Google Ads Audit Checklist for Agencies (2026)
A 12-point Google Ads audit checklist built from 400+ account reviews. Covers conversion tracking, bidding, match types, attribution, and tag health — for agencies managing multiple clients.
Most agencies audit Google Ads the same way: pull a performance report, look at ROAS or CPA, flag the campaigns that are underperforming, and call it done. I've reviewed hundreds of accounts that passed that kind of audit — and found conversion tracking problems that had been inflating reported performance for months, negative keyword lists with 40 entries blocking converting queries, and Smart Bidding strategies optimizing to the wrong event entirely.
A real Google Ads audit checklist for agencies doesn't start with performance. It starts with the plumbing. Once you know the data is trustworthy, the performance conversation is productive. Until then, you're debating numbers built on a broken foundation.
This is the 12-point framework I use across every Google Ads account review, whether it's a $5K/month e-commerce account or a $200K/month enterprise client. The order matters — work top to bottom.
Why Most Agency Google Ads Audits Miss Critical Issues
The failure mode I see most often isn't laziness. It's the wrong starting point.
Agencies open the Ads interface, look at the last 30 days of data, and diagnose performance. But if conversion tracking is broken — and in my experience, it's misconfigured in roughly 60% of accounts I inherit — then every performance conclusion drawn from that data is wrong. You're optimizing against noise.
The second failure is single-platform thinking. Google Ads doesn't exist in isolation. It depends on GTM firing correctly, GA4 receiving the right events, Search Console confirming organic traffic trends, and the landing pages actually converting. An audit that stays inside the Ads interface will miss the issues that live in those seams.
The third is confusing activity with accuracy. Running a search term report and adding negatives is an activity. Confirming that your conversion tag fires exactly once on the order confirmation page and nowhere else is accuracy work. Both matter. Most audits only do the first.
The 12-Point Google Ads Audit Checklist
1. Conversion Tracking Accuracy
What to check: Pull the conversion action list. For each action marked "Include in conversions," verify it's tagged and verified, not just imported. Cross-reference the Ads conversion count against GA4 key event count for the same date range.
What good looks like: Primary conversion action is verified, firing on the correct page, and the count discrepancy between Ads and GA4 is under 15%. Secondary actions (phone clicks, form starts) are excluded from the primary conversion column.
What bad looks like: Conversion counts in Ads are 30–50% higher than GA4 — typically caused by the tag firing on multiple pages or duplicate actions. Or the reverse: GA4 shows conversions but Ads shows zero, meaning auto-tagging is off and GCLID isn't being passed.
Why it matters: Smart Bidding trains on conversion data. If that data is wrong, your bidding algorithm is learning from fiction. I've seen accounts where tROAS was hitting its target perfectly — against a conversion event that was firing on every page load.
2. Campaign Structure Alignment
What to check: Are campaigns segmented by objective (acquisition vs. retention vs. brand defense), network (Search vs. Shopping vs. Display vs. PMax), and audience temperature? Or are all objectives crammed into one campaign because "it was easier to set up that way"?
What good looks like: Each campaign has a clear, single objective. Brand and non-brand traffic are separated. Performance Max is isolated from Standard Shopping so you can see cannibalization.
What bad looks like: One campaign running Search and Display simultaneously. Non-brand and brand keywords in the same ad group competing for budget. PMax running alongside Shopping with no exclusions and no visibility into which is actually driving revenue.
Why it matters: Campaign structure determines what Smart Bidding can optimize for, how you allocate budget with intention, and whether your reporting is legible. Poor structure doesn't just make the account harder to manage — it actively degrades performance by giving the algorithm conflicting signals.
3. Keyword Match Type Strategy
What to check: What's the match type distribution? Is it intentional, or is it a historical accident? Accounts set up before 2021 often have Broad Match Modified keywords that defaulted to Phrase when Google changed the behavior — and nobody noticed.
What good looks like: A documented match type strategy. If you're using Broad + Smart Bidding (the approach Google recommends and that works well at scale), the account has sufficient conversion volume and the search term reports confirm relevant traffic. If you're using Phrase/Exact, the keyword list is actively maintained and the search terms are tight.
What bad looks like: A mix of all three match types with no strategy behind it. Broad match keywords in a manual CPC campaign with no negative list and no conversion history — burning budget on irrelevant queries.
Why it matters: Match type is the primary lever controlling traffic quality. An undisciplined match type strategy is usually the fastest path to wasted spend.
4. Negative Keyword Lists
What to check: Does the account have a shared negative keyword list applied at the account level? Campaign-level negatives to prevent cross-campaign cannibalization? And critically — are any negatives accidentally blocking converting queries?
What good looks like: A tiered negative structure: account-level list (job seekers, students, "free," generic research terms), campaign-level negatives for intent separation, and a documented review process for adding new negatives from the search term report.
What bad looks like: No shared negative list — every campaign managing negatives independently, creating inconsistent coverage. Or a shared list that was bulk-imported from a template without verification, now blocking "buy," "pricing," and other high-intent terms I've seen killed this way.
