Skip to content

Google Ads audit checklist: review your account systematically

maitiq · Published

A Google Ads audit checklist is useful when it records evidence instead of judging every setting against a blanket standard. maitiq runs the recurring areas at a configured cadence, daily by default. Below is the self-assessment, area by area; the service is described in maitiq’s Google Ads account audit.

What should you establish before reviewing the account?

Before you open the account, put the following in writing:

  • business goals and the selected conversion actions;
  • market, language and location logic;
  • the monthly, annual or period budget and who approves it;
  • campaign types and the job each one does;
  • brand and competitor rules;
  • the people responsible for marketing, analytics, the website and data protection;
  • the final approval authority for changes.

Without this framework the checklist records technical states but does not replace a business decision.

How do you assess a single audit item?

StatusMeaningNext step
Verified and compliantevidence meets the test defined in advancedocument it, review again later
Confirmed deviationevidence contradicts the approved stateprepare the risk and a proposal
Context decisiondata is available; whether it is right depends on the businessthe responsible person decides
Not verifiablethe source, period or definition is missingobtain evidence; do not treat it as passed

The four states stop missing data from quietly passing as unremarkable.

The working list is a copyable static template with the same placeholders for all ten areas:

IDAreaCheckEvidenceStatusRiskDecisionResponsibleVerificationRollbackReview date
01Accesscompany access, access levels and ownership of the accountfill inone of four statesfill infill innametest and success criterionreturn statedate
02Conversionbusiness goals, the selected conversion actions, counting method and attribution modelfill inone of four statesfill infill innameID reconciliationreturn statedate
03Settingsmarket, language, location option, networks and automatic recommendationsfill inone of four statesfill infill innamecontrol reportreturn statedate
04Structurecampaign role, naming logic and landing-page mappingfill inone of four statesfill infill innamebefore/after checkreturn statedate
05Search termsobserved queries, visibility limit and maturity of the samplefill inone of four statesfill infill innamenew reportreturn statedate
06Negativesmatch type, level, variants and conflict with positive keywordsfill inone of four statesfill infill innameconflict checkreturn statedate
07Budgetapproved period plan, budget type and spend limitsfill inone of four statesfill infill innameplan reconciliationreturn statedate
08Adsmessage, assets, destination path and platform noticesfill inone of four statesfill infill innamepath testreturn statedate
09Changesactor, purpose, approval and limits of the change historyfill inone of four statesfill infill innamefollow-up reportreturn statedate
10Reportingdefinition, period, time zone, currency and source per metricfill inone of four statesfill infill innamereconciliationreturn statedate

Copy the table and replace the placeholders with the evidence for your account, instead of pre-empting a finding with default values.

What maitiq turns the checklist into

This is how maitiq keeps this under review: areas 5 to 7 run as recurring workflows at a configured cadence, daily by default. The search-term workflow checks search terms and negatives with a conflict check that compares text against existing positive and negative keywords (where they overlap, a broad or phrase exclusion is downgraded to exact or discarded; Google’s own matching rules must be checked separately from this), and the pacing workflow checks the budget against the approved period plan. Every change applied through maitiq is logged per applied action with the decision, the time, and the old and new value where such a comparison of values exists; this logging is not itself a daily audit. Every proposal names the proposed action, the rationale and the supporting data; implementation is decided by a manual decision or a configured, bounded Autopilot approval, and the application is logged separately. An explicit approval to exclude specified keywords already covers the guarded exclusion action; no second authorisation is needed. The audit itself stays read-only and authorises no change.

1. Are the account, access and ownership settled?

Google Ads has five access levels. The interface names them “Email only”, “Billing”, “Read only”, “Standard” and “Admin”. “Read only” can view campaigns and use planning tools but cannot edit campaigns. Users with “Read only” or “Standard” can also grant email-only access; “Admin” manages the broader invitations and access levels.

So record the customer ID, all direct users with their level, linked manager accounts and their ownership status:

  • Is there an active admin on a controlled company address?
  • Is there a second person who can act if someone is unavailable?
  • Which agency or tool gains access through a manager account?
  • Does a manager account hold ownership rights, and are they needed?
  • Who controls payment, Tag Manager, analytics and other systems?
  • Are former employees or contractors still active?

