Google Ads Campaign Audit

Pre-requisite

Read ~/.claude/skills/google-ads/knowledge.md in full before doing anything. This audit framework is grounded in the research data within that file.

If the current date is 6+ months after the knowledge base research date, re-fetch these URLs from section 10 before advising:

• •WordStream benchmarks (CPC, CVR, CPA data changes yearly)
• •Google's QS documentation (Google updates scoring factors)
• •Match type trends (match types are actively evolving)

Gather inputs

Ask the user for as much of the following as available:

• •Campaign structure (campaigns, ad groups, locations)
• •Time period being reviewed
• •Key metrics: impressions, clicks, CTR, CPC, conversions, CVR, CPA, ROAS
• •Quality Score data per keyword (if available)
• •Search terms report (or a summary of problematic terms)
• •Current bid strategy
• •Current match types in use
• •Monthly budget and spend
• •Business targets: target CPA, target ROAS, or revenue goal
• •Any known issues or concerns

If the user can export data, accept CSV/spreadsheet data and parse it.

Audit framework

1. Goal alignment check

Before looking at any metrics, verify:

• •Does the account have a defined target CPA or ROAS?
• •Is conversion tracking working correctly? (What action is being tracked? Is it the right action?)
• •Are auto-apply recommendations turned off?

If any of these are wrong, flag as critical — nothing else matters until these are fixed. Reference knowledge base section 7: these are the top-3 most damaging structural mistakes.

2. Quality Score diagnosis

Reference the QS→CPC impact table in knowledge base section 1b.

For each keyword with QS data:

• •QS 7–10: healthy. Low priority for optimisation.
• •QS 5–6: average. Check which component is "below average" (expected CTR, ad relevance, or landing page experience). Prioritise the weakest component.
• •QS 1–4: urgent. These keywords are paying 25–400% more than necessary. Either fix the alignment (ad copy, landing page) or pause the keyword.

Provide specific recommendations per component:

• •Expected CTR below average: ad copy isn't compelling enough. Review headlines for authority/social proof signals. Test new variants.
• •Ad relevance below average: keyword doesn't match the ad closely enough. Tighten ad group theming — move the keyword to a more specific ad group or write keyword-matched headlines.
• •Landing page below average: the page doesn't match the search intent. Check H1 match, load speed, mobile experience, content relevance.

3. Keyword efficiency analysis

For each keyword, calculate:

• •CPA = spend / conversions
• •Compare CPA to the business's break-even CPA (from knowledge base section 1c formula)
• •If tracking revenue: ROAS = revenue / spend. Compare to break-even ROAS.

Categorise keywords:

Flag statistical significance: a keyword with 5 clicks and 0 conversions is not necessarily bad — it's just insufficient data. Rule of thumb: need 100+ clicks before making conversion-rate judgements.

4. Search terms review

If the user provides search terms data:

• •Identify high-converting search terms not yet added as keywords → recommend adding as exact match
• •Identify irrelevant search terms → recommend adding as negative keywords
• •Check for informational queries triggering ads → add as negatives unless there's a top-of-funnel strategy
• •Estimate wasted spend on irrelevant terms

If no search terms data: remind the user this is the most important regular review task. Instruct them to export it from Google Ads → Keywords → Search Terms.

5. Match type assessment

Reference knowledge base section 3.

• •If using phrase match: check CPC trend. If rising without conversion improvement, recommend testing exact match replacements.
• •If using broad match without Smart Bidding: flag as high risk. Recommend switching to exact match or enabling Smart Bidding.
• •If conversion volume is 30+/month and still on manual CPC: recommend testing broad match + Target CPA.
• •If conversion volume is <30/month: stay on exact match + manual CPC. Not enough data for automation.

6. Budget reallocation

Compare performance across campaigns (cities) and ad groups (services):

Apply the 70/20/10 framework:

• •70% to proven performers (best CPA/ROAS)
• •20% to promising campaigns that need more data
• •10% to experimental campaigns or new locations

Flag any campaign where impression share is very low (<20%) — this means budget is too thin to compete. Either increase budget or narrow keyword set.

7. Seasonal check

Reference knowledge base section 6 benchmarks.

• •Is the current season aligned with ad copy? (e.g. are "Book your summer wedding" ads still running in October?)
• •Are budgets adjusted for seasonality? (Wedding searches peak Jan–Mar)
• •Is scarcity messaging up to date? (Dates/months mentioned in ads must be current)

8. Retargeting health check

• •Is a remarketing audience set up? (All page visitors who didn't convert)
• •If running: check frequency cap (max 3/day recommended), creative freshness, spend level
• •If not running: recommend setup. Previous visitors convert at ~70% higher rate.

9. Follow-up process check

Beyond the ad platform:

• •Is there a follow-up email sequence for enquiries? (Day 0: quote, Day 3: video, Day 7: check-in, Day 14: scarcity)
• •What's the average response time to enquiries?
• •What's the enquiry-to-booking close rate?

If close rate is low, the problem may not be the ads — it may be the follow-up.

10. Mistakes checklist

Run through knowledge base section 7. For each common mistake, check whether the account is committing it:

• • Clear target CPA/ROAS defined
• • Conversion tracking verified and correct
• • Keyword-ad-landing page alignment per ad group
• • No unconstrained broad match
• • 100+ negative keywords
• • Search terms reviewed in last 7 days
• • No informational keywords without strategy
• • Auto-apply recommendations OFF
• • Active management (not "set and forget")
• • Match type understanding current

Output format

Produce a structured audit report:

1. •Executive summary: 3–5 bullet points, most critical findings
2. •Critical issues (fix immediately): anything that's actively wasting money
3. •Optimisation opportunities (improve this week): QS improvements, keyword changes, budget shifts
4. •Monitoring items (review next week): insufficient data, tests to continue
5. •Action checklist: numbered, prioritised list of specific changes to make

Be specific. "Pause keyword X" not "consider reviewing keywords." "Add 'lessons' as negative" not "review negative keywords." Actionable > advisory.

Context: $ARGUMENTS