Monitorização Diária de Desempenho e Pacing
Monitorize campanhas ativas para anomalias de desempenho (picos de CPM, quedas de CTR, violações de CPA) e problemas de pacing de orçamento numa base diária.
Porquê Usar Este Modelo
Monitorização diária automatizada deteta anomalias antes que esvaziem o seu orçamento. Apanhe picos de CPM, quedas de CTR e violações de CPA precocemente para poder intervir antes que gasto significativo seja desperdiçado.
Para Quem É
Gestores de conta que gerem múltiplas campanhas e precisam de uma forma sistemática de apanhar problemas de desempenho sem rever manualmente cada métrica todos os dias.
Como Usar
Execute diariamente ou defina um agendamento. Reveja o resumo de alertas primeiro — problemas críticos surfaceiam no topo. Verifique o estado de pacing para conformidade de orçamento. Tome ação em campanhas sinalizadas antes do fim do dia.
CONTEXT
Monitor active campaigns for performance anomalies and budget pacing issues. Run this check daily, weekly, or monthly based on your configured schedule.
CONFIG
| Parameter | Default | Description |
|---|---|---|
| Analysis Type | daily | How often this runs: daily, weekly, or monthly |
| CPM Spike Alert | 30% | Flag if CPM exceeds 7-day average by this much |
| CTR Drop Alert | 20% | Flag if CTR drops below 7-day average by this much |
| CPA Breakeven | — | Alert if CPA exceeds this value |
| Zero-Conv Spend Threshold | 2× daily budget | Flag if spend exceeds this with zero conversions |
| Under-pace Threshold | 45% at 50% elapsed / 56% at 80% elapsed | Flag if expected spend not reached |
| Over-pace Threshold | 110% | Flag if spend exceeds this vs expected |
PROCESS
Learning Phase Check
For each campaign, determine whether it is still in the learning phase:
- Still learning: spend is less than 50× the daily budget AND it has been less than 7 days since launch
- Exited learning: either condition has been met
Learning-phase campaigns are excluded from CPA anomaly checks, but always included in pacing analysis.
Anomaly Detection
Compare current performance against 7-day averages. Flag the following:
- CPM spike: current CPM is elevated beyond the configured threshold vs its 7-day average
- CTR drop: current CTR has declined beyond the configured threshold vs its 7-day average
- CPA above breakeven: CPA has exceeded the breakeven threshold
- Zero-conversion overspend: no conversions recorded and spend has exceeded the configured multiplier of daily budget
Budget Pacing
For each campaign, calculate expected spend based on time elapsed and compare against actual spend.
| Time Elapsed | Expected Spend | Flag if Under | Flag if Over |
|---|---|---|---|
| 50% of day | 50% of daily budget | < 45% = under-paced | > 110% = over-paced |
| 80% of day | 80% of daily budget | < 56% = under-paced | > 110% = over-paced |
Severity
- Critical: multiple anomalies detected, or zero-conversion overspend
- Warning: a single anomaly
OUTPUT
If no active campaigns: ✅ No active campaigns to analyze.
Otherwise output:
🚨 ALERT SUMMARY
| # | Entity | Metric | Current | vs 7-Day Avg | Severity | Cause | Action |
|---|
📊 PACING STATUS
| Campaign | Budget | Hours Elapsed | Expected | Actual | Pace% | Status |
|---|
- Critical: show recommended action only
- Warning: include suspected cause and recommended action
NOTES
- CPA checks are skipped when conversion data is unavailable — report
CPA: N/A - Always include learning-phase campaigns in pacing analysis
- If data is insufficient to evaluate any metric, skip that check without flagging