$1,152,935/Sales with HIGGSFIELD on GPT-6 ASTRA - 0 to $100k Dropshipping Challenge

Business outcomes / case results

• Live example brand: A fashion dropshipping store built from scratch using Higgsfield + GPT-6 Astra
• Total sales: £848,000+ (≈ $1.1M+)

Performance examples

• September: $300,000+ in a single month
• Single day: $15,000+ on the 20th (example)
• Single day: £1,900/day on the 31st (example), and nearly $3,000/day on the 1st (example)

Brand hero creative

• A hero slideshow/trailer video described as: “made by supercomputer on Higgsfield with Astra”

Positioning/claim

• Ads and creative generation become largely “autopilot” after initial setup
• Large-scale testing is enabled quickly

Emphasis: “speed vs rushing” and understanding each step before moving on.

---

What the workflow is (playbook-style)

Core stack / components

• Shopify (store theme + product pages)
• Cloud Code + AI sourcing to add hundreds of products quickly
• Higgsfield for creative generation (images/videos)
• GPT-6 Astra to prompt/drive Higgsfield outputs
• Optional: Higgsfield “Supercomputer” for cinematic trailers
• Meta Ads Manager for launching and testing

---

Execution steps (repeatable process)

1) Setup the store

• Create a Shopify account (example promo mentioned: free first 3 days, then $1/month for first 3 months)
• Use a specific plug-and-play theme template (not from the Shopify Theme Store)
• Swap:
  • brand name
  • text referencing the creator
  • product catalog content
  • logos
• Add products at scale:
  • Cloud Code sources and inserts hundreds of products in minutes

---

2) Generate ad creative

• Use Higgsfield → generate creatives from image inputs (prompt-driven)
• Connect Higgsfield ↔ ChatGPT via an “install plugin” link
• Use GPT-6 Astra to generate ad concepts and outputs, described as:
  • “10 images + 5 videos per product” (per their prompt)
• Example output scale (from the content):
  • Select 3 products → generate 45 ads total in “a few minutes”
• Require diversity of marketing angles in outputs (avoid identical/samey creative)

---

3) Generate the hero/trailer video

• In Higgsfield Supercomputer:
  • click GPT-6 Astra
  • prompt: “create a cinematic trailer… for my brand” + brand link
  • download and place as hero media on the store

---

Ads testing + launch mechanics (operations details)

Testing/learning loop

After launching, the AI workflow is described as:

• prompt to review ad performance
• optionally connect to Meta Ads Manager so performance can feed improvements (“learn from it”)

---

Meta campaign setup (concrete parameters)

• Example ad account scale shown by presenter:
  • Meta spend: $507,000+ (not fully trustworthy because “Meta does not track everything”)
  • Example tracking metric: ~20,000 orders (as tracked by Meta)

Launch approach for a “new sales ad set”

• Campaign objective: Sales
• Budget: $30/day (example)
• Optimization event: Conversion / Purchase
• Avoid: using “Add to Cart” even if Meta suggests it
• Schedule: launch for the next day, ideally 8:00–9:00 AM in the selling timezone
• Website/product targeting: uses the specific product page URL per ad
• Ad creation: manual upload
  • upload/select the generated video
  • set:
    • Primary text from product description (copy/paste)
    • Headline = product name
    • Description includes offer (example: “free shipping in the UK”)
    • CTA = “Shop Now”
• Media handling: bulk testing by duplicating ads

---

Bulk testing workflow

• Create one ad structure
• Duplicate (e.g., duplicate 10x) and swap only the media (different generated videos) to test many variants quickly

---

Metrics & KPIs mentioned (and implied focus)

Revenue/cashflow KPIs

• $1.1M+ lifetime sales (example brand)
• $300k+ month
• $15k+ day
• £1,900+ day; ~$3,000 day (examples)

Advertising/ops KPIs (ad account screen examples)

• Spend: $507,000+
• Tracked orders: ~20,000 orders

Creative testing scale

• 45 ads in a few minutes from one prompt (3 products × 10 images × 5 videos)
• Claim: outputs can scale to hundreds of ads/products

No explicit CAC/LTV/churn targets were provided.

---

Actionable recommendations extracted from the content

• Choose a repeatable store template (plug-and-play theme) to reduce design friction
• Build product catalogs at scale:
  • use AI sourcing via Cloud Code to add hundreds of products quickly
• Generate creative in bulk:
  • provide AI with 2–3 product images + product details (title, price, offer/discount)
  • in the prompt, demand diverse angles to avoid “samey” ads
  • example batch size: 10 images + 5 videos per product
• Use Meta launch settings aligned to purchases:
  • optimize for Purchase, not Add-to-Cart
  • launch at 8–9 AM local selling time (example)
  • target using product page URLs per ad
• Rapid experimentation:
  • duplicate ad structures and swap only the generated media to run many variants quickly

---

Presenters / sources

• Presenter/tooling reference: Higgsfield on GPT-6 Astra (referred to as “Higgsfield”)
• Main speaker/host: “I” / channel owner (name not provided)
• Mentioned platforms/partners/tools:
  • Higgsfield
  • GPT-6 Astra (via ChatGPT integration)
  • Shopify
  • Cloud Code
  • Meta Ads Manager (Facebook/Instagram ads)
• Creator/theme DM reference: @realsimob on Instagram
• Contact mentions: Instagram, Telegram, Discord