Google Ads Audit: How I Audit a Google Ads Account End-to-End

If you Google "google ads audit," the first ten results all say the same thing.

They tell you to check your Quality Score. Review your negative keywords. Look at your conversion tracking. Verify your bid strategies. They give you a 47-point checklist and a free PDF.

I want to skip that.

Google itself reports that businesses earn $8 in profit for every $1 spent on Google Ads on average. Most of the accounts I audit are nowhere close to that number. They don't fail because the bid strategy is wrong. They fail because the structure is broken, the keyword intent is mismatched, and the budget is silently being eaten by placements that were never going to convert. Quality Score is a downstream symptom. If you fix it without fixing the structure underneath, it comes back inside a month.

This post is the actual order I run when a new Google Ads account lands on my desk. Eight stages. The patterns I see most often. The diagnostic at each step that tells me whether to keep going or write the first finding.

You can run it yourself in three to five hours. The diagnostic takes longer to build instinct for. The structure does not.

Why most Google Ads audits skip the parts that actually matter

Two failure modes show up in nearly every Google audit I read.

The first is the surface audit. The auditor opens the account, runs Google's own pre-built recommendations, exports the keyword report, screenshots a few high-CPC search terms, and writes "your negative keyword list needs work." The findings are real. They're also shallow. The auditor has rearranged the furniture in a house with broken plumbing.

The second is the agency sales audit. The auditor is also the salesperson for the retainer that comes after the audit. The findings get shaped to create just enough anxiety to close. "Your account has 47 critical issues. We can fix them for $4,500 a month." The audit was never the product. The retainer was the product. The agency-audit-as-sales-funnel mechanic gets a closer look in this breakdown of how free audits actually work.

What's missing in both is the diagnostic order.

Quality Score is a symptom. Negative keywords are a symptom. Performance Max underperformance is a symptom. They show up because something upstream is broken: the structure does not match how the site sells, the nomenclature is unreadable, or the campaigns are bidding against each other and inflating the account's own CPCs.

Fix upstream, downstream resolves itself. Fix downstream, upstream keeps generating new symptoms.

For the broader category critique that frames why most Google audits operate this way, see the Honest Audit Manifesto.

Stage 1: Conversion tracking and data trust

Before I trust any number on the dashboard, I check that the data is real.

This is the gate. If conversion tracking is broken, every recommendation downstream is built on fiction. I'd be optimizing for a fantasy.

What I check:

• Is the Google Ads conversion pixel firing on the right page (the thank-you page, not the cart)?
• Is GA4 connected and importing conversions correctly?
• Are enhanced conversions enabled with first-party data?
• Are conversion values passing through correctly, or is every purchase counted as $1?
• Is offline conversion import set up if the brand has a meaningful phone or email sales channel?

The pattern I see most often: a brand reports 4x ROAS on Google. Their P&L disagrees. When I dig in, the conversion event was firing on every cart visit, not on purchase. The account had been optimized for cart visits for nine months. The algorithm got very good at finding people who clicked Add to Cart and never bought.

That is not a bid strategy problem. That is a data problem. And it makes every other number in the account a lie.

If stage 1 fails, the audit stops here and the first finding is written.

Even when conversion tracking is set up correctly inside Google Ads, the platform's last-click attribution still causes systematic ROAS inflation against the founder's P&L. The audit method catches both the platform-level tracking issues and the cross-platform attribution overlap. The two leaks compound. Fix Stage 1 first, then reconcile to the bank account.

For the structural reason cross-platform numbers never reconcile cleanly, see why your ad reports show 3 different numbers. Google's data-driven attribution only sees Google. Meta's 7-day-click plus 1-day-view sees more than it should. Shopify counts orders. The math is broken on purpose.

Stage 2: Account structure

The most under-discussed input to good Google Ads management is the campaign structure. Most agency audits skip it.

My standard: the campaign structure should mirror the website's information architecture exactly.

If the site has 7 categories, there should be 7 campaigns in Google Ads. If those 7 categories break into 21 subcategories, there should be 21 ad groups under those campaigns. If the catalog has 400 to 500 SKUs, those SKUs should be mapped at the product or product-group level inside the ad groups.

Why this matters. The structure is what gives you control. Without it, you cannot answer the simplest question an operator should be able to answer in 30 seconds: "Which category is making us the most money this month, and at what return?"

When the structure mirrors the site, that question takes one filter. When it doesn't, you spend an afternoon building a custom report and still don't trust the answer.