Why it matters: Negative keyword conflicts are invisible in the main dashboard. You won't see a warning that your account-level negative is suppressing a campaign's best converting query. You have to look for it.
5. Audience Targeting Configuration
What to check: Are remarketing audiences built and applied? Are Customer Match lists current (check membership counts — stale lists shrink to zero)? Are in-market and affinity segments in Observation mode on Search campaigns, not Targeting mode?
What good looks like: Remarketing audiences segmented by funnel stage (site visitors, cart abandoners, past purchasers). Bid adjustments informed by actual performance data from the observation period. Customer Match refreshed at least quarterly.
What bad looks like: Audiences in Targeting mode on Search — which restricts the campaign to only showing ads to people in that audience segment. This is a common setup mistake that tanks impression volume without any error message. Also: Customer Match lists with 0 active members that nobody noticed had expired.
Why it matters: Audience signals are increasingly important as match types broaden. Without them, you're flying blind on who you're actually reaching.
6. Ad Copy and Asset Quality
What to check: Do all active ad groups have at least one RSA with Good or Excellent ad strength? Are headlines pinned where they should be? Are the assets (formerly extensions) complete — sitelinks, callouts, structured snippets, and where applicable, call and image assets?
What good looks like: Each RSA has 15 headlines and 4 descriptions, at least 2–3 include keywords, at least one is pinned to position 1, and no two headlines are redundant. Asset coverage is complete. Ad strength is Good or Excellent across active ad groups.
What bad looks like: Ad groups with a single RSA at "Poor" strength that was auto-created by Google's recommendation engine and never touched. No callout assets. Sitelinks pointing to the homepage from every campaign including tightly-themed product campaigns.
Why it matters: Google's auction considers ad relevance and expected CTR as components of Quality Score. Weak ad copy doesn't just get lower CTR — it raises your CPCs.
7. Landing Page Relevance
What to check: Does the landing page match the ad group's keyword theme? Is the page actually converting in GA4, or does traffic from this ad group have a significantly higher bounce rate than the site average? Does the page load in under 3 seconds on mobile?
What good looks like: The landing page uses the same language as the ad. The offer in the ad matches what's on the page. Mobile load time is under 3 seconds. The conversion rate for paid traffic to this page is at or above the site average for organic.
What bad looks like: A campaign about "enterprise HR software" sending traffic to the homepage. Or a seasonal promotion ad pointing to a page that's still running last quarter's offer. These are always attributed to "Google Ads performance" when the problem is the landing page.
Why it matters: Landing page experience is a Quality Score component. Beyond that, the best-optimized Ads account in the world can't fix a page that doesn't convert. I've doubled client ROAS by fixing the landing page without touching a single bid.
8. Quality Score Patterns
What to check: Pull the Quality Score columns at the keyword level. Look for keywords with QS below 5, particularly those with high spend. Diagnose whether the issue is Expected CTR, Ad Relevance, or Landing Page Experience.
What good looks like: Active keywords are at QS 7 or above. The distribution skews toward 8–10 for core converting terms. Low-QS keywords are either paused or have a documented reason for running.
What bad looks like: High-spend keywords at QS 3–4. Keywords with "Below Average" landing page experience that nobody has investigated. This directly inflates CPC — a QS 4 keyword can cost 2–3× what a QS 8 keyword costs for the same position.
Why it matters: Quality Score is Google's signal that your keyword, ad, and landing page are relevant to each other. Low QS means you're paying more for every click than your competitors with better-configured accounts.
9. Budget Allocation
What to check: Which campaigns are hitting their daily budget limit more than 10 days per month? Those campaigns are leaving revenue on the table. Which campaigns are underspending vs. their allocation? Is the budget distribution aligned with actual contribution to business goals?
What good looks like: High-ROAS campaigns have uncapped or ample budgets. Shared budgets are used intentionally, not as a default. Budget decisions are driven by performance data, not by equal distribution across campaigns.
What bad looks like: Your best-performing campaign is budget-limited every day while a low-performing brand campaign has twice the budget. Or shared budgets allocated across dissimilar campaign types, letting the algorithm favor cheap clicks over valuable ones.
Why it matters: Budget-limited campaigns can't scale. If your best campaign hits its daily budget by 2pm, you're capping your own results — regardless of how well the targeting and bidding are configured.
10. Bidding Strategy Alignment
What to check: Is the bidding strategy appropriate for the campaign's conversion volume and objective? Are tCPA or tROAS targets set to realistic values based on historical data? Are any campaigns stuck in the learning phase?
What good looks like: Smart Bidding campaigns have at least 30–50 conversions per month to inform the algorithm. Targets are set within 10–15% of historical averages to start, then adjusted incrementally. No campaign has been in "learning" for more than 7 days.
What bad looks like: A campaign with 8 conversions in the last 30 days running Target ROAS. The algorithm has no data to learn from and will oscillate wildly. Or a tCPA target set at $15 for a product that has never converted below $45 — the campaign starves itself trying to hit an impossible number.
Why it matters: Smart Bidding is powerful when given good signals. When it's starved of data or given impossible targets, it actively makes performance worse. The strategy needs to match the account's maturity.
