charlieviettq/awesome-agent-skill / .claude/skills/algo-ad-bidding
"algo-ad-bidding"
Verified"Implement and select ad bidding strategies from manual CPC to automated target-CPA and target-ROAS. Use this skill when the user needs to choose a bidding strategy, set up automated bidding, or optimize bid parameters — even if they say 'what bidding strategy should I use', 'target CPA setup', or 'smart bidding configuration'.".
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.3 KB |
| examples/sample_scenario.md | markdown | 5.6 KB |
| references/learning-period.md | markdown | 8.5 KB |
| references/migration-playbook.md | markdown | 11.5 KB |
examples/sample_scenario.md
View on GitHub# Example: 家居品牌從手動 CPC 轉換至 Target CPA
## Scenario
**公司:** 居家生活電商 HomePlus(台灣本地品牌,年營收約 NT$8,000 萬)
**廣告帳戶狀況(2026 年 3 月):**
- 平台:Google Ads,搜尋廣告為主
- 預算:NT$150,000 / 月
- 目前策略:手動 CPC,平均出價 NT$12–35(依關鍵字分組)
- 月轉換數:112 筆(追蹤事件:完成結帳)
- 目前平均 CPA:NT$580
- 目標 CPA:NT$500(行銷主管設定的 KPI)
- 轉換追蹤:Google Ads Tag + GA4 匯入,已驗證無重複計算
**使用者問題:**
> 我們手動出價很費時,帳戶有 112 筆月轉換,想改用自動出價。Target CPA 設多少合適?什麼時候可以切?切完要觀察哪些指標?
---
## Analysis
### Phase 1:Input Validation
| 檢核項目 | 數值 | 判斷 |
|----------|------|------|
| 月轉換數 | 112 筆 | ✅ 超過 50 筆門檻,可用 Target CPA |
| 轉換追蹤準確性 | Google Tag + GA4 驗證通過 | ✅ 資料品質可信 |
| 有無收益資料 | 有(結帳金額),但各 SKU 差異大 | Target ROAS 可考慮,但先以 CPA 穩定基線 |
| 業務目標 | 降低 CPA(580→500)+ 維持量 | 效率優先 → Target CPA 適合 |
| 月預算 | NT$150,000 | CPA 目標 NT$500 → 理論最大量:300 筆;目前量 112 筆,預算非瓶頸 |
**Gate 通過:** 轉換資料充足、追蹤正確、策略選擇明確。
---
### Phase 2:Core Algorithm 應用
**策略選擇:Target CPA**
依 Iron Law 數據門檻:
- 112 conv/month >> 50 門檻 → Target CPA 可用
- 有收益資料 → 未來可升級 Target ROAS,但目前 SKU 均價差異(NT$800–NT$4,500)會讓 ROAS 目標難以設定,先以 CPA 控制效率
**Target CPA 初始值設定:**
不應直接設目標值 NT$500(過於激進),應從當前表現開始緩步收斂:
```
初始 Target CPA = 當前 CPA × 1.0 ~ 1.1
= NT$580 × 1.05
= NT$609(建議設 NT$600)
```
原因:演算法在學習期需要足夠的拍賣空間找到轉換,若目標設得比現況低 14%,學習期曝光量會大幅萎縮,難以蒐集足夠訊號。
**收斂計畫(3 個月):**
| 時段 | Target CPA | 預期操作 |
|------|------------|----------|
| 第 1–2 週(學習期) | NT$600 | 禁止調整,僅監控 |
| 第 3–4 週 | NT$575 | 學習期結束後,若 CPA < NT$600 則降 |
| 第 5–8 週 | NT$530 | 穩定後再收斂 |
| 第 9–12 週 | NT$500 | 達成 KPI 目標 |
**出價邏輯(Target CPA 運作方式):**
每次拍賣,演算法根據以下訊號預測轉換機率(P_conv):
- 使用者裝置、時段、地點
- 搜尋意圖(查詢語意)
- 受眾屬性(再行銷清單、相似受眾)
- 歷史互動行為
出價公式:
```
bid = Target CPA × P_conv
= NT$600 × P_conv
高意圖用戶(P_conv = 0.08)→ bid = NT$48
低意圖用戶(P_conv = 0.01)→ bid = NT$6
```
目標:跨所有拍賣平均 CPA 趨近 NT$600。
---
### Phase 3:Verification 指標
學習期(前 14 天)監控清單:
| 指標 | 基準值 | 警戒線 | 行動 |
|------|--------|--------|------|
| 日轉換數 | ~3.7 筆/天 | < 1 筆/天連續 5 天 | 檢查預算是否限制、CPA 是否過緊 |
| 曝光份額 | (切換前記錄) | 下降 > 40% | Target CPA 可能過低,暫緩調降 |
| 實際 CPA | NT$580 | > NT$800 | 學習期波動允許,超過才介入 |
| 轉換追蹤正常 | 每日有數據 | 連續 2 天 0 轉換 | 立即查 Tag 是否失效 |
**Gate:** 第 14 天後,若實際 CPA 落在 NT$480–720(目標 ±20%),視為學習期通過。
---
### Gotchas 應對(本案適用)
1. **學習期波動**:告知行銷主管第 1–2 週 CPA 可能飆至 NT$800+,這是正常現象,不要在此期間更動目標值。
2. **轉換延遲**:HomePlus 分析顯示,消費者平均在點擊後 1.8 天完成購買。Google Ads 轉換窗口應設為 **30 天**(非預設 7 天),確保演算法拿到完整訊號。
3. **預算非瓶頸**:目前 112 筆 × NT$580 = NT$64,960 實際花費,遠低於 NT$150,000 預算。Target CPA 不會因預算充裕就放量——它只在找得到符合 CPA 目標的拍賣時才出價。勿期待切換後花費自動倍增。
4. **季節性衝擊**:6 月父親節、雙 11 前後,競價環境劇變。需提前在 Google Ads 設定**季節性調整(Seasonality Adjustment)**,避免演算法誤判。
---
## Result
```json
{
"recommendation": {
"strategy": "target_cpa",
"target": 600,
"currency": "TWD",
"confidence": "high",
"rationale": "112 conv/month exceeds 50-conversion threshold; start at NT$600 (current CPA x1.05) to preserve learning signal, then step down to NT$500 over 12 weeks"
},
"expected_performance": {
"cpa_range": [480, 720],
"volume_change": "-5% to +20%",
"note": "Volume may dip during 14-day learning period before recovering"
},
"convergence_plan": {
"week_1_2": {"target_cpa": 600, "action": "freeze — learning period"},
"week_3_4": {"target_cpa": 575, "action": "reduce if actual CPA < 600"},
"week_5_8": {"target_cpa": 530, "action": "reduce if stable"},
"week_9_12": {"target_cpa": 500, "action": "KPI target reached"}
},
"metadata": {
"monthly_conversions": 112,
"current_cpa": 580,
"learning_period_days": 14,
"conversion_window_recommended_days": 30,
"budget_monthly_twd": 150000,
"budget_is_binding": false
}
}
```
**切換時機建議:** 避開促銷檔期(如雙 11、節慶),選在流量平穩的週二或週三上午切換,以確保學習期資料不被異常流量污染。
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |