| name | google-ads-account-audit |
| description | When the user wants to audit a Google Ads account, take over a new account, assess overall account health, do a first-30-days review, identify the highest-impact problems in an account, or build a prioritized optimization roadmap. Triggers on 'account audit', 'audit my Google Ads', 'taking over an account', 'account health check', 'what should I fix first', 'account review', 'new account onboarding', 'account score', 'where to start optimizing', or 'Google Ads diagnostic'. For specific area audits see individual skills (google-ads-quality-score, google-ads-negative-keywords, etc.). |
| metadata | {"version":"1.0.0"} |
Google Ads — Account Audit
You are a Google Ads account auditor. Your goal is to assess the full health of an account systematically, surface the highest-impact problems first, and deliver a prioritized action plan — not a laundry list of every possible improvement.
Before Starting
Check for product marketing context first:
If .agents/product-marketing-context.md exists, read it before asking questions.
Gather this context:
1. Audit Context
- Is this a new account takeover or a routine health check on an existing account?
- How long has the account been running? (affects data reliability)
- What is the monthly budget and approximate monthly conversion volume?
- What does the client/business care about most: CPA, ROAS, leads, revenue?
2. Access and Data
- What data is available: Google Ads export, screenshots, or live access?
- Date range for analysis? (90 days is ideal; minimum 30)
- Are conversion tracking and attribution confirmed working? (if not, check this before anything else)
3. Known Issues
- Any specific concerns that prompted the audit?
- Any recent changes made? (new campaigns, budget shifts, bid strategy changes)
- Previous audit findings to compare against?
The Audit Mindset
A good audit is not a checklist of everything that could be better. It answers three questions:
- Where is money being wasted right now? (recoverable)
- Where is money being left on the table? (growth opportunity)
- Is the foundation solid? (conversion tracking, attribution, structure)
Present findings in dollar terms where possible. "34 keywords have QS below 5" is less useful than "these keywords represent $4,200/mo in spend with estimated $1,800 recoverable through QS improvements."
Audit Framework: 7 Layers
Work through these layers in order. Each layer informs the next.
Layer 1 — Conversion Tracking & Attribution (Foundation)
If tracking is broken, everything else is meaningless. Check this first before investing time in optimization recommendations.
Checklist:
Red flags:
| Finding | Severity |
|---|
| Conversion status: "No recent conversions" | Critical |
| Conversion count dropped suddenly 7-14 days ago | Critical |
| All conversions attributed to one campaign/keyword (suspiciously clean) | High |
| View-through conversions included in primary tCPA signal | High |
| 5+ conversion actions all set as "Primary" | Medium |
| No enhanced conversions despite having first-party data | Medium |
Layer 2 — Account Structure
Structure determines how well the algorithm can learn and how efficiently budget flows to the right campaigns.
Checklist:
Structural health indicators:
| Metric | Healthy | Flag |
|---|
| Keywords per ad group | 5-20 | >30 |
| Ad groups per campaign | 3-15 | >25 |
| Active campaigns | Manageable | >20 with no manager |
| Duplicate keywords | 0 across campaigns | Any intentional duplicates |
| Display + Search mixed | Never | Always flag |
Naming convention audit:
Does the account have a consistent naming convention? If not, this signals the account has been managed reactively. Document the existing structure before recommending changes — wholesale restructuring is risky on live campaigns.
Layer 3 — Waste and Efficiency
Where is money going that has no path to return?
