Run Campaigns Overnight Safely: The Rule Stack That Replaces Midnight Checks
Alessandro Conti
シニア・パフォーマンスマーケター
The right ad automation rules run campaigns overnight safely — so you can close the laptop at 11 PM and wake up to controlled outcomes rather than budget disasters. The fantasy version of running Meta ad campaigns is: set them up, walk away, and come back to money. The reality is more anxious — most media buyers check their phone at midnight, 2 AM, or 3 AM because they do not trust what might have happened while they were asleep.
Ad automation rules that let you run campaigns overnight safely are not about trusting Meta. They are about building a system that handles every overnight failure mode automatically so the question "what went wrong while I slept" is answered before you ask it.
Quick answer: The overnight rule stack that lets you close Ads Manager at 11 PM has four parts: a zero-conversion spend cap that pauses ad sets burning without converting, a CPA circuit breaker for sustained underperformance, an account anomaly guard that alerts on spend spikes individual rules miss, and Telegram routing across two channels — emergency alerts that ring, non-urgent alerts that wait until morning.
This guide covers the exact configuration for each component, the Telegram routing setup, and the morning review ritual that closes the loop.
Why Media Buyers Check Their Phones at 3 AM
Two things drive overnight anxiety: the zero-conversion burn and the silent account anomaly.
The zero-conversion burn is the most common overnight disaster. An ad set is running at normal spend pace, but the pixel is broken, the landing page is down, or the audience is exhausted. No conversions are recording. Without a rule to catch it, this runs all night — €50/hour, €100/hour depending on the budget — and you wake up to a €400-€800 loss with zero revenue.
The silent account anomaly is subtler. A campaign duplication error, a budget misentry, or an unexpected Meta delivery spike sends the account's total spend to 200% of your planned daily budget. Individual campaign rules do not catch this because each campaign looks within limits — the problem is at the account aggregate level.
Broken tracking is a documented, widespread problem: Adobe reported in 2023 that companies estimate roughly 30% of their marketing budget is wasted on the wrong channels or due to measurement gaps, and a 2024 HubSpot State of Marketing analysis found that a large share of marketers still struggle to attribute conversions accurately across their funnel. A misfiring pixel overnight sits squarely inside that wasted slice. The fix for both failure modes is structural, not behavioral — rules that act autonomously when you cannot.
The Overnight Rule Stack: Four Components
Component 1: Zero-Conversion Spend Cap
This is the most important overnight rule. It catches the zero-conversion burn scenario within the first few hours.
Configuration:
Conditions (ALL must be true):
- Spend today > 3x target CPA
- Conversions today = 0
- Ad set status = Active
- Ad set running time > 4 hours (prevents pausing brand-new ad sets before they ramp up)
Action: Pause ad set + Telegram alert to emergency channel
Evaluation frequency: Every 30 minutes
Cooldown: 12 hours
Why 3x and not 1x or 2x: Meta's conversion attribution has a 1-3 hour lag for most pixel events. An ad set with €30 target CPA that has spent €60 and shows "0 conversions" may have 2 conversions in the attribution pipeline. At €90 (3x), the statistical likelihood of a full recovery is low enough to justify pausing. At €60 (2x), pausing creates too many false positives.
The 4-hour running time condition is equally important. New ad sets in their first 4 hours have irregular spend patterns as Meta's delivery algorithm calibrates. Applying the zero-conversion rule during this ramp-up period creates false positives from normal early delivery behavior.
The zero-conversion spend cap is the single most important overnight rule a media buyer can configure. A zero-conversion burn running unchecked for 8 hours is a budget disaster that ruins the month. A zero-conversion burn caught at €90 is a €90 lesson. One rule, configured correctly, is the difference.
See ad spend cap automation rules for the complete five-layer spend cap architecture.
Component 2: CPA Circuit Breaker
The CPA circuit breaker catches ad sets that are converting but at an unsustainable cost — the scenario the zero-conversion rule misses.
Configuration:
Conditions (ALL must be true):
- CPA (last 3 days) > 2x target CPA
- Spend (last 3 days) > 5x target CPA (enough data for confidence)
- Conversions (last 3 days) > 3 (confirms there is actual conversion data, not noise)
Action: Pause ad set + Telegram alert to non-urgent channel (for morning review)
Evaluation frequency: Every 3 hours
Cooldown: 24 hours
Why 3-day lookback instead of today: A 3-day CPA gives you the trend, not the moment. Ad sets have natural daily variance — a €40 CPA Tuesday might be €28 CPA Wednesday. Acting on a single day's CPA creates false positives from normal daily variance. Three days of sustained above-target CPA is a real signal.
Why to non-urgent channel: A CPA circuit breaker firing overnight is usually confirming a trend that started yesterday. The pause is appropriate, but it is not a 3 AM emergency — it is a morning review item.
