TL;DR
A real Google Ads audit checklist takes 4 to 8 hours, not 30 minutes. It audits strategy and account architecture, not just settings and checkboxes. Most accounts we audit have technically clean settings but strategic problems: misaligned campaign architecture, conversion tracking that’s measuring the wrong things, bid strategies running below the data threshold needed to work, and search-term waste that has shifted shape since AI assistants changed query patterns.
This article walks through the 10-section framework we use across audited accounts. It covers what to check, why it matters, what failure looks like at each stage, and where to focus time when you only have 2 hours instead of 8.
Why most Google Ads audit checklists audit the wrong things
Most audit checklists are settings audits. They walk through the Google Ads UI, check whether boxes are ticked correctly, confirm that conversion tracking exists, glance at Quality Score, and produce a report listing what’s “set up properly” versus what isn’t.
This kind of audit misses the real problems. An account can have every checkbox ticked correctly and still be strategically bleeding. Settings audits answer “is the account configured according to best practice.” Strategy audits answer “is the account architected correctly for what this business needs.” Those are different questions, and Google Ads accounts fail far more often at the second one.
Here’s a concrete pattern we see repeatedly in our work as Google Ads specialists for small businesses. An e-commerce account has clean conversion tracking, well-organised campaigns, decent Quality Score distribution, and Smart Bidding running. By any settings-audit standard it looks well-managed. But the campaigns are structured around product categories rather than margin tiers, so the bid algorithm is treating the account’s $400-margin SKUs the same as its $40-margin SKUs. Smart Bidding can’t optimise for profit it can’t see. The account is “well-optimised” by tactical standards and structurally bleeding by strategic standards.
The trade-off worth naming upfront: thorough settings checklists are easier to apply but produce surface-level results. Strategy audits require more judgment and more account context. Both have a place. The audit framework below tries to do both, but if you only have time for one, do the strategy audit. Settings problems are usually visible in the UI and get caught eventually. Strategic problems compound for years without anyone noticing.
What’s changed about Google Ads audits since 2023
If you’re applying a 2020-era audit checklist to a 2026 account, you’re missing the most common failure modes. Four shifts matter for how audits should work in 2026.
Smart Bidding has obscured keyword-level visibility. Three years ago, an audit could review keyword-level performance, identify which keywords were driving conversions versus draining budget, and adjust bids accordingly. Today, with most accounts running tCPA, tROAS, or Maximise Conversions, the bid algorithm overrides keyword-level signals based on user-level prediction data the auditor can’t see. The audit needs to shift up a level. Instead of asking “is this keyword bid correctly,” ask “does this campaign have enough conversion volume for the algorithm to learn?” Different question, different answer.
PMax and AI Max obscure search-term data. Performance Max campaigns surface a fraction of the search terms they spend on, and AI Max (rolling out through 2026) extends similar opacity to Search campaigns. A 2020 audit would pull a 90-day search-terms report and find every wasteful query. A 2026 audit pulls the same report and gets a sample. The audit method has to change to compensate: more emphasis on negative keyword strategy upfront (because reactive cleanup catches less waste), more attention to bid signals and conversion modelling, less reliance on the search-terms report as the primary data source.
GA4 attribution differs materially from Universal Analytics. Most accounts migrated to GA4 between 2022 and 2023, but many audit frameworks still assume Universal Analytics behaviour. GA4’s data-driven attribution (the default), combined with its event-based model and consent-mode-aware reporting, produces conversion counts and source attribution that don’t match what Google Ads reports natively. An audit that compares GA4 conversions to Google Ads conversions and flags the discrepancy as a “tracking issue” is often misreading the situation. The discrepancy is usually correct behaviour, not a bug.
Post-iOS 14 and post-CAPI reality changed what’s measurable. For accounts running both Google and Meta, the 2024-2025 reality is that some conversions can be tracked deterministically and some have to be modelled. An audit that treats all conversions as equally trustworthy will misallocate effort. The 2026 audit needs to ask: which conversions are server-side tracked, which are client-side, which are modelled, and how much weight does each carry in the bid signal?
The honest framing: a thorough 2026 audit has to do more inferential work than a 2020 audit. Less direct data, more strategic judgment. That’s why audit time has gone up, not down, despite Google Ads UI improvements.
The 10-section senior operator audit framework
Each section below covers what to check, why it matters at the strategic level, what failure looks like, and what to do if it fails. Apply in order. The earlier sections are foundational. Later sections depend on the earlier ones being correct.