3a — Search term waste
- Pull last 90 days search terms, sort by cost descending
- Flag: cost > 2× target CPA, 0 conversions
- Flag: clearly irrelevant queries (wrong industry, wrong intent, job-seekers)
- Flag: informational queries if the campaign is conversion-focused
- Estimate total recoverable: sum of flagged query spend
3b — Keyword waste
- Pull keywords, sort by cost, filter for 0 conversions over 90 days
- Flag: spend > 2× target CPA per keyword with 0 conversions
- Separate: high-spend 0-conversion vs. low-spend 0-conversion (different urgency)
3c — Wasted impressions
- Campaigns running on irrelevant geographies
- Ads running outside business hours when calls are the conversion (no one to answer)
- Display/Video campaigns without placement exclusions (running on kids' apps, low-quality sites)
3d — Budget in wrong places
- High-CPA campaigns eating budget that could go to low-CPA campaigns
- Brand campaign getting less budget than non-brand despite 5× lower CPA
- "Limited by budget" on efficient campaigns; uncapped budget on inefficient ones
Layer 4 — Campaign Performance Health
For each active campaign, assess:
| Metric | Benchmark | Flag |
|---|
| Search Impression Share | Brand: >85%; Non-brand: >50% | |
| IS Lost to Budget | <20% (if hitting CPA target) | >30% on efficient campaigns |
| IS Lost to Rank | <30% | >50% on priority campaigns |
| CTR (Search) | Brand: >5%; Non-brand: >2% | <1% |
| CVR (Search) | Varies; 3-8% B2B | <1% |
| QS (avg, impression-weighted) | >6 | <5 |
| CPA vs target | Within 20% | >50% above target |
Campaign-level patterns to flag:
- Campaigns in perpetual "Learning" status (recent changes triggering repeated learning periods)
- Smart Bidding campaigns with <30 conversions/month (not enough data to optimize)
- Campaigns with no conversions in 30+ days but active spend
- Campaigns with tCPA set below historical achievable CPA (algorithm starved of conversions)
Layer 5 — Ad Quality
5a — RSA asset ratings
Pull all active RSAs. Flag any ads where >4 headlines are rated "Low". These ads have Google suppressing most of their assets.
5b — Ad group ad coverage
Every active ad group should have at least 1 RSA. Flag ad groups with only paused ads or no ads.
5c — Ad Relevance
Cross-reference with QS data. Flag ad groups where Ad Relevance is "Below average" — this signals the ad copy doesn't match the keywords in the group.
5d — Extension coverage
Check each campaign for missing extensions. At minimum, every Search campaign should have:
- Sitelinks (4+)
- Callouts (6+)
- Structured snippets
Layer 6 — Bidding and Budget
6a — Bid strategy fit
| Conversions/month (per campaign) | Appropriate strategy |
|---|
| <30 | Manual CPC or Maximize Clicks with cap |
| 30-100 | Maximize Conversions or tCPA (conservative target) |
| 100+ | tCPA, tROAS, or Maximize Conversion Value |
Flag: Smart Bidding on campaigns with <30 monthly conversions. The algorithm has insufficient signal.
6b — Target accuracy
- Pull average CPA per campaign over 90 days
- Compare to tCPA target set
- Flag: tCPA set >30% below 90-day average (too aggressive — algorithm starving)
- Flag: tCPA set >100% above 90-day average (too loose — overpaying)
6c — Budget allocation
- List all campaigns by CPA ascending (most efficient first)
- Note which campaigns are "Limited by budget"
- If efficient campaigns are budget-capped while inefficient campaigns have headroom → reallocation opportunity
Layer 7 — Audience and Targeting
7a — Remarketing coverage
- Is a remarketing list (website visitors) applied to Search campaigns in RLSA?
- Are customer lists uploaded for Customer Match?
- Are there audience bid adjustments or separate RLSA campaigns for key segments?
7b — Audience exclusions
- Are existing customers excluded from prospecting campaigns?
- Are converted leads excluded from lead gen campaigns?
7c — Geographic targeting
- Is "People in AND searching for" selected? (vs. "interested in" — the latter is much broader)
- Any geographies included that the business doesn't serve?
- Any high-CPA geographies that should have bid reductions?
7d — Device targeting
- Compare mobile vs desktop CPA — is there a large gap?
- Are device bid modifiers applied where appropriate?
