The Most Egregious Findings From 500 Google Ads Audits

Most advertisers assume their Google Ads account is "fine" because the reported numbers look acceptable. Smart Bidding is on, conversions are coming through, the agency sends a monthly report. In reality, when PPC.co auditors open the back of these accounts, the same handful of structural failures turn up again and again, often costing more than the agency fee ever did.

What follows is a field report from several hundred Google Ads account audits conducted over the last few years. Not opinions. Patterns. Each finding below is one the audit team sees across industries and spend tiers, with the diagnostic step that surfaces it and the dollar impact when it goes unaddressed.

So where is the money actually leaking?

Conversion Tracking Is Lying To The Bidder

The first place any serious PPC audit starts is the conversions table, because every downstream decision rests on it. Disruptive Advertising's review of 2,000 accounts found that only 57.7% had any conversion tracking in place, and of those, only about half were actually tracking anything meaningful. That means roughly seven in ten accounts are bidding against a signal that is either missing or wrong.

The specific errors are predictable. A native Google Ads tag and a GA4 import fire for the same event. A thank-you page triggers both a page-view conversion and a form-submit conversion. One auditor documented a case where a client's reported 90 monthly leads were in fact 32 real inquiries after duplicate and triple firing was stripped out. Smart Bidding does not know this. It learns from the inflated signal and overbids on whatever correlates with the duplicates.

The diagnostic is boring and effective. Segment the All Campaigns view by conversion action, compare 30 days of reported conversions against the CRM or Shopify back end, and flag any ratio above 1.2:1. Then open each conversion action and check the counting method, the primary/secondary flag, and whether the same business outcome is being imported from two sources.

Performance Max Is Buying Brand Searches You Already Owned

Performance Max is the single largest source of recoverable spend the audit team identifies in 2026. The pattern: PMax reports a 7x or 10x ROAS, the client is thrilled, and nobody notices that branded queries are being absorbed into the black box at a premium.

Optmyzr's February 2025 review of 503 managed accounts found Search keywords overlapping with Performance Max search terms in 91.45% of them. That is not a rare edge case. It is the default state of accounts running PMax alongside Search. The reported ROAS looks spectacular because PMax is cashing checks the brand campaign would have cashed anyway, usually at a lower CPC.

The audit workflow is straightforward:

• Pull the PMax channel performance report and look at how much budget is going to Search inventory versus Shopping, YouTube, and Display.
• Open the PMax search terms report and bucket branded versus non-branded queries.
• Compare the brand Search campaign's CPA against blended PMax CPA. If brand Search converts at a fraction of the cost, PMax is almost certainly re-capturing that traffic.
• Check whether an account-level brand exclusion list exists at all. Negative keywords alone do not solve this.

When the fix is implemented correctly, PMax-attributed conversions typically drop 10 to 20 percent, brand Search impression share recovers, and blended Google ROAS holds steady or improves because the account stops paying itself. That recovered budget then funds non-brand prospecting, which is what PMax was supposed to be doing in the first place.

Broad Match Running Without A Negative Keyword List

Match type discipline has eroded quickly. An Optmyzr analysis of 30,000 accounts found that exact match has lost roughly 9.5 percentage points of spend share since 2022, with broad match now the dominant match type by budget. That shift is defensible when it is paired with a mature negatives list and a working feedback loop. It is not defensible when the account has three negatives total and the search terms report shows the budget being spent on queries unrelated to the product.

WordStream's analysis of 15,000 accounts quantified the gap: accounts with at least one negative keyword had conversion rates nearly triple those without, and the average account wasted $1,127 per month. Wasted spend concentrates. A separate review of 32 accounts and 5.2 million zero-conversion search terms found that the most expensive 10% of those terms hold a median 48.8% of the waste. Half the leak sits in a tenth of the queries. Catching it is a one-afternoon exercise.

The audit team's shortcut: sort the search terms report by cost descending, filter to zero conversions over 90 days, and work down from the top. Most accounts need twenty negatives to recapture the majority of the waste. The keyword match types you choose matter less than whether the account is actually learning from what it bought last month.