If you only have an hour: the five places to look first
A first pass on an inherited account is reconnaissance, not optimisation. It should produce a prioritised list of problems with rough impact and effort, not changes to ship. Check these five places in order.
- Spend against conversions at search-term level: sort the search terms report by cost for the last 30 days, then check the top 10 terms by spend with no conversions and high-impression, low-CTR terms that signal targeting too broad for the ad (see the wasted spend benchmarks).
- Conversion tracking: count the conversion actions (five or more primary actions feeding Smart Bidding means a confused goal), check purchases count “Every” (Google’s recommended setting for sales) and pass a transaction ID so a reloaded confirmation page can’t count the same order twice, and confirm the Enhanced Conversions match rate is above 50%, with fixes in the conversion tracking setup guide.
- Bid strategy against conversion data: test whether each target is realistic for the campaign’s history (an 800% Target ROAS on a campaign that has never passed 350% throttles delivery) and whether campaigns hitting budget daily are on the right strategy, using the bid strategy decision framework.
- PMax hygiene: check asset group text ratings, brand exclusions (without them PMax can claim branded conversions and inflate ROAS) and audience signals matched to the campaign goal, with structure covered in the PMax ecommerce structure guide.
- Quality Score outliers: filter to Quality Score 4 or below, treating brand keywords under 7 as a structural fault and scores of 4 to 5 across many ad groups as an architecture problem, then use the Quality Score diagnostic guide.
Run this scan before any client conversation on an inherited account, or on your own account if you are weighing an agency change.
1. Account architecture
The first thing to audit is whether the account is structured to match the business’s actual conversion economics. Account architecture means campaign segmentation logic: are campaigns split by intent (brand vs non-brand), by margin tier (high-margin SKUs vs low-margin), by conversion type (lead-gen vs e-commerce purchase), by product category, or by some combination?
Most accounts inherit a structure that worked at an earlier scale and never get re-architected as the business grows. The fail mode looks like this: a campaign called “All Products” running across the full catalogue, with bid strategy targeting one ROAS for everything regardless of margin, customer LTV, or stock priority. Checking that the structure still fits the business is part of how Adgenix runs ecommerce accounts.
What to check: does the campaign structure separate things that have different optimisation goals? Brand keywords (where you want maximum impression share at low CPC) need to be in their own campaign, separated from non-brand. High-margin product lines should be in different campaigns from low-margin lines, so bid targets can differ. Lead types with different LTV (qualified leads versus newsletter signups) should not share a conversion column.
If architecture is wrong, the most aggressive optimisation work downstream cannot fix it. Smart Bidding will optimise toward whatever conversion signal you give it. If the signal mixes high-value and low-value conversions, the algorithm averages them and underbids on the valuable ones.
2. Conversion tracking integrity
The second priority. If conversion data is wrong, every downstream optimisation is steering off a faulty compass.
The audit checks: are conversions firing where they should, only where they should, with the correct values, attributed to the correct sources? Common failures include double-counting (where both GTM and a plugin fire the same conversion), missing conversions on key thank-you pages, conversion values set to a static figure instead of dynamic order value, and Enhanced Conversions configured but not actually receiving hashed user data.
A specific failure pattern worth naming: GA4 reports significantly different conversion totals from Google Ads native conversion tracking. Some discrepancy is expected (different attribution models, different reporting windows, different consent handling). But a 30%+ gap usually indicates one of three issues: a tracking implementation problem, a Consent Mode configuration that’s silently dropping events, or Enhanced Conversions sending data Google can’t match.
The audit needs cross-reference data to test this. Without access to the client’s CRM or back-end orders data, you can’t verify whether either reporting system is correct. Audits without that access can identify the discrepancy but not which system is right. Note this as a flag, not a finding.
3. Bid strategy fit-for-purpose
Most accounts above $5k per month run Smart Bidding (tCPA, tROAS, or Maximise Conversions). Most accounts below $1k per month should not. The middle range is where bid strategy audits have the most impact.
The audit checks: does this campaign have enough conversion volume for Smart Bidding to actually learn? A working rule, not a Google requirement: about 30 conversions in 30 days before moving to Target CPA, and about 50 before Target ROAS. Below those thresholds the algorithm runs in a degraded state, often producing worse performance than thoughtful manual bidding would.
The fail mode: an account running tCPA on a campaign generating 8 conversions per month, with the operator wondering why performance is volatile. The algorithm doesn’t have enough data to optimise. It’s making best-guess decisions on sparse signal.
