How to Build a Bulletproof E‑commerce Promotion System Using Simple Math

The author, a product manager who works on e‑commerce systems by day and explores game mechanics by night, reveals how to configure a promotion system that appears complex to the boss but is actually simple.

1. Fatal case: a 2‑million‑yuan loss

During a mother‑and‑baby platform anniversary event, the stacking rule “full‑300‑50 + full‑200‑30 + 10% member discount” had no upper limit, causing a baby‑formula order to be priced at –52.4 CNY, leading to massive loss when scalpers bought 2,000 units.

2. Three root causes of the vulnerability

Uncontrolled discount verification order : applying discounts in the wrong sequence (stack then cap) creates negative prices; the correct order is platform coupon → shop coupon → full‑reduction → discount, from high to low, offsetting step by step.

Insufficient user risk controls : no detection of multiple orders from the same device ID, virtual numbers, or identical shipping addresses.

Broken circuit‑breaker mechanism : inventory of 1,000 sold 2,000 units without triggering alerts.

3. Defensive system three‑part solution

Technical rule :

<code>if (final_price < cost_price * 0.7) {
  trigger_manual_review;
  freeze_suspicious_account;
  send_alert_sms_to_CEO;
}</code>

Process tools :

Discount firewall: develop a “promotion combination calculator” and set gradient caps (e.g., max discount ≤ 50 % of product price).

Scalper‑identification matrix: device fingerprint, behavior path, order features, social network.

Circuit‑breaker escape route: automatically disable discounts when inventory falls below 80 %, limit each user to 3 items and enforce a 15‑minute payment window.

4. Financial rescue toolkit

Promotion combination risk calculator includes negative‑price warning, profit‑margin red‑line alerts, scalper risk scoring, support for up to five layers of discount stacking, automatic conflict detection, real‑time final price and margin display, product price range settings, cost‑price protection (no lower than 70 % of cost), and selectable discount types (full‑reduction, discount, coupon, points).

Scalper characteristic checklist covers device dimensions (fingerprint, multiple accounts, abnormal location), IP dimensions (multiple accounts per IP, proxy IPs, IP‑address mismatch), device behavior (time‑zone changes, unusual resolution, network switches), behavior dimensions (order speed, fixed intervals, shallow browsing, rapid cart clearing), social dimensions (account similarity, lack of social activity, batch registration), and order dimensions (multiple orders per account, abnormal amounts, concentrated timestamps, identical shipping info, single payment method, high failure rate).