The pattern I see most often: a brand started with 2 SKUs and one campaign three years ago. They added products, duplicated campaigns, created variants of variants, and now nobody can read the account. The structure didn't fail. It was never designed.

A useful test. Open the campaigns view in Google Ads. Sort by spend. If the top five campaigns share names, contain duplicate keywords, or split a single category across multiple campaigns, you have a structure problem. The audit notes that as a finding before any keyword analysis happens.

Stage 3: Nomenclature and readability

Naming conventions sound trivial. They're not.

A campaign name like Search 5 tells me nothing. A campaign name like SRCH_Skincare_Cleanser_BMM_US_Mar26 tells me the campaign type, the category, the subcategory, the match type, the geo, and the launch month. I can audit ten campaigns at a glance instead of clicking into each one.

A simple convention I use as a starting point:

Apply the convention once, at campaign creation, and the account becomes readable forever. Skip it, and the readability problem compounds with every new campaign launched into the junk drawer.

The benchmark: take the account and ask whoever runs it, "Show me the top performing Shopping campaign for the supplements category, launched in Q1." If they can do it in under 30 seconds without opening any individual campaign, the nomenclature is working. If not, that's the next finding.

Stage 4: Cannibalization audit

This is the stage that surfaces the leak most agency audits miss entirely.

When two campaigns target similar keywords or overlapping audiences, they bid against each other. Your own campaigns drive up your own CPCs. The auction tightens because you are the auction.

How it shows up in the data: the account's average CPC drifts up over six to twelve months. The operator sees the new CPC and treats it as the new normal. They never realize that without the cannibalization, the same keyword would clear at 30 to 40% lower CPC.

A composite example. A few months ago I audited an electronics brand running about $25K/month on Google. Founder named Priya. She had eight Search campaigns covering roughly the same product categories, set up over two years, each with overlapping keyword lists. Her average account CPC was $2.10. After deduplicating the keywords across campaigns, consolidating the structure, and adding cross-campaign negatives, the account's clearing CPC dropped to $1.40 inside three weeks. Same volume. 33% less spend. No bid strategy change.

That money was not lost to Google. It was lost to her own structure.

What I check:

• Run a search-term overlap report across campaigns
• Check if the same query is triggering multiple campaigns
• Verify whether match types are layered (broad in one campaign, exact in another, with no negatives between them)
• Look at brand-vs-non-brand split (most accounts pay too much for their own brand traffic because of cannibalization between brand and category campaigns)

If the account has been running for more than 12 months and nobody has done a cannibalization audit, the leak is almost always there. Almost always.

For the full defense of brand keyword campaigns as defensive infrastructure, see the Self-Defense Campaign methodology.

Stage 5: Budget allocation by campaign type

Now I look at where the money is going by campaign type.

Google's own dashboard recommends putting 25 to 30% of budget into Performance Max. Their CSMs push the same advice. I disagree, and the disagreement matters.

Performance Max distributes spend across Search, Shopping, Display, YouTube, Discovery, and Maps. The placements that drive most of the conversions on a typical paid ads account are Search and Shopping. The placements where Performance Max often spends a meaningful share of the budget are Display and YouTube.

The order of priority for Google ads budget:

1. Search (highest intent)
2. Shopping (high intent, product-specific)
3. Display and YouTube (retargeting only, not acquisition)
4. Performance Max (only after the manual structure is healthy and only as an additive layer, not as the primary engine)

If a brand is spending more on Performance Max than on manual Search and Shopping combined, that is a finding before any other diagnostic runs. For the full PMax-specific audit, see the Performance Max audit checklist. For a deeper breakdown of what monthly Google Ads spend looks like by brand stage, see the cost breakdown by spend tier.

Stage 6: Search term hygiene and negative keywords

Google is the most intent-based platform in performance marketing. The reader has typed exactly what they want. If your keyword is matching their query, you have the audience. If your keyword is matching the wrong query, you are paying for clicks that will never convert.

What I check:

• Pull the Search Terms report for the last 90 days
• Sort by spend
• Identify queries that triggered ads but are clearly off-intent (the wrong category, wrong audience, wrong product)
• Check the negative keyword list at campaign and account level
• Look for cross-match-type leaks (a broad match keyword catching queries that an exact match keyword in another campaign should own)

The pattern I see most often: a brand spending $40K/month on Google Search has 800 to 1,200 unique search queries triggering ads each month. Of those, 200 to 300 are clearly off-intent. The negative keyword list has 47 entries. That's the leak.

Search Term Negation should be a daily habit from day two of running a campaign. Most accounts I audit have not added a negative keyword in the last 60 days. That alone tells me the optimization cadence has stopped.