The uncomfortable complication: an analysis of 13 B2B accounts found that 61% of Search spend went to search terms Google never disclosed in the search terms report, and no account had even 75% of its terms reported. The negatives you can build are the negatives you can see. The rest requires n-gram analysis of what is reported, inference from landing-page behavior, and tighter match-type discipline on themes that keep leaking.

Performance Max Is Overspending On Low-Margin SKUs

For ecommerce accounts, the shopping feed is the second-largest audit finding after conversion tracking. A JudeLuxe audit of 50 ecommerce accounts spending between £10,000 and £350,000 per month found that 43% of Performance Max spend went to SKUs in the bottom two margin quintiles, with a median £4,200 per month wasted per account.

PMax optimizes for revenue, not margin. If the feed includes high-revenue, low-margin SKUs with no listing group segmentation or custom labels indicating profitability, the bidder will chase the gross-dollar sales and the P&L will quietly suffer. The audit fix is to add custom labels for margin tiers, split listing groups by tier, and either exclude the bottom quintile or give it a hard ROAS target that reflects contribution margin rather than top-line revenue.

Search Partners And Display Expansion Are Dragging ROAS Down

Two checkboxes buried in campaign settings quietly cost accounts a measurable share of their performance. Industry research cited by PPC Land puts Search Partner Network return on ad spend 37% below Google Search proper. Display expansion on Search campaigns tends to look even worse once filtered to real conversions rather than view-through credit.

The diagnostic is a thirty-second toggle check across every campaign in the account. Segment by network, compare Search Partners and Display versus Search proper on CPA and ROAS, and turn off what cannot defend its own numbers. For most accounts, that single change recaptures three to seven percent of total spend with no other work.

View-Through Conversions Are Inflating Display And YouTube Credit

View-through attribution is the most quietly destructive setting in the account for brands that run Display, YouTube, or Discovery alongside Search. By default, an impression served to someone who later converts through any channel can earn credit, and that credit flows into Smart Bidding as a signal.

Audit steps: in each video and display campaign, check whether view-through conversions are counted as primary or observation-only. Pull the "conversions by conversion action" view and compare click-through to view-through volume. When view-through runs more than roughly 15 to 20 percent of click-through on a brand-safe display campaign, the attribution window is likely too long, the counting is too generous, or both. Move view-through to secondary so the bidder stops chasing it, and the picture of what actually drives new customer acquisition sharpens immediately.

Ranking The Repeat Offenders By Recoverable Spend

Not every finding carries the same financial weight. Across the audits the team runs, the ranking of recoverable spend is remarkably stable, which is why the audit process starts with tracking integrity and campaign structure before anyone touches ad copy or bids. The downstream work does not matter if the signal is wrong and the money is going to the wrong queries.

If you want a benchmark for how the fundamentals compare across channels before the audit even starts, the Google Ads vs Meta Ads ROI benchmarks are a reasonable baseline, and the Smart Bidding vs manual CPC comparison frames when automation is working for you versus against you.

What A Rigorous Audit Actually Produces

The deliverable from a credible Google Ads account audit is not a slide deck of generic advice. It is a prioritized list of findings, each tied to a dollar figure and a diagnostic trail, so the client can see exactly what the problem is, how it was identified, and what the recovered spend buys. The sequencing matters: fix the conversion tracking first, because every optimization downstream depends on a true signal, then address the structural waste in Performance Max and broad match, and only then touch bids, creative, and audiences.

Agencies lose accounts because they report on the metrics the platform hands them rather than the metrics the business cares about. An audit is the correction. If the numbers in your monthly report feel too clean to match the pipeline you actually see, the account is probably hiding one or more of the patterns above. The team at PPC.co runs this diagnostic on request, and the recoverable spend tends to pay for the engagement several times over before any new strategy is layered on top.

1. 11. Search terms reportFilter last 30 days, sort by cost descending, scan top 50 for irrelevant terms and add as negatives
2. 22. Network segmentationSegment campaign view by network and compare Search Partners CPA against Google Search
3. 33. Hourly performancePull the Hour of day report and flag hours with high spend and zero or near-zero conversions
4. 44. Campaign overlapCheck auction insights on Search campaigns and compare against Performance Max for cannibalization