Component 3: Account-Level Spend Anomaly Guard
This catches the aggregate problems that individual campaign or ad-set rules miss.
Configuration:
Conditions (ANY must be true — OR logic):
- Total account spend today > 140% of planned daily budget
- Any single campaign spend today > 200% of that campaign's daily budget
Action: Telegram alert to emergency channel (human reviews before acting — no auto-pause at account level)
Evaluation frequency: Every 1 hour
Cooldown: 3 hours
Why alert-only (no auto-pause): Account-level automatic pauses are too blunt. If a single campaign is misbehaving, pausing the entire account stops your best-performing campaigns along with the problematic one. Account-level anomalies need a human decision about which specific entity to address.
Why 140% and not 120%: Meta's daily budget delivery has natural variance — actual daily spend ranges from 90% to 110% of budget on normal days. 140% flags a genuine anomaly, not normal variance. 120% would create frequent false alerts on high-variance accounts.
The account-level guard exists because individual campaign rules are blind to aggregate failure. A duplication error or budget misentry can push total spend far past plan while every single campaign still reads as in-bounds. Watching the account total, not just the parts, is what catches the spend spike no per-campaign rule will ever see.
Component 4: Telegram Routing — Two Channels
The alert system is only as good as the notification routing. A useful overnight alert needs to reach you when necessary, but not create alert fatigue that leads to notifications being turned off.
Channel A: Emergency (sound on, all night)
What routes here:
- Zero-conversion spend cap triggers
- Account-level spend anomaly > 150%
- Landing page errors (if you have external monitoring)
Notification setting: Sound on, badge count visible, banner alerts.
If this channel fires between 11 PM and 7 AM, it is worth waking up for. The rule filters have ensured that only genuine crises reach this channel.
Channel B: Morning Review (notifications off overnight)
What routes here:
- CPA circuit breaker pauses
- Frequency/creative fatigue alerts
- Scaling rule confirmations
- Budget pacing warnings
- All informational alerts
Notification setting: Silent overnight, check first thing in the morning as part of the morning ritual.
The split ensures you sleep through the non-urgent queue while remaining reachable for genuine emergencies. See how to set up ad alerts on five platforms via Telegram for the full Telegram configuration.
Pre-Sleep Checklist: What to Verify Before Closing the Laptop
The overnight rule stack provides coverage, but a 5-minute pre-sleep check ensures no edge cases slip through.
Check 1: Active ad sets confirmed. Are the ad sets you expect to be running actually active? A brief scan of campaign status prevents the scenario where you think something is running but it paused earlier in the day.
Check 2: Budget levels reviewed. Verify that daily budgets are set correctly. An ad set with a €1,000 budget instead of €100 (a common data-entry error) will burn significantly overnight. The account-level anomaly alert should catch this, but a pre-sleep budget scan is a faster catch.
Check 3: Landing pages working. Open your landing page(s) in a browser and confirm they load correctly. Many budget burns happen because a landing page went down at 7 PM and the zero-conversion rule needs several hours of spend to confirm the issue. Catching a down page before midnight eliminates the overnight zero-conversion scenario entirely.
Check 4: Pixel firing. Spot-check that your pixel is firing by checking Meta Events Manager for recent events. If no events show in the last 2 hours on an active campaign, something is wrong.
Check 5: Rules enabled. Confirm the four overnight rules are active — it is surprisingly easy to disable a rule while editing it and forget to re-enable.
This checklist takes 5 minutes and eliminates most of the edge cases that slip through the rule stack.
The Morning Review: Closing the Loop
The overnight rule stack is not "set and forget" forever — it needs a daily review to stay calibrated.
Morning review process (10-15 minutes):
-
Check Emergency channel first. Did anything fire? If yes, what was the situation? Was the alert a true positive or a false positive?
-
Check Morning Review channel. Scan all overnight alerts in order. For each: appropriate action or false positive?
-
Check the rule audit log. Open the automation rule execution log and scan for any rule that fired more than once on the same entity (may indicate missing cooldown), any rule that paused a high-performing ad set (may indicate threshold calibration needed), any expected rule that did not fire (may indicate a broken condition).
-
Restart appropriate paused ad sets. If the zero-conversion rule paused ad sets that you now know had a legitimate reason for zero conversions (landing page down from 11 PM to 2 AM, now fixed), restart them.
-
Update kill/scale list. Products that the CPA circuit breaker paused for the third time in 7 days probably need a human kill decision, not another restart.
The morning review is not a monitoring session — it is an exception handler. The rule stack produces a queue of items that need human judgment. Your job each morning is to process that queue, not to re-examine all the data the rules already processed.
Calibrating the Stack for Different Account Types
The thresholds above work as defaults, but calibrate based on your specific account characteristics.