11. Attribution Model
What to check: What attribution model is the account using? Is it appropriate for the sales cycle length and the conversion volume? If using data-driven attribution, does the account have sufficient volume to support it (typically 300+ conversions per month)?
What good looks like: Data-driven attribution on high-volume accounts. Last-click only on accounts where the conversion window is short and the path is simple. The team understands what model is active and why — it wasn't just left on the default.
What bad looks like: Last-click attribution on a B2B account with a 30-day sales cycle and multiple touchpoints — systematically undervaluing top-of-funnel campaigns that drive consideration. Or data-driven attribution on an account with 40 conversions per month, where the model is making up segments with insufficient data.
Why it matters: Attribution model determines what Smart Bidding learns to optimize for. It also shapes how you evaluate campaign performance. An agency making budget decisions based on last-click attribution in a multi-touch environment is defunding their own best campaigns.
12. Tag and Pixel Health
What to check: Is the Google Ads conversion tag firing via GTM, hardcoded on the page, or both? Is the Conversion Linker tag present on all pages? Is auto-tagging enabled? Does the GCLID parameter appear in GA4 session reports, confirming the handoff is working?
What good looks like: One tag implementation method (GTM preferred). Conversion Linker present and firing. Auto-tagging enabled. GCLID visible in GA4. Tag Assistant shows clean firing on the conversion page with no duplicate events.
What bad looks like: Both a GTM-fired tag and a hardcoded page tag firing simultaneously — doubling conversion counts. No Conversion Linker, which breaks attribution on iOS Safari and other browsers with strict cookie policies. Auto-tagging disabled with no documented reason, meaning Google Ads can't see which clicks led to sessions.
Why it matters: Tag health is the foundation of everything else. If the tag is broken, the data is broken. If the data is broken, every optimization decision is wrong. This is always the first thing I check on a new account, and I find problems here more often than anywhere else.
How to Turn Audit Findings into a Client Deliverable
An audit that lives in a spreadsheet nobody reads is worthless. Here's the structure that actually gets action:
Lead with the business impact, not the technical finding. Don't open with "your Conversion Linker tag is missing." Open with "your Ads account is likely underreporting conversions from Safari users, which we estimate is affecting 20–30% of your traffic based on your device mix." The technical explanation comes second.
Tier everything. Critical — fix this before we do anything else. Recommended — fix this month. Informational — context for the record. Clients who receive a 30-item list of equal weight fix zero items. Clients who receive a 3-item Critical list fix all three.
Show before and after. For every major finding, project what changes after the fix. "Removing these 12 overlapping negative keywords is estimated to restore impressions to 3 previously blocked converting queries. Based on historical CPL for those keywords, this could generate an additional 8–12 leads per month."
Deliver as PDF, not slides. PDFs get saved, forwarded, and referenced. Slide decks get presented once and forgotten. Your audit report should be something the client's CFO can read six months later.
Automating Your Google Ads Audit Process
Running this 12-point framework manually takes 3–5 hours per account. At 10 clients, that's 30–50 hours per month of audit work before you've done any actual optimization.
The data collection layer — pulling reports from Ads, cross-referencing GA4, checking GTM, reviewing Search Console — is automatable. The judgment layer (what does this mean for this client, what's the priority, what do we recommend) is not.
StackXray handles the data collection and cross-platform comparison automatically. It connects to all six Google platforms via OAuth, runs the full audit framework simultaneously, and surfaces findings ranked by severity — so you walk in with the diagnostic already done and spend your time on the analysis and client conversation.
For agencies running recurring audits across a portfolio, automating the data layer is the difference between audits that happen consistently and audits that get deprioritized when the team is busy.
Run This Audit on Your Next Client Account — Free
StackXray connects to GA4, GTM, Google Ads, Search Console, Core Web Vitals, and Merchant Center simultaneously — and cross-references findings across all six platforms in one pass. No spreadsheets. No switching between dashboards. The first audit is free.
Audit your first client account free →
Final Thoughts
The 12 points above aren't theoretical — they're the specific checks I've run on every Google Ads account I've touched for the last 14 years, from JP Morgan Chase's digital analytics infrastructure to small agency clients spending $5K a month.
The common thread across every broken account I've inherited: the problems weren't hidden. They were just in places nobody had looked. Conversion tags firing twice. Negative keywords blocking top performers. Smart Bidding starved of data and oscillating wildly against an impossible target.
A structured audit framework makes those problems visible before they compound. Run this against your next client account. Most agencies find at least two or three Critical issues in accounts they thought were clean.
If you want to run the cross-platform checks — the ones that require comparing Ads data against GA4, GTM, and Search Console simultaneously — StackXray does that automatically. Your first audit is free.
Run your first Google marketing stack audit free.
StackXray audits GA4, GTM, Google Ads, Search Console, Core Web Vitals, and Merchant Center in one pass — with Claude AI cross-referencing findings across all six platforms simultaneously.
Audit your first client account free14+ years in enterprise martech across Adobe AEP, GA4, GTM, and Google Ads. Former Solutions Consultant at Adobe, embedded with digital agencies across the enterprise stack. Founder of StackXray.