What to do if it fails: either consolidate campaigns to pool conversion volume (combine three small campaigns into one, splitting the budget intelligently), shift to a more permissive bid strategy (Maximise Clicks while you build conversion volume), or revert to manual CPC with conservative bids while the account scales toward Smart Bidding readiness.
The trade-off here is real. Smart Bidding outperforms manual on accounts above the data threshold. Manual outperforms Smart Bidding below it. Switching too early or too late both cost performance.
4. Search terms and negative keyword hygiene
This section gets less time than it used to, because PMax and AI Max obscure search-term data, and Smart Bidding’s user-level signals override keyword-level decisions in many accounts. But the principle still matters: is the account spending money on queries that have no commercial value for this business?
What to check: pull the 90-day search-terms report (knowing it’s a sample). Filter by spend descending. Look at the top 100 queries by spend and classify each into one of five intent gates: commercial mismatch, informational without buying intent, wrong product or wrong-fit attribute, wrong geography, or brand confusion.
If more than 30% of search-term spend falls into one of these gates, the account has reactive negative keyword work to do. If more than 60% does, the account has a systematic negative keyword strategy problem and needs to shift from reactive cleanup to proactive intent-gating before more budget drains.
The 2026 angle worth flagging: the share of waste in Gate 2 (informational without buying intent) has grown substantially because AI assistants have absorbed pure-informational queries, leaving research-stage queries on Google that look commercial but rarely convert. This is covered in detail in our intent-gate negative keyword framework.
5. Quality Score distribution
Account-level Quality Score averages hide more than they reveal. A 7/10 average can sit on top of a distribution where 30% of keywords are at 9-10, 40% are at 7-8, and 30% are at 4-5. The 4-5 tail is where the budget bleed happens, because low-Quality-Score keywords pay 2-4x the CPC of high-Quality-Score keywords for the same position.
What to audit: pull keyword-level Quality Score data, sorted ascending. Look at the bottom 20% of keywords by QS. Check three things: are these keywords getting impressions despite low QS (Google is showing them, and they’re paying premium CPC), are the ad relevance and landing page experience scores both poor (suggesting structural issues), and is the expected CTR component low (suggesting the keyword shouldn’t be in this account at all).
The fail mode: an account with a respectable account-level QS that has 50+ low-QS keywords driving 20-30% of total spend. Fixing those keywords (either improving the landing page match, rewriting ads, or removing the keywords entirely) typically reduces CPC across the affected keywords by 30-50%.
What’s harder in 2026: Smart Bidding has reduced the ranking influence of Quality Score for many accounts, which makes QS a less direct lever than it used to be. But the underlying signals (ad relevance, landing page experience) still matter for cost efficiency, even if the QS number itself matters less.
6. Ad copy and creative testing
The audit checks the structure of responsive search ad assets, not just whether ads exist. The framework: 15 headlines per RSA, structured as 6 keyword-focused headlines, 5 CTR-driver headlines that demonstrate value or urgency, and 4 call-to-action variants of which at least one is short (15 characters or less for mobile display).
The fail mode: an RSA with 15 headlines that all sound the same. The algorithm has nothing meaningful to test. Ad strength shows “Excellent” because the system is rewarding asset count, not asset diversity.
What to check beyond headline structure: are headlines being pinned in ways that constrain the algorithm without strategic reason? Are ad extensions populated and current? Are sitelinks pointing to live pages with relevant content? Has the account been A/B testing creative concepts (different angles, different value propositions) or just iterating on word choice within the same concept?
The trade-off: tighter pinning gives more control but limits testing. Looser pinning gives the algorithm more room but risks ad combinations that contradict the campaign strategy. Most accounts over-pin, then wonder why their RSAs don’t outperform their old expanded text ads did.
7. Audience and geo-targeting
The audit checks whether audience signals and geographic targeting align with where the business actually converts.
For audience signals: are first-party audiences (customer match, website visitors, converters) uploaded and active? Are they applied to campaigns where they would help (typically as observation, not targeting, in Search, and as targeting in PMax and Display)? Are similar-audience signals being used to expand reach to people who look like converters?
For geo-targeting: open the location report. Look at performance by region. The fail mode: a campaign nominally targeting “Australia” that’s actually getting 60% of impressions in one metro area, with conversion rates that vary 3-4x between the highest-converting and lowest-converting regions. Bid modifiers should reflect this. Most accounts don’t apply geo bid modifiers at all, leaving uniform bidding across regions with very different conversion economics.
What’s worth checking specifically: negative locations. Many accounts target Australia broadly but never exclude offshore traffic that slips through (queries in Australian English from overseas users, or campaigns running on placements outside the geo target). Negative locations clean this up.