For the full defense of when broad match works and when it bleeds money, see the Match-Type Discipline.

Stage 7: Ad and landing page congruency

This is where the fastest CRO win usually lives, and it's the one most "Google audits" never check.

The principle: the ad communication and the landing page must tell the same story.

If the ad headline promises "30% off your first order," the landing page headline should reinforce that exact offer. If the ad shows a specific product, the landing page should show that exact product, not a category page where the user has to find it. If the ad emphasizes free shipping, the landing page should display the free shipping badge in the first viewport.

The more discrepancy between the ad and the landing page, the lower the conversion rate. The more congruency, the higher.

I have seen the extreme version of this exact pattern in real audits, where 6 of 18 active ads were driving traffic to 404 pages. That is the obvious version of an ad-page congruency leak. The more common version is subtler: the ad makes a promise the page does not honor, the user feels the small disconnect, and they bounce.

What I check, in order:

• Click every active ad and land on the page it serves
• Compare the ad headline to the page headline (exact words, similar words, different words?)
• Compare the offer in the ad to the offer on the page
• Check mobile rendering of the landing page (most ads serve to mobile, most landing pages get tested on desktop)
• Check page load time (anything above 3 seconds on mobile is a leak)

This is the half of the audit that lives between the ad and the cash register. Most agency audits never go there. That's why their findings stop short of the actual revenue lever. The cross-platform reality check usually points to a checkout break. For the full audit method, see the mobile-first CRO diagnostic.

Stage 8: Quality Score and bid strategy

I save Quality Score for last on purpose.

Quality Score is a downstream metric. It reflects expected click-through rate, ad relevance, and landing page experience. Each of those three is the product of decisions made at stages 2, 3, 6, and 7. If you fix the structure, the keywords, and the ad-page congruency, Quality Score lifts on its own.

Auditing Quality Score in isolation is like grading a student on their final exam without checking what they were taught all semester.

What I look at, briefly:

• Quality Score distribution across the top spending keywords
• The three component scores (CTR, ad relevance, landing page experience) so I know which one is dragging the average down
• The bid strategy in use (Manual CPC, Maximize Conversions, Target ROAS, Maximize Conversion Value)
• Whether the bid strategy matches the account's stage (a new account on Target ROAS without enough conversion data is being asked to optimize without the data to do it)

For most paid ads accounts under $50K/month, manual CPC with deliberate bid adjustments outperforms automated bid strategies. The reason is simple: automated strategies need data, and most SMB accounts do not generate enough conversions per week for the algorithm to learn properly. Above $50K/month, Target ROAS or Maximize Conversion Value can outperform manual, but only if stages 1 through 7 are clean.

If they're not, the bid strategy is optimizing on bad data. Same problem as stage 1, just downstream.

For the full breakdown of when each smart bidding strategy works vs when manual CPC outperforms, see the Smart Bidding Audit.

For the full breakdown of what each component score actually measures and the order of operations to lift Quality Score sustainably, see what Quality Score actually measures.

If you want a quick gut-check on whether your AOV and margin can absorb your category's CPC, the break-even ROAS calculator works the math in 30 seconds.

What good and bad look like, by stage

A self-check table you can use on your own account:

If you walk through this table honestly and find yourself in the bad column on more than three rows, the account does not need a bid strategy change. It needs a structural rebuild.

A note on the other half: most brands also run Meta

The 8-stage method above is built for Google. Most paid ads brands also run Meta. The audit logic is similar but the diagnostic order changes, because Meta is interruption-based and Google is intent-based. The leaks live in different places.

I documented the full 10-stage Meta audit method for that side of the picture. If you run both platforms, audit them both. Treating Meta and Google as one bucket is one of the most common reasons brands cannot figure out where their actual returns are coming from.

The 4 account stages framework (Launch, Optimization, Scaling, Cost-Cut) applies equally to Google accounts. The specific tactics shift between platforms, the stages do not. See the 4 stages of a Meta account for the full breakdown.

For brands selling on Amazon as well, see the Amazon ads audit method. The audit order flips on a marketplace: listing health and reviews come before bids, because the ad rides on top of the listing instead of standing on its own.

For the comprehensive self-audit checklist with dollar impact per item, see the complete Google Ads audit checklist (2026). It is the line-item version of the 8 stages above, with a yes-or-no check and a price tag on each one.

For the 5 highest-leverage levers consolidated into a single optimization sequence, see how to maximize Google Ads (5 levers that move ROAS).