentions “Alex” and a team involved.
  • References TrendTrack and Claude/Anthropic integration.
  • Only Givan is explicitly named as a presenter.