一键导入
skill-040-target-cpa-roas-realism-check
Verify that Target CPA and Target ROAS values are achievable based on historical performance.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Verify that Target CPA and Target ROAS values are achievable based on historical performance.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Verify that Search, Display, Shopping, Video, and Performance Max campaigns are never mixed in ways that compromise optimization signals.
Ensure branded and non-branded keywords live in separate campaigns with independent budgets and targets.
Audit naming conventions across all campaigns, ad groups, and labels for a standardized taxonomy.
Evaluate whether each ad group contains tightly themed, closely related keywords (ideally 5-20 per group).
Segment and analyze Search Partners performance separately from Google Search.
Check whether Search campaigns have inadvertently opted into the Google Display Network.
| name | skill-040-target-cpa-roas-realism-check |
| description | Verify that Target CPA and Target ROAS values are achievable based on historical performance. |
| allowed-tools | Read, Grep, Glob |
Unrealistic bidding targets cause significant issues: too aggressive targets lose auctions and volume drops; too conservative targets cause overspending and inefficiency. This skill validates whether current targets match historical performance and identifies campaigns where targets are limiting volume or wasting budget.
Data Source: Standard
Standard Data:
data/account/campaigns/*/campaign.md - Target CPA, Target ROAS settingsdata/account/bidding_strategies.md - Portfolio strategy targetsdata/performance/campaigns/*/campaign_metrics_30_days.md - Actual CPA, ROAS achievedReference GAQL:
SELECT
campaign.id,
campaign.name,
campaign.bidding_strategy_type,
campaign.target_cpa.target_cpa_micros,
campaign.target_roas.target_roas,
metrics.cost_micros,
metrics.conversions,
metrics.conversions_value,
segments.date
FROM campaign
WHERE segments.date BETWEEN '{start_date}' AND '{end_date}'
AND campaign.status = 'ENABLED'
Use /google-ads:get-custom for trend analysis over longer periods. Replace {start_date} and {end_date} with actual dates (e.g., 90-day window). Note: LAST_90_DAYS is not a valid GAQL date range; use BETWEEN with explicit dates instead.
| Condition | Severity |
|---|---|
| Target >30% below actual | Critical (unrealistic) |
| Budget utilization <50% + aggressive target | Critical |
| Target 20-30% below actual | Warning (aggressive) |
| Target >30% above actual | Warning (over-conservative) |
| Target within 10% of actual | Info (well-calibrated) |
Short (default):
## Target Realism Audit
**Account:** [Name] | **Campaigns with Targets:** [X] | **Misaligned:** [Y]
### Critical ([Count])
- **[Campaign]**: Target CPA $[X] vs actual $[Y] ([Z]% gap), [W]% budget used -> Increase to $[Rec]
### Warnings ([Count])
- **[Campaign]**: Target ROAS [X]% vs actual [Y]% -> Adjust to [Rec]%
### Target vs Actual
| Campaign | Target | 30d Actual | Variance | Budget Used | Status |
|----------|--------|------------|----------|-------------|--------|
### Recommendations
1. [Priority action with expected impact]
Detailed adds: