An audit should tell you what to fix, why it matters and how you will know the change worked. A longer checklist or a higher account score is not the outcome. Reliable measurement and better commercial decisions are.
This guide covers the same task whether you call it a Google Ads audit or an AdWords account audit. It separates platform requirements from practical recommendations, with links to Google's documentation where the distinction matters.
Start with the business question
Write down the target market, budget, sales cycle and definition of a qualified lead. Export recent campaign performance, conversion actions, search terms and change history. Include a comparable prior period and allow for conversion lag before judging recent results.
Use read-only access for the initial review. Make a separate change plan with an owner, evidence, priority and review date. Do not apply every recommendation simultaneously: you will lose the ability to tell which change helped.
Measurement: establish what the account is learning from
1. Test the complete conversion path
Follow an authorised test enquiry from the landing page to the confirmation, conversion event and CRM record. Check rejected forms and repeat visits as well as a successful submission. A button click is not proof that the enquiry was accepted or delivered.
Record: the conversion action, trigger, test time and receiving record. A status such as "No recent conversions" can reflect low activity; it does not establish that tracking is broken. Use Google's tracking guidance alongside an end-to-end test.
2. Separate enquiries from qualified opportunities
List primary and secondary conversion actions and the goals each campaign uses. Check whether bidding is rewarded for a lightweight interaction, a submitted form or a genuinely qualified lead.
Action: agree a qualification definition with sales. Where appropriate, connect CRM outcomes through offline conversion imports. Check consent, identifiers, timestamps and import diagnostics. Our HubSpot tracking guide provides additional implementation context.
3. Check duplicate counting and goal overlap
One customer action can be measured by several tags or conversion actions. Check direct Google Ads tracking, imported analytics events, calls and repeated confirmation-page visits together.
Action: choose counting settings appropriate to each action and use supported transaction or event identifiers where applicable. Do not assume an identifier deduplicates every action across every integration. Reconcile a small sample against the underlying records.
4. Review conversion lag, windows and values
A B2B opportunity may take longer to qualify than an ecommerce purchase. A short review window can make recent traffic look unproductive when conversions are still arriving.
Action: inspect time-to-conversion and CRM sales-cycle data. Choose supported windows suited to actual customer behaviour. For value-based bidding, confirm currency, accuracy and the business meaning of each value. There is no single correct window for every account.
Campaign settings: check where the budget can go
5. Confirm location and network settings
Compare intended service areas with actual user locations and location reports. Review location options, language settings, Search partners and other available network settings in the context of the campaign type.
Action: investigate spend outside the serviceable market. Do not exclude a network solely because its average differs from Search; assess lead quality, volume and measurement reliability.
6. Check structure against business constraints
Campaigns may need separate budgets, regions, languages, goals or products. Fragmentation is a concern when similar campaigns cannot collect useful data or require duplicated management—not when an arbitrary campaign-count limit is exceeded.
Action: map each campaign to a distinct business purpose. Consolidate only where shared goals, settings and reporting make sense. Preserve necessary budget and geographic controls.
7. Review Search and Performance Max together
Running both does not automatically mean you are bidding against yourself. Google documents Search and Performance Max prioritisation, including priority for an eligible Search keyword identical to the search term. Other situations depend on eligibility and prioritisation rules.
Action: inspect query coverage, brand traffic and incremental outcomes. Use brand exclusions or negatives where they support the strategy; do not automatically block every successful Search term from Performance Max.
8. Distinguish audience Observation from Targeting
An audience in Targeting can restrict who sees the ads; Observation lets you monitor an audience without narrowing reach in that way. They are not interchangeable settings for giving an audience higher priority.
Action: check intended reach before changing a setting. Follow Google's guidance, and review campaign-type behaviour and policy eligibility separately.
Queries and bidding: connect spend to relevant demand
9. Review search terms for intent, not just keywords
Group visible terms into buying intent, useful research, irrelevant intent and uncertain intent. Look at spend and qualified outcomes, not just whether the query contains a preferred keyword. Reporting does not reveal every query.