High-AOV e-commerce (€150+ orders):
- Zero-conversion trigger: 4x target CPA (higher AOV = slower attribution, more false positives at 3x)
- CPA circuit breaker: 1.7x target (e-commerce margins are tighter)
- Account anomaly: 130% (less expected variance)
SaaS trial campaigns:
- Zero-conversion trigger: 5x target CPA (trial signups have 3-7 day attribution windows)
- CPA circuit breaker: 2x target (trial-to-paid conversion is expected to take longer)
- Account anomaly: 150% (more acceptable variance given longer evaluation windows)
Dropshipping / impulse products (sub-€50 orders):
- Zero-conversion trigger: 3x target CPA (standard — impulse products convert quickly)
- CPA circuit breaker: 2x target (more variance is acceptable)
- Account anomaly: 130% (fast-burning ad sets need tighter spend control)
For the dropshipper-specific overnight rule configuration, see dropshipper tiered budget scaling rules which pairs scaling rules with the safety net above.
Building the Stack in Wevion
Wevion's rule engine provides the compound AND/OR logic that these overnight rules require, plus the Telegram notification integration and rule audit log that make the morning review practical.
Setup sequence:
-
Create the zero-conversion spend cap rule first (highest priority). Test in alert-only mode for 3 days before enabling auto-pause.
-
Create the CPA circuit breaker. Set 3-day lookback, add the conversion floor condition, set 24-hour cooldown.
-
Create the account-level anomaly alert with OR logic (140% OR any campaign at 200%). Alert-only, no automatic action.
-
Configure Telegram integration: connect your Telegram channels, label Channel A as "emergency" and Channel B as "review." Map rules to the appropriate channel.
-
Run the full stack in alert-only mode for 7 days. Review every alert for appropriateness. Adjust thresholds until your false positive rate is below 15%.
-
Enable auto-pause on the zero-conversion rule after the alert-only validation period.
For a step-by-step rule creation walkthrough, see how to automate Meta ads rules step by step. For the advanced multi-condition logic that extends these rules, see advanced multi-condition automation rules for Meta ads.
What the System Cannot Cover
Transparency about the limits is important. The overnight rule stack handles the automated-detectable scenarios. It does not handle:
Meta platform outages. If Meta's delivery system goes down overnight, campaigns pause naturally. Rules cannot accelerate or improve this — it is a platform event.
Sudden policy reviews. If Meta's system flags a creative or account for policy review overnight, the review happens regardless of rules. Rules cannot prevent or accelerate policy reviews.
Complex audience depletion. A Lookalike audience that reaches saturation overnight shows up as declining CTR and rising CPM, which the creative fatigue alert will eventually catch — but there may be a 24-48 hour lag before the alert fires.
Competitor activity. If a competitor launches an aggressive campaign that drives up CPM across your target audience overnight, rules will catch the downstream effects (CPA rising) but not the upstream cause. The morning review is where you diagnose causes.
Key Takeaways
Running Meta campaigns overnight safely requires four components working together: a zero-conversion spend cap (pauses burning without converting), a CPA circuit breaker (pauses sustained underperformance), an account-level anomaly guard (catches aggregate problems individual rules miss), and a Telegram routing system with separate emergency and review channels.
Add a 5-minute pre-sleep checklist and a 10-15 minute morning review to close the loop, and you have a system that replaces overnight dashboard anxiety with controlled exception handling.
The measure of success: if the emergency Telegram channel fires fewer than twice a week on average, your thresholds are calibrated correctly. If it fires nightly, thresholds are too sensitive. If you are discovering overnight problems during the morning review that no rule caught, you have a coverage gap.
For the complete automation rules architecture that this overnight stack is part of, see Wevion automation rules deep dive.
This guide is part of our automation rules hub.
よくあるご質問
The Ad Signal
推測を拒否するメディアバイヤーのための週刊インサイト。1通のメール。シグナルのみ。
関連記事
広告費上限の自動化ルール：予算超過を未然に防ぐ方法
広告費上限の自動化ルールは、管理された予算と午前3時の災害との分かれ目です。このガイドでは、すべてのタイプの費用上限ルール、使用すべき正確なしきい値、そしてそれらを完全な予算保護システムに組み込む方法を解説します。
Wevion自動化ルール：TheOptimizerより強力か？
Wevionの自動化ルールエンジンは、複合AND/ORロジック、カスケードルールチェーン、カスタムクールダウン、クロスアカウント適用を提供します。TheOptimizerのメトリクス深度とConvertBombのカスタム数式と比較します。
Automation Rules Meta Ads: 5 Mistakes That Waste Spend and How to Fix Each
Automation rules promise to save you time and protect your budget. But badly configured rules can pause your best ad sets, trigger conflicting actions, and cost more than they save. This guide covers the five most common mistakes and the exact fix for each.