Gunjo · Business Intelligence for the AI Era
← Sticker Wall AGENT · DETAIL

Claude Code-Powered DTC E-commerce Ad AI Optimizer - Monthly Agency Fee of 20K-60K RMB

Workflow: Daily, Claude Code agents pull yesterday's spend, clicks, conversions, and ROAS data across platforms via Meta Ads API,

AGENT

Key Fields

FIELD STAMPS
IndustryE-commerce / Retail
RegionGlobal
ScaleSME
ChannelOnline

🔧 Workflow

Daily, Claude Code agents pull yesterday's spend, clicks, conversions, and ROAS data across platforms via Meta Ads API, Google Ads API, and Ocean Engine Ads API, automatically calculating ROI for each ad set. Ad sets with an ROI below 1.5 are automatically paused or have their budgets reduced by 30%, while those above 3 have their budgets scaled by 20%. Simultaneously, it invokes the Midjourney API to batch-generate new creative A/B testing variations based on high-converting asset features. Humans only need to review a one-page decision summary daily to approve or reject abnormal adjustments, completing the day's work in about 30 minutes.

🛠 Setup Requirements

Requires grasping basic Claude Code invocation logic, familiarizing with permission application and calling rules for Meta Marketing API, Google Ads API, and Ocean Engine Ads API, and pairing with n8n to build cross-platform data pipelines. Initial setup takes about 2 to 3 weeks, requiring configuration of ad account permissions, cross-platform conversion tracking pixels, API keys, and exception alert rules. Subsequent weekly maintenance for API adaptation and strategy iteration takes about 4 to 6 hours.

🧰 Toolchain

  • 🔧 Claude Code
  • 🔧 Meta Ads API
  • 🔧 Google Ads API
  • 🔧 n8n
  • 🔧 AdValet
  • 🔧 Shoplazza

💰 Revenue

1. Overseas DTC sellers (Primary revenue): Commission-based ad agency service fees. Client monthly ad spend of 30,000 to 80,000 RMB × service fee rate of 12% to 15% = single client monthly fee of 3,600 to 12,000 RMB (estimated) × 4 to 6 managed agencies = monthly revenue of about 20,000 to 60,000 RMB, which constitutes almost all of the monthly revenue (approx. 100%, estimated, based on merchant self-reported figures independently unverified). 2. Incremental ROI commission: Commissions based on incremental ad performance brought by creative and landing page A/B testing. Commission ratios are not public, and the number of cooperating clients is unverified, making the exact revenue share unclear. 3. Lightweight self-service packages: Monthly subscription for optimization templates and creative tag libraries offered to small and medium sellers. Monthly subscription fees have not been published, subscriber counts are unverified, and revenue share is unspecified. 4. Opportunity point - Cross-platform optimization SaaS: Productizing cross-platform data pulling and ROI calculation scripts into a monthly subscription (basic API calls for Meta, Google, and Ocean Engine ad platforms are free, advanced data interfaces are billed by call volume at approx. 100 RMB/month, platform pricing is public). The potential scale of this remains unproven.

💸 Cost

Claude Code subscription and multi-platform API calls cost about 800 to 2,500 RMB per month; n8n self-hosted server costs about 50 to 100 RMB per month; Midjourney batch image generation subscription costs about 100 to 200 RMB per month. Basic API calls for Meta, Google, and Ocean Engine ad platforms are free, while advanced data interfaces are billed by call volume at approximately 100 RMB per month.

⏱ Time Investment

Dedicate 0.5 to 1 hour daily to review AI-generated decision summaries and handle exception alerts; spend 4 to 6 hours weekly setting up ad systems for new clients, updating ad platform API adaptation rules, and optimizing creative tag libraries.

🚀 Getting Started

Step one is to apply for developer permissions for Meta Marketing API, Google Ads API, and Ocean Engine Ads API, preparing enterprise qualifications and past ad campaign case studies as proof. Use Claude Code to write cross-platform data pulling and ROI calculation scripts, run a two-week test on your own or a friend's DTC store to verify optimization performance, and after accumulating 3+ successful case studies, start taking clients in vertical communities and cross-border service provider platforms.

🔑 Keys to Success

  • ✅ Only take on clients with monthly ad spend over 30,000 RMB to ensure service fees cover AI tool costs and leave a profit margin.
  • ✅ Convert the first batch of clients to paid status after a two-week free trial run to verify results, yielding the highest conversion rate.
  • ✅ Ad platform API permission application is the entry barrier; prepare enterprise qualifications and historical ad performance proof in advance.
  • ✅ Continuously train the AI creative understanding model to build an exclusive creative tag library tailored to the client's product category, achieving an ROI optimization accuracy over 30% higher than general models.

⚠️ 风险

  • ⚠️ Frequent adjustments to API interfaces and advertising policies by Meta, Google, and Ocean Engine require ongoing code maintenance and adaptation to prevent campaign interruptions.
  • ⚠️ Significant fluctuations in client ad performance may lead to service fee disputes; contracts must explicitly state performance-guarantee disclaimers, charging solely based on a percentage of actual ad spend.
  • ⚠️ AI misjudgments can lead to wasted ad budgets, necessitating the setup of daily maximum spend thresholds and emergency pause rules, with mandatory human review of anomalous campaigns during daily audits.

📌 Real Cases

  • 📌 Shoplazza has officially integrated the AdValet ad-serving AI agent, achieving fully automated Meta ad growth and validating the feasibility of AI-managed DTC store ad campaigns.
  • 📌 A developer used Claude Code to automate high-ticket affiliate marketing, reaching a monthly revenue of 1 million JPY (approx. 50,000 RMB), proving that the AI-driven ad optimization arbitrage model is viable.
  • 📌 Silicon Valley AI advertising startup Kiri AI reached an annual revenue of 72 million RMB solely by providing automated ad systems for DTC clients, completing a 200 million RMB financing round and validating the scalable value of the AI ad optimization arbitrage model.