According to Google, only one directly linked manager account is designated owner of a client account; rights can be inherited higher up in its hierarchy. The client account keeps its data and can remove the link subject to checking billing and conversion dependencies (see switching agencies). That says nothing about the number of direct users; additional rights exist only when they are understood and needed.

2. Is the account optimising for the right outcome?

List every conversion action with its name, source, category, use, counting method, value, attribution model, window, status and the campaigns that use it.

Primary actions appear in the “Conversions” column and are used for bid optimisation provided the standard goal they belong to is used for bidding. Actions with the use “Secondary (observe only)” appear exclusively in “All conversions” and do not feed into bid optimisation – unless they belong to a custom goal that the campaign uses. Check the goals, therefore, not just the role of each action. For counting, Google distinguishes “Every” and “One” conversion per ad interaction. This platform count does not replace server-side idempotency: unique storage in your own system and counting in Google Ads are two separate controls.

Two points have changed compared with older checklists. First, of the attribution models Google now supports only “Last click” and “Data-driven”; “First click”, “Linear”, “Time decay” and “Position-based” are no longer supported, and “Data-driven” is the default for most conversion actions. Second, Google’s EU user consent policy applies explicitly to the European Economic Area, the United Kingdom and Switzerland. Record how consent is collected, documented and withdrawn. The method is covered in more depth in the conversion tracking audit.

3. Have the campaign settings been chosen deliberately?

For each campaign, record the type, status, goal, language, location option, networks, schedule, bidding strategy, budget, conversion goals and landing page. Do not ask whether a setting deviates from a supposed best practice, but whether it fits the documented purpose.

Three settings deserve particular attention in the Swiss context:

  • Location. Google currently offers “Presence or interest” and “Presence”. Both options can be right. Record which one was chosen and why.
  • Networks. Search Network partners are included by default in a new Search campaign and are deselected per campaign in the network setting. Since August 2025, Google has documented a complete placement report for Search, Shopping and App campaigns: it lists the websites in the Search Network with the impressions for each website. A website appears only if it reaches at least one threshold in the selected period (at least 5 impressions, more than 0 clicks, more than 0 video views or more than 0 conversions); a missing row therefore does not prove zero activity.
  • Automatically applied recommendations. According to Google, “Apply recommendations automatically” is available only at account level, and Google states that it does not increase the budget. Record every active type, who enabled it, its purpose and the most recent applications. The “History” tab and the “Change history” page show what was applied.

4. Is the campaign structure understandable?

Document the separation of brand and non-brand where it is commercially relevant, unambiguous names, landing-page mapping, separate control for markets or products with their own budgets, ongoing tests including the hypothesis, and paused legacy structures with the reason for them.

Do not restructure for a neater diagram: rebuilds make comparisons harder and influence bidding strategies. An audit first explains how the existing account works.

5. Which search queries trigger your ads?

Keywords are targeting definitions; search terms are the actual queries, as far as Google makes them visible. You can find the search terms report under “Campaigns” ▸ “Insights and reports” ▸ “Search terms”.

What matters is the limit of this report: Google lists search terms that were used by a significant number of people and omits terms with too few searches for data protection reasons. “Not in the report” therefore does not mean “not searched for”. Check irrelevant or informational terms, new profitable patterns, brand and competitor queries, match type expansion, the partner share and conversion delay.

Not every row without a conversion belongs straight on an exclusion list. Period, clicks, costs, measurement quality and business context all belong in the same decision.

6. Are negative keywords effective and free of conflicts?

Negative keywords behave differently from positive keywords. For them, Google automatically accounts for capitalisation and spelling mistakes. By contrast, you have to add synonyms and singular or plural forms yourself, and close variants are not excluded automatically. Check:

  • the match type of each excluded term;
  • the level: ad group, campaign, account or shared list;
  • whether the list is actually applied to the campaigns it is intended for;
  • conflicts with positive keywords;
  • singular, plural, word order and language variants;
  • outdated exclusions after changes to the product range or offer;
  • the documented reason and approval for each entry.

