Google Ads Automated Rules 2026: Practical Guide

Most Google Ads money is not lost while you are looking at the account. It is lost in the hours when nobody is: overnight, on weekends, during a site deploy, while the buyer is on a plane. The campaign keeps serving after the lead form silently stops submitting. The product sells out and the feed keeps advertising it for another day. A keyword that produced leads all month burns a week of budget across one Friday and returns nothing.

Google Ads automated rules are the built-in mechanism for writing that reaction down in advance: if this condition is true, do that, and tell me about it. No code, no server, no maintenance — the rule runs on Google’s side on a schedule and keeps running while you sleep. This guide covers how a rule is actually constructed, which frequency and data window to choose, nine scenarios with concrete thresholds, and — separately — the rules you must never put on an account that runs automated bidding.

A rule is not an autopilot. It is a circuit breaker. A good rule fires rarely, and when it does, it saves far more than it cost to set up.

The anatomy of a rule

Every automated rule is assembled from five mandatory blocks. Understanding each of them removes most of the reasons rules “don’t work the way I expected”.

Block	What it defines	What to watch for
Action	Pause, enable, change bid or budget, send an email	Email-only is the most underrated mode: oversight without interference
Apply to	Campaigns, ad groups, ads, keywords, audiences, asset groups	Scope by labels rather than by listing campaigns manually
Conditions	Metrics and thresholds joined with AND logic	There is no OR — if you need one, build two rules
Frequency	Hourly, daily, weekly, monthly, plus the run time	The run time follows the account time zone, not yours
Email results	Every time / only when changes are made / only on errors	“Every time” on a frequent rule turns your inbox into noise nobody reads

Rules live under Tools → Bulk actions → Rules, together with their run history. That history shows, per execution, whether the rule fired, how many items it touched and whether it errored out. Open it weekly: a rule that silently fails is more dangerous than no rule at all, because you believe you are protected.

The data window is where most rules break

Every condition carries a lookback period: yesterday, last 7 days, last 30 days (not including today). That last phrase matters. If a rule runs hourly on a seven-day window that excludes today, it sees the exact same frozen picture all day and can fire on the same object twenty times in a row.

The second failure mode is conversion lag. Recent days always undercount conversions, because some of them have not been written back to the click yet. A rule that looks at “yesterday” and pauses anything with zero conversions will reliably kill healthy objects. How to quantify that back-fill and at what depth data becomes honest is covered in our breakdown of conversion windows and conversion lag — without that correction, any conversion-based threshold is guesswork.

What a rule can and cannot do

It can	It cannot
Pause and enable campaigns, ad groups, ads, keywords	Create new objects — campaigns, ad groups, ads
Change daily budgets by percentage, by amount, or to a value	Reliably manage strategy targets (tCPA/tROAS) in most setups
Change bids on manual bidding	Influence bids inside an automated strategy
Email you on a condition — effectively custom alerting	Reach external data: CRM, inventory, exchange rates
Run on a schedule down to the hour	React instantly — hourly is the smallest step
Scope itself with labels and filters	Express branching logic or loops

That last row is the dividing line between rules and scripts. Anything requiring an external source, non-trivial math or object creation belongs in code — the mechanics are covered in our guide to Google Ads scripts and automation. Rules cover the simple half of the work, and that half happens to be the daily operational half.

Nine rules worth having in almost any account

The thresholds below are directional for a mid-sized account doing a few dozen conversions a month. Recalculate them against your own economics: a correct spend threshold is almost always a multiple of your target CPA, not a round number.

1. Kill switch for spend without conversions (keyword level)

Condition: cost over the last 30 days (not including today) is greater than three target CPAs AND conversions equal zero. Action: pause the keyword, send an email. Frequency: weekly. Three CPAs is a statistically meaningful signal rather than a bad streak, and at a 30-day depth conversion lag barely interferes.

2. Abnormal daily spend alert (campaign level)

Condition: yesterday’s cost is more than 1.5× the daily budget. Action: email only. Frequency: daily, in the morning. Intraday overdelivery is normal system behaviour; a 1.5× jump usually means someone raised a budget, an old campaign switched on, or a bid strategy changed.

3. Budget-limited campaign detector

Condition: search lost impression share (budget) above 30% over the last 7 days. Action: email. Being budget-limited hurts profitable campaigns more quietly than it seems: it truncates volume and distorts how impressions are distributed across the day. How to read those metrics correctly is in our piece on impression share and auction insights.

4. Conversion collapse monitor

Condition: conversions yesterday equal zero AND yesterday’s cost is above two target CPAs. Action: email. Frequency: daily at 11:00 account time. This rule does not catch a bad day — it catches breakage: a dropped tag, a broken form, a redirect to a 404. The spend threshold keeps small campaigns, where a zero day is normal, out of your inbox.

5. Low-CTR ad review (ad level)

Condition: impressions over 30 days above 3,000 AND CTR below half the ad group average. Action: email, not pause. Auto-pausing ads is risky: you can leave an ad group with no active responsive search ad variant and collapse its delivery.

6. Out-of-hours campaign shutdown

No condition — a scheduled pause in the evening and enable in the morning. Relevant for call-centre funnels where a lead outside operating hours costs the same and converts worse. The finer instrument is time-of-day adjustment; the trade-offs are compared in our article on ad scheduling and dayparting.

7. Budget increase on campaigns beating target

Condition: CPA over the last 14 days is 25% below target AND conversions above 15 AND lost impression share (budget) above 10%. Action: raise daily budget by 10%, email. Frequency: weekly. The hard constraint is the cadence — no more than once a week and no bigger than 10–15% per step, otherwise you permanently restart the Smart Bidding learning period and never see stable data.

8. Disapproved ads and assets alert

Condition: ad status is disapproved. Action: email. Frequency: daily. A disapproved ad inside a three-ad group is easy to miss in the interface but removes part of your reach. A daily check costs nothing and catches silent traffic loss.

9. Weekly digest of enabled campaigns with zero impressions

Condition: impressions over the last 7 days equal zero AND status is enabled. Action: email. An enabled campaign with no impressions is nearly always a symptom: a failed payment, an account-level negative keyword conflict, an empty asset group, a broken geo setting. Reviewing that list weekly is basic hygiene, the same class of work as a regular Google Ads account audit.

The rules that break automated bidding

This is where automation starts costing money. Smart Bidding is a learning system: it builds a model from account history and from the stability of its inputs. A rule that changes those inputs every day removes its ability to learn at all.

Dangerous rule	What happens	Replace it with
Daily budget swings of ±20%	The strategy constantly rebuilds its pacing model; CPA oscillates and periods are not comparable	Steps of 10–15%, no more than weekly
Hourly pause/enable of campaigns by CPA	Learning resets, accumulated history fragments	Dayparting or adjustments, without pausing the campaign
Auto-pausing keywords on “yesterday”	Conversion lag makes it kill working keywords	A 30-day window and a spend threshold of 3+ target CPAs
Changing tCPA/tROAS with a rule	Every target change starts a new learning period	Manual 10–15% steps with 7–10 days between them
Bulk-pausing ads on CTR	Ad groups left without active variants, delivery drops	Alert plus a manual rebuild of the responsive ad

The general principle: rules belong at the periphery of the account, where decisions are binary and do not touch learning — alerts, obvious waste, breakage detection. Anything that touches auction mechanics should be validated with a test rather than a rule; the correct method is in our walkthrough of Experiment Center and Google Ads experiments.

A safe rollout sequence

Write the rule as a sentence. “If X, then Y, because Z.” If you cannot finish the “because”, you do not need the rule.
Build it in email-only mode first. For two weeks it acts as an observer. You see which objects it would have hit and whether your thresholds are sane.
Run the preview. The interface offers a dry run showing exactly which items the rule would apply to right now. If that list surprises you, your conditions are wrong.
Narrow the scope with labels. A rule applied to “all campaigns” will eventually hit the campaign you launched yesterday, which it knows nothing about.
Enable the action and set notifications to “only when changes are made”. Then an email means an event, not background noise.
Record it in the team change log. Three months later nobody will remember why a campaign paused overnight if the rule is undocumented.
Review quarterly. Thresholds age along with your economics. A rule carrying last year’s CPA does quiet damage.

Rules, scripts and native automation: what handles what

Task	Tool	Why
Pause a keyword that burned three CPAs with nothing to show	Rule	Simple condition, all data inside the account
Stop a campaign when the product goes out of stock	Script	Requires an external data source
Manage bids inside the auction	Smart Bidding	A rule cannot see query and user context
Rewrite URLs across hundreds of ads	Google Ads Editor	One-off bulk work that needs a human review before upload
Alert on conversion drops across all client accounts	Rule at manager account level	One rule instead of thirty identical ones
Account for a two-day seasonal spike	Seasonality adjustment	A rule cannot warn the strategy in advance

Five mistakes that cost money

A rule with no volume floor. “Conversions = 0” without “cost above N” fires on every new keyword on its first day alive. Always add the second condition that filters out small numbers.

Account time zone is not your time zone. A “pause at 23:00” rule on an account set to a US time zone will stop delivery in the middle of your working day. Check the account setting before, not after.

Overlapping rules. One rule raises budget on CPA, another lowers it on spend. The budget oscillates daily, the strategy never exits learning, and the cause is invisible in every report except the change history.

“All campaigns, forever” scope. Every new campaign inherits old logic it was never designed for. Labels solve this in one line.

“Email every time”. Within a month all Google Ads mail is filtered into a folder nobody opens. Keep only “on changes” and “on errors”.

A 30-minute checklist

Open the rules section and list every existing rule: author, creation date, last time it fired.
Delete rules that never fired in a quarter or that fire every single run — the first is dead, the second is broken.
Verify the account time zone and each rule’s run time.
Confirm no rule changes a budget or a strategy target more than once a week.
Recalculate thresholds against your current target CPA and margin.
Add the missing alerts: zero conversions with meaningful spend, disapproved ads, enabled campaigns with no impressions.
Switch all notifications to “only when changes are made”.
Document the rule set in a shared team doc with one line of rationale each.

A rule only its author knows about is not automation — it is a delayed-action fault. Document every one of them.

Rules at manager account level

If you run more than five accounts, the highest-value place for rules is not inside an individual account but at the manager level. One rule instead of thirty identical ones means you adjust thresholds in a single place rather than touring thirty interfaces after every change in economics.

What genuinely belongs at manager level:

A single conversion-collapse alert — one email listing every account that spent yesterday and converted nothing. It replaces the morning tour of dashboards.
Portfolio-wide abnormal spend monitoring. Especially valuable after handing an account to a new buyer: the first week is when unplanned overspend most often happens.
Enabled campaigns with zero impressions. Across thirty accounts these appear constantly and nobody catches them manually.
Disapproved ad tracking. One daily email across the portfolio instead of thirty open tabs.

Manager-level rules do not replace account-level ones: actions tied to a specific account’s economics — CPA thresholds, budget steps — belong lower down. How to divide responsibility and the other quirks of manager accounts are covered in our piece on the Google Ads manager account.

What a healthy rule set looks like, in numbers

Take two hypothetical accounts with identical spend — $12,000 a month, a $40 target CPA, roughly 300 conversions a month. The only difference is their approach to automation.

Parameter	Account A: rules as autopilot	Account B: rules as circuit breaker
Active rules	19	7
Rules that change budgets	6, daily	1, weekly, 10% step
Notification-only rules	2	5
Executions per month	around 400	around 12
Change history entries	hundreds of automated edits	a handful
Bid strategy status	regularly back in learning	stable
Time spent explaining “what happened”	hours per month	minutes

Account A looks more automated, but cause and effect cannot be connected inside it: any change in results is explained by a dozen automated edits. Account B is duller, and every execution there is an event with a traceable cause, so any deviation is visible immediately.