8. Attribution and reporting
The audit examines what attribution model the account is using, what lookback window is configured, and whether those settings match the actual sales cycle of the business.
For most e-commerce DTC brands, the conversion happens in days, not weeks, and the attribution model should reflect that. For longer-cycle businesses (B2B lead-gen, high-consideration purchases), the attribution model needs to capture the longer journey, which means data-driven attribution becomes more important and lookback windows need to be longer.
The fail mode: a B2B SaaS account with a 60-day average sales cycle running last-click attribution with a 30-day lookback. Half the conversions are being attributed to the last touchpoint, missing the multi-touch reality that produced the sale.
Worth checking specifically for accounts running both Google and Meta: how is cross-channel deduplication being handled? If both platforms are claiming credit for the same conversion in their respective dashboards, the operator has to decide which platform “owns” the conversion for budget allocation purposes. Most accounts fudge this. Senior operator audits name it explicitly.
9. Strategic vs tactical waste
This is the section that distinguishes operator-grade audits from junior audits. Tactical waste is visible in the data: bad search terms, poor ad copy, wrong bid strategy, broken tracking. Strategic waste is invisible in the data: campaigns optimised toward the wrong outcome, account architecture that mixes incompatible goals, conversion signals that don’t reflect what the business actually values.
A specific example: an e-commerce account using “all purchases” as the conversion signal, with Smart Bidding optimising toward maximum purchase volume. The operator notices ROAS has slowly declined over six months. The tactical audit finds no obvious issues. The strategic audit notices that the account has been gradually skewing toward the lowest-priced products, because those have the highest conversion rate, and the algorithm has been doing exactly what it was asked: maximise conversion count. Revenue is up, profit is down, and the account looks “well-optimised.”
What to look for: misalignment between the conversion signal and the business outcome the operator actually wants. Goals stated as ROAS when the business actually optimises for profit. Lead volume as the goal when lead quality is what the sales team measures. Campaigns running for two years against a goal that no one has revisited.
The counter-conventional point worth naming: a tactically well-optimised account can still be strategically bleeding. The fix is rarely a tactical change. It’s usually a re-think of the conversion strategy, the account architecture, or what the campaign is actually for.
10. Account hygiene and operational readiness
The last section covers operational signals that don’t directly affect performance but indicate whether the account is being actively managed.
What to check: are notes and labels being used to track changes, tests, and rationale? Is change history showing regular activity (someone is in the account weekly or fortnightly), or has the account been on autopilot for months? Are user permissions appropriate (does the agency still have admin access after the engagement ended)? Are budget pacing alerts configured? Is there a documented process for who responds when conversion volume drops or CPC spikes?
These checks won’t show up in performance data, but they predict future performance. An account with no notes, no change history for 90 days, and four people with admin access who don’t communicate is going to have problems eventually. An account with weekly notes, deliberate change history, and clean access controls is being managed by an operator who treats the account as a system, not a destination.
How long a real audit takes
The honest framing most checklists won’t give you: a thorough Google Ads audit takes 4 to 8 hours of focused work, not 30 minutes.
Time breakdown for a typical audit on a $10-25k per month account:
- Section 1 (architecture): 30-45 minutes
- Section 2 (tracking): 45-90 minutes if cross-referencing GA4 and CRM data
- Section 3 (bid strategy): 30 minutes
- Section 4 (search terms): 60-90 minutes for a real review of 90 days of data
- Section 5 (Quality Score): 30-45 minutes
- Section 6 (ad copy): 30-45 minutes
- Section 7 (audience and geo): 30 minutes
- Section 8 (attribution): 30 minutes
- Section 9 (strategic waste): 60-90 minutes: this is judgment work, not checklist work
- Section 10 (hygiene): 15 minutes
- Synthesis and prioritisation: 60-90 minutes
That’s 7 to 10 hours for a thorough audit. A 2-hour quick audit covers Sections 1, 2, 3, and 9 only, the high-impact strategic checks, and explicitly defers the rest. A 4-hour audit adds Sections 4, 5, and 6.
Account size affects this. A $5k per month account audits faster because there’s less data to review. A $50k+ per month account takes longer because the volume of data means manual review of every signal isn’t feasible. The auditor has to use scripts, dashboards, or sampling to cover the same ground.
The trade-off: 2-hour audits are appropriate when the goal is “find the obvious wins quickly.” 8-hour audits are appropriate when the goal is “understand whether this account is set up to scale.” Different goals, different time investments. Pretending a 30-minute checklist can do either is the most common audit mistake of all.
