China Ecommerce Operator

Problem

China e-commerce decisions are often distorted by platform GMV, paid ROI, and attributed orders that ignore refunds, fulfillment, commissions, and ad spend. Operating actions can also slip into external writes—pricing, ads, listings, refunds—without evidence, stop lines, or rollback. This skill frames merchant problems as verifiable net revenue, contribution profit, platform-native actions, and reviewable experiments, protecting compliance and profit before GMV.

How It Works

The workflow starts with a compact merchant brief, then gathers evidence by priority: store back-office, official APIs, user-provided CSV/XLSX, platform rules, and public web sources. Time-sensitive rules are logged with URL, capture date, effective date, and confidence. It builds the operating base with GMV, net revenue after refunds, per-order contribution profit, and breakeven ROAS, and can use scripts/commerce_math.py to show inputs, formulas, and sensitivity. It then isolates the primary constraint and outputs do / don’t / verify / insufficient-info, plus 1–3 experiments with hypothesis, single variable, sample/period, stop line, and success line. External writes require a preview of account, object ID, content diff, impact, and rollback before explicit approval for that action.

Boundaries

Use it for platform strategy, selection, profit modeling, competitive research, compliance checks, and experiment review. It is not for bypassing anti-automation, fake orders/reviews, infringing reposting, or high-risk categories without qualifications. For food, medical, livestream, and commercial AIGC use, follow current official rules and human review; any numeric thresholds proposed by the plan should be marked as [strategy assumption].