A conflict can block valuable demand. “More negative keywords” is not a quality goal; relevance, evidence and controlled reach are what matter. The decision logic is covered in more depth in the guide to search terms and exclusions.

7. Is the budget tied to a real plan?

First document the approved period plan and compare it with the account. For most campaigns with an average daily budget, Google states daily spend of up to twice that amount and, as the monthly limit, the average daily budget multiplied by the average number of days per month (30.4); the monthly calculation assumes the budget is unchanged for the period. Other budget and bidding modes and budget changes follow their documented rules.

Also check the budget type. Alongside the average daily budget, Google documents a shared budget across several campaigns and a campaign total budget with a start and end date. A shared budget automatically reallocates amounts that are not fully used to campaigns that are limited by budget, is available only for Search, Shopping, Display and Video campaigns, and is a single budget shared by several campaigns rather than one budget per campaign. Also record the source of the plan, budget changes, monthly spend, projection, seasonality, currency and account time zone. A historical average is a comparison, not automatically the current business plan.

8. Are ads and landing pages consistent?

Check the status, policy notices, assets, final URL, mobile rendering, message, form and conversion path. Does the ad match the search intent? Is the conversion triggered only after the actual outcome?

Ad strength is rated from “Incomplete” to “Excellent” and indicates how varied and combinable your assets are. Google states explicitly that it does not feed into the calculation of Ad Rank, Quality Score or auction wins. A high rating is therefore no proof of success, and a high click-through rate does not prove that the contacts are qualified.

9. What does the change history show – and what does it not?

Under “Campaigns” ▸ “Change history”, Google Ads provides the changes of the past two years. The “User” column shows the address of the person who made a change in the interface; automatic changes and changes triggered through interfaces appear under system or tool names. Check major budget and bid target changes, keywords, automatically applied recommendations, conversion goals, access rights and how closely they coincide in time with performance movements.

Two limits belong in the documentation. First, Google limits the detail per entry, in particular when a change affects a very large number of objects; when the necessary detail is therefore unavailable, the finding is “not verifiable”. Second, closeness in time is not causation. Record parallel changes, seasonality, competition and tracking interventions, and add the business purpose, approval, risk and verification to the history.

10. Is the report fit for decision-making?

For every finding, the report needs the question, evidence, period, status, risk, recommended decision, the responsible person together with the approval, the verification date and a rollback or stop plan where applicable.

Also define which point in time each metric refers to. Record explicitly whether a conversion is attributed to the date of the ad interaction or the date of the completion, which conversion window applies and which time zone and currency the account uses. As long as this convention is not in writing, every period comparison is “not verifiable”.

Why does the checklist name no thresholds?

Platform metrics look like audit results but are not: according to Google, the optimisation score is an estimate between 0 and 100%, recommendations may be dismissed, and it plays no part in Quality Score; ad strength does not influence Ad Rank or the auction. Google’s own averages describe a platform, not your account – which is why the checklist works without a score, interval or metric threshold.

Frequently asked questions

How often should a Google Ads audit take place?

There is no universal interval. Major changes to the website, tracking, agency, budget or bidding logic justify a targeted review; a recurring rhythm depends on risk and the volume of change. For the recurring areas, maitiq runs the checks at a configured cadence, daily by default.

Must an audit contain changes?

No. An audit produces findings and a basis for decisions. An identified shortcoming is not yet approval to change anything.

What does “not verifiable” mean?

The required evidence is missing, too short, not clearly attributable or technically unavailable – and it must count neither as zero nor as passed.

Which details go out of date fastest?

Interface labels, recommendation types and selection lists; four attribution models that were once common, for example, are no longer supported. Re-check every interface detail immediately before you use it.

Can a checklist replace a professional audit?

It structures an internal stocktake. Where measurement is complex, several accounts are involved, spending responsibility is high or governance is unclear, the read-only maitiq audit delivers the same ten areas with evidence, a coverage status for each check and prioritised proposals – without changing a single setting.

Sources and how to read them

The sources are Google platform documentation and Google’s user consent policy; they explain features, limits and the legal context. They are not independent legal authorities, nor supplier or association evidence of market prices or success. Information about how maitiq works is available at maitiq.com.

Have maitiq review your specific case.