Common audit mistakes (operator-grade)
Six patterns we see repeatedly on audits that have technical depth but miss the strategic picture.
Mistake 1: Auditing without account access. Some audits are run from preview screenshots or PDF exports the prospect provides. These miss critical context: change history, conversion data quality, audience configurations, and the dozens of small settings that don’t show up in summary views. A real audit needs read-only access at minimum. Without it, the audit can identify symptoms but not root causes.
Mistake 2: Auditing without conversion data context. A Google Ads audit that doesn’t reconcile against the client’s CRM or back-end sales data is auditing one half of the conversion economy. The Google Ads dashboard says 200 conversions. The CRM says 140 of those were qualified leads and 60 were spam. Without the second number, the audit will recommend optimising toward the wrong signal.
Mistake 3: Auditing settings without strategic intent. This is the settings-vs-strategy distinction the article opened with. Checking that conversion tracking is configured correctly is necessary. Understanding whether the conversions being tracked are the right conversions for the business is the actual audit work.
Mistake 4: Auditing without fortnightly comparison data. Single-snapshot audits miss drift patterns. An account looks fine today but has been declining steadily for three months. The audit needs to compare current performance against historical baselines, not just review current state. Pull at least 90 days of data, broken by week, for every metric that matters.
Mistake 5: Auditing without first-party data context. The post-iOS 14, post-CAPI reality changed what’s auditable. An audit that treats all conversions as equally trustworthy will misallocate optimisation effort. The audit needs to ask: which conversions are server-side tracked, which are client-side, which are modelled, and how much weight does each carry in the bid signal?
Mistake 6: Auditing keyword-level data when Smart Bidding has obscured it. This is the 2026-specific mistake. Three years ago, keyword-level audits produced clear actions: this keyword underperforms, lower the bid or pause it. Today, with Smart Bidding overriding keyword-level signals based on user-level prediction data, the keyword-level review tells a less complete story. The audit needs to shift up a level to bid signals, conversion modelling, and audience signal quality. Auditors who haven’t updated their method since 2020 produce reports that miss most of what’s actually wrong.
The pattern across all six: assuming audits are about what’s visible in the Google Ads UI. A senior operator audit treats the UI as one input among several, and treats strategic intent as the lens through which everything else is evaluated.
When to do this yourself vs when to bring in an outside operator
If you’re running a single Google Ads account at $5k+ per month, with relatively stable performance, this framework is genuinely applicable as a self-managed quarterly audit. The 4-hour version (Sections 1, 2, 3, 4, 5, 6, and 9) is the right depth for that account size. Most owner-operators have the capacity to do this once per quarter. The compounding work isn’t in the framework, it’s in the discipline of returning to it on a regular cadence.
Where it gets harder:
8 or more campaigns running simultaneously. The cross-campaign analysis becomes a real workload. Patterns that are obvious in a single campaign get hidden by averaging across many.
Multi-platform attribution. When you’re also running Meta, LinkedIn, or other paid channels, the audit work has to extend across platforms. Single-platform audits miss the channel-mix questions that are often the most valuable.
Already deep in optimisation cycles. If your account has had aggressive optimisation work and performance has plateaued, the next gains usually require a fresh operator perspective rather than more of the same approach. Self-audits are good at finding obvious problems. Outside audits are good at finding blind spots.
Conversion economics aren’t clear. If you don’t have a confident answer to “what’s our break-even ROAS” or “what’s a qualified lead actually worth to us,” the audit work has to start with that conversation, not with the Google Ads UI.
Time pressure. If running the account is taking time away from running the business, the framework is technically applicable but practically not. The audit is 4-8 hours of focused work. If those hours are unavailable, the audit doesn’t happen.
The honest framing: most $5k to $25k per month accounts can do a competent audit themselves following this framework. Most $25k+ accounts benefit from outside execution because the time investment is harder to justify against the cost of a senior operator’s attention. And accounts at any spend level that have plateaued usually need a fresh look more than they need more methodology. If the audit shows you need help, see what a Google Ads specialist does and how to hire one.
If you’d rather skip the 8 hours of audit work, our Free Revenue Leak Report applies this exact framework as a 48-hour deliverable. We’ll cover all 10 sections, identify which intent gates the account is leaking through, and prioritise the findings by potential impact. No follow-up required. Keep the report either way.
For broader Google Ads work, see our Google Ads consultant service, where the same audit framework runs continuously across all managed accounts.
Adgenix · Senior Paid Media · Melbourne, Australia
