The 30-Point Google Ads Audit Checklist (the one we automated)
The full checklist Advestigate runs, published openly. Work through it by hand if you have an afternoon — or see why we turned it into software.
This is the actual checklist behind Advestigate's audit engine, published in full. Nothing here is secret: every item is something a competent PPC person would check by hand. The product exists because 'by hand' takes an afternoon per account and gets skipped exactly when accounts need it most.
Work through it top to bottom. The order matters: conversion tracking comes first because a broken tag quietly invalidates every other number you'll look at.
Conversion tracking (do this first)
1. At least one active conversion action exists. 2. Every primary conversion action has recorded a conversion in the last 30 days — a tag that fired last quarter and stopped is the most expensive silent failure in Google Ads. 3. No duplicate primary actions in the same category (double counting inflates results and misleads Smart Bidding). 4. Every campaign with meaningful spend has attributed conversions — if one hasn't, either measurement or the campaign is broken. 5. Auto-tagging is enabled. 6. GA4 is linked so you have an independent view of post-click behavior.
Wasted spend
7. No broad match keywords running under manual CPC or Maximize Clicks — broad match without conversion-based Smart Bidding is Google matching loosely with nothing reining it in. 8. No keywords with significant spend and zero conversions over 90 days. 9. Display Network expansion is off on Search campaigns. 10. Search partners are reviewed against their actual conversion rate. 11. No ad groups burning budget inside otherwise-converting campaigns. 12. Location targeting is set to 'Presence', not 'Presence or Interest' — the default leaks spend to people merely searching about your area.
Keywords and negatives
13. Negative keyword lists exist and are applied to every spending campaign. 14. No negative keyword blocks a keyword that converts (this happens more than anyone admits — a broad negative added in a hurry). 15. Quality Score distribution is healthy; a large share of 4-and-below keywords means ads and landing pages don't match the searches. 16. No duplicate keyword/match-type pairs competing across ad groups. 17. Ad groups hold tight themes, roughly 5–15 keywords, so one ad can actually be relevant to all of them.
Ad copy and assets
18. Every ad group has at least one active ad (keywords with no ad silently stop serving) and ideally a test running. 19. RSA ad strength is Good or better on most ads. 20. No ad group depends solely on frozen legacy expanded text ads. 21. No disapproved or policy-limited ads sitting unnoticed. 22. At least 4 sitelinks and 4 callouts per campaign — assets are free ad real estate. 23. RSA headlines are genuinely distinct, not the same message rephrased eight times.
Budget and bidding
24. No profitable campaign is losing meaningful impression share to budget while unprofitable ones spend freely — this is the most common growth left on the table. 25. Budget allocation broadly follows performance. 26. Target CPA/ROAS values sit within ~20% of trailing actuals; aggressive targets don't make traffic cheaper, they make it disappear. 27. Campaigns with 15+ monthly conversions have tested Smart Bidding. 28. Brand campaigns hold 90%+ impression share — brand clicks are the cheapest traffic you can buy.
Structure and settings
29. Significant campaigns have had their hour/day performance reviewed (24/7 serving is a choice, not a default). 30. Spend isn't overwhelmingly parked in Performance Max with no Search campaigns for proven keywords. 31. The change history shows the account is actually being managed — fewer than a handful of changes in 90 days on a spending account is a finding in itself. 32. Old paused campaigns are archived so real problems stay visible.
That's the list. Advestigate runs all of it in about two minutes against 90 days of data, names the specific campaigns and keywords for each finding, and estimates the monthly cost where the data supports it. If you'd rather do it by hand: block out an afternoon per account, and actually do it quarterly — the decay between audits is where the money goes.
More like this on the blog index, or see the audit itself on the sample report page.