ads-budget

AgriciDaniel ｜ 行銷與 SEO ｜ 原文 ｜ MIT

規劃預算、出價和 pacing，用邊際報酬做分配決策

---
name: ads-budget
description: "規劃並檢視付費媒體預算、出價、pacing、邊際報酬、目標、限制條件，以及在支援的平台之間的分配。適用於廣告預算分配、媒體預算、出價策略、放大規模、花費 pacing、ROAS 目標或投資取捨。計算與預測請用 ads-math。"
---

# 預算與出價

1. 先確認目標、轉換價值、毛利、現金限制、銷售承載量、歸因不確定性、季節性，以及平台所需的最低證據量。
2. 把花費和成果正規化到可比較的窗口與定義。
3. 計算損益平衡的界線，並列出公式、輸入、不確定性和敏感度情境。
4. 區分已承諾的基準、受控的實驗和預備額度。
5. 比較維持、重新分配、放大、縮減或做實驗這幾個選項，要用邊際證據，不要只看混合平均。
6. 回傳能直接決策的計畫，包含各平台/廣告活動的金額、時程、負責人、護欄、成功衡量和 rollback 觸發條件。

70/20/10、固定的 CPA 倍數、固定的預算對 CPA 比例，以及固定百分比的放大，都只是選配的 heuristic，絕對不能當成通用的授權依據。

---
name: ads-budget
description: "Plan and review paid-media budgets, bidding, pacing, marginal return, targets, constraints, and allocation across supported platforms. Use for ad budget allocation, media budget, bidding strategy, scaling, spend pacing, ROAS target, or investment tradeoffs. Use ads-math for calculations and forecasts."
---

# Budget and Bidding

1. Establish objective, conversion value, gross margin, cash constraints, sales
   capacity, attribution uncertainty, seasonality, and platform minimum evidence.
2. Normalize spend and outcomes to comparable windows and definitions.
3. Calculate break-even boundaries and show formulas, inputs, uncertainty, and
   sensitivity cases.
4. Distinguish committed baseline, controlled experiments, and reserve capacity.
5. Compare hold, reallocate, scale, reduce, or experiment options using marginal
   evidence rather than blended averages alone.
6. Return a decision-complete plan with platform/campaign amount, timing, owner,
   guardrails, success measure, and rollback trigger.

Rules such as 70/20/10, fixed CPA multiples, fixed budget-to-CPA ratios, and fixed
percentage scaling are optional heuristics, never universal authorization.