Action: investigate material mismatches and improve the relevant ad or landing page. Keep informational traffic only where its role and economics are justified. Our negative-keyword guide is a starting point, not a list to apply blindly.
10. Audit negative keywords and their scope
A useful negative list can be short or long. There is no minimum count that proves good management. Words such as "free", competitor names and research terms may be valuable or irrelevant depending on the offer.
Action: review account, campaign and ad-group exclusions, match behaviour and conflicts. Apply a shared exclusion only when it should affect every campaign within that setting's supported scope.
11. Evaluate match types with conversion evidence
The percentage of broad-match keywords is not, by itself, a diagnosis. Assess query relevance, bidding, budget and measurement quality together.
Action: investigate expensive irrelevant traffic over a sufficient period. Do not pause every zero-conversion keyword without considering spend, conversion lag and uncertainty. Document a test rather than promising one match type always wins.
12. Check bidding eligibility and target realism
Avoid treating one conversion threshold as a requirement for all Smart Bidding strategies. Requirements and recommended data volumes vary by campaign and strategy. Google's Target ROAS guidance describes campaign-specific requirements.
Action: compare targets with recent, lag-adjusted performance, value quality and diagnostics. Investigate unrealistic targets or unstable measurement before changes. Allow for learning and conversion delays when reviewing results.
13. Allocate budget using business outcomes
Cheap clicks and a low cost per form do not necessarily produce profitable customers. Compare qualified lead cost, opportunity value and capacity by campaign, market and service.
Action: identify where additional demand is economically useful. Treat impression-share diagnostics as context, not a command to raise budgets. Record assumptions when there are too few qualified outcomes for a reliable comparison.
Ads and landing pages: make the next step credible
14. Review responsive search ads beyond Ad Strength
Check whether headlines and descriptions explain the offer, work in different combinations and match the landing page. Review policy status and avoid unsupported claims.
Action: use Ad Strength as creative feedback, not an auction score. Google says it is not used to calculate Ad Rank or Quality Score. Test meaningful messaging differences and qualified outcomes; do not add near-identical ads to meet a supposed universal quota.
15. Check assets and destinations
Review sitelinks, callouts, business details, phone numbers and other relevant assets for eligibility and accuracy. Follow every destination, including mobile links, and check prices, offers and availability.
Action: replace misleading or broken destinations. Select assets useful to the buyer and appropriate to the campaign. Adding assets does not guarantee a fixed percentage increase in CTR.
16. Test landing-page relevance and usability
Explain who the offer is for, what is included, why the provider is credible and what happens after an enquiry. Test mobile layout, form errors, confirmation behaviour and consent handling.
Action: use field data and a reproducible lab test to investigate slow rendering and layout shifts. Record the device and page tested. Core Web Vitals describe useful experience measures, but a performance score alone does not prove conversion improvement.
17. Reconcile the account with sales feedback
Sample leads with sales: was the business a fit, could the team contact the person, did an opportunity form and was revenue recorded? Separate spam and duplicates.
Action: connect each change to a measurable business outcome. Keep a change log and revisit results after enough conversion lag and activity. An audit is a decision process, not a guarantee of immediate savings or additional leads.
Turn the findings into a prioritised action list
Fix broken measurement and customer journeys first. Then address documented irrelevant spend, weak offer alignment and settings that conflict with the business goal. Treat creative and bidding improvements as testable hypotheses.
For each finding, write: evidence → affected campaign or page → proposed change → owner → success measure → review date. Keep uncertainty visible when the sample is small.
Want a second opinion? Request a free Google Ads audit and tell us your market, spend range and main concern. Review Google Ads management services, our management pricing and how PPC management fees work before deciding whether ongoing support is appropriate.
Editorial review: 3 October 2026. Platform references were checked for this update. Examples are diagnostic guidance, not audited client results or performance guarantees.