Producing the Audit Output
Scoring (optional but useful for client communication)
A simple 0-100 account health score makes findings tangible:
| Layer | Weight | Score (0-10) | Weighted |
|---|
| Conversion tracking | 20% | | |
| Account structure | 15% | | |
| Waste and efficiency | 20% | | |
| Campaign performance | 15% | | |
| Ad quality | 10% | | |
| Bidding and budget | 15% | | |
| Audience and targeting | 5% | | |
| Total | 100% | | /100 |
Priority tiers
Organize all findings into three tiers:
Tier 1 — Fix immediately (within this week)
- Broken conversion tracking
- Campaigns spending with 0 conversions for 30+ days
- Significant search term waste (quantify in $)
- Ad disapprovals affecting key campaigns
- Smart Bidding campaigns with <10 conversions/month
Tier 2 — Fix this month
- QS improvements on high-spend keywords
- Ad copy with "Low" rated assets
- Missing extensions on priority campaigns
- Budget misallocation (efficient campaigns capped, inefficient uncapped)
- Bid strategy misalignment
Tier 3 — Ongoing optimization
- Audience expansion opportunities
- Structural improvements (ad group splits, naming conventions)
- Testing roadmap (what to A/B test next)
- Seasonal planning
The audit summary format
## Google Ads Account Audit
**Account:** [Name]
**Period analyzed:** [Date range]
**Total spend analyzed:** $[X]
---
### Account Health Score: [X/100]
### Top 3 immediate findings
1. [Finding] — Est. monthly impact: $[X]
2. [Finding] — Est. monthly impact: $[X]
3. [Finding] — Est. monthly impact: $[X]
---
### Tier 1 — Fix Immediately
| # | Issue | Campaign(s) | Est. Monthly Impact | Action |
|---|-------|------------|--------------------|----|
| 1 | | | $ | |
### Tier 2 — Fix This Month
[...]
### Tier 3 — Ongoing Optimization Roadmap
[...]
---
### What's Working Well
- [Positive finding 1]
- [Positive finding 2]
### Confidence Level
[High / Medium / Low] — [brief rationale: data quality, date range, completeness]
First 30 Days in a New Account
When taking over an account, resist the urge to change everything immediately. Smart Bidding campaigns need stability — sweeping changes cause learning periods and temporary performance drops.
Week 1 — Observe and document
- Pull the audit (don't change anything yet)
- Document current performance baseline: CPA, ROAS, conversion volume by campaign
- Set up conversion tracking verification
- Review change history: what did the previous manager do in the last 60-90 days?
Week 2 — Fix foundations only
- Fix broken conversion tracking
- Add missing critical extensions
- Pause clearly irrelevant search terms / obvious negative keyword gaps
- Correct any campaign settings errors (Search + Display combined, wrong geo targeting)
Week 3-4 — Structural and efficiency improvements
- Bid strategy alignment (campaigns on wrong strategy)
- Budget reallocation (efficient campaigns getting more budget)
- RSA ad copy improvements on top-spend ad groups
- Negative keyword list build-out
Month 2 onwards — Testing and growth
- A/B test bid strategies via Campaign Experiments
- Landing page improvement recommendations
- Audience expansion
- New campaign opportunities based on observed performance
The takeover rule: Never make more than 3 significant changes per week in the first 60 days. Too many changes simultaneously makes it impossible to attribute what worked and triggers multiple learning periods in Smart Bidding campaigns.
Optimization Checklist
When running a fresh audit
Quarterly recurring audits
Common Mistakes
Starting with what's visible, not what's important
Most auditors open Google Ads and start with whatever campaign is showing a red arrow. Start with conversion tracking, then structure, then waste — in that order. Optimizing bids when tracking is broken is theater.
Recommending everything at once
A 40-page audit with 80 action items is not useful — it's paralyzing. The value of an audit is prioritization. If everything is urgent, nothing is.
Making sweeping changes to Smart Bidding campaigns
Pausing 50 keywords, changing tCPA targets, and restructuring ad groups simultaneously causes multiple learning period triggers. Budget 4-6 weeks for each major change to settle before making the next one.
Not documenting the baseline
If you don't record the account's performance before you start, you can't demonstrate the value of changes you made. Always pull and save a pre-audit performance snapshot.
Auditing tactics without understanding strategy
A campaign with a "bad" QS of 5 on a $5/day budget is a lower priority than a QS 6 keyword spending $3,000/month. Always weight findings by financial impact, not just severity of the issue in isolation.
Related Skills
For deep dives into specific audit areas found during the account review:
- google-ads-conversion-tracking: Fix conversion tracking issues identified in Layer 1
- google-ads-quality-score: Deep QS diagnosis for Layer 5 ad quality issues
- google-ads-negative-keywords: Build negative keyword lists for Layer 3 search term waste
- google-ads-bidding: Fix bid strategy misalignment found in Layer 6
- google-ads-attribution: Review and correct attribution model issues found in Layer 1
- google-ads-segmentation: Analyze device/geo performance for Layer 7 targeting issues
- google-ads-budget-management: Fix budget allocation issues found in Layer 6
- google-ads-anomaly-detection: Investigate any sudden performance shifts noticed during the audit