The practical takeaway worth holding on to whenever you write a new rule: a rule’s value is measured not by how often it fires but by how many losses it prevented. A rule that fires daily is almost certainly describing normal account life rather than an emergency — which means the threshold is wrong.

What to do after a rule has done damage

The scenario is familiar: you wake up to half the campaigns paused, or a budget that doubled overnight. The recovery sequence:

Disable the rule, do not delete it. A deleted rule loses its run history, and that history is what you need for the post-mortem.
Open the change history for the period and filter to automated changes. It shows the exact time, object and edit.
Restore state manually — re-enable what was paused, return budgets to their previous values.
Do not compensate with an aggressive budget increase. A sharp jump after a dip is a second shock to the strategy; return to the previous level, not above it.
Let the strategy relearn before assessing results, and adjust for conversion lag when you do.
Rewrite the rule as notification-only and watch it for two weeks before letting it act again.

It is also worth checking whether the rule collided with a seasonality adjustment or a running experiment: edits from several sources inside one period are the hardest combination to untangle.

Where PPC Rebels fits

Automated rules are part of operational discipline, and discipline only pays off on an account that runs consistently and does not stop for external reasons. If you need a working account with history and a sane access structure, look at our Google Ads agency accounts; if you want to understand the full data-strategy-control loop, we also run Google Ads training and publish a hands-on Google Ads guide. Tools do not replace understanding — they amplify it. Related pieces from this series: primary and secondary conversions on the signal your algorithm learns from, and Google Ads brand lists on separating brand from non-brand traffic.

FAQ

How many automated rules can one account have?

The limit is high enough that you will not realistically hit it — a healthy account usually runs 5 to 15 live rules. The problem is almost never quantity; it is that old rules never get reviewed.

Do rules run when I am not logged in?

Yes. They execute on Google’s side on a schedule, independent of whether your interface is open. That is exactly why they cover nights and weekends.

Can a rule change my target CPA or ROAS?

The available actions for strategy targets are limited and depend on campaign type and strategy. Even where it is technically possible, automating target changes is a bad idea: each change restarts learning.

Why did a rule fire on an item that does not meet the conditions?

Usually the data window: the rule evaluated a different period than the one you are looking at. The second explanation is that the numbers at run time differed from current ones because conversions were written back later.

What is the rule preview for?

It is a dry run without applying actions: the system lists the items that would be affected right now. It is the only safe way to validate conditions before switching a rule on.

Can I limit a rule to one group of campaigns?

Yes, through filters and labels. A label is safer than a campaign list: a new campaign without that label stays outside the rule by default, which is usually what you want.

Do rules conflict with Smart Bidding?

Not inherently. Conflict appears when a rule frequently changes inputs — budget, target, campaign status. Observer rules and rules that remove obvious waste coexist with Smart Bidding without issue.

Rule or script — which should I use?

A rule when the logic is simple and all data lives inside Google Ads. A script when you need external sources, real computation, object creation or spreadsheet reporting. Start with rules and graduate to scripts when a rule can no longer express the task.

How do I know a rule is hurting me?

Three signs: it fires on nearly every run; the change history shows frequent budget or status edits; the bid strategy keeps returning to a learning state. Any one of them is a reason to switch it off and investigate.

Should I automate negative keywords with a rule?

You cannot add negatives automatically with a rule, and that is a feature rather than a limitation. Search term work needs semantic judgement — the method is described in our article on negative keywords and search term analysis.

Do rules work with Performance Max?

Partially. Campaign-level actions and some elements are available, but the depth of control in PMax is fundamentally lower than in Search. The real levers there are different — they are covered in our complete Performance Max guide.

Where should I start if the account has no rules at all?

With three observer rules: zero conversions at meaningful spend, abnormal daily spend, and enabled campaigns with no impressions. All three email-only. After two weeks you will know which of them deserve to become actions.