Scan your account in 90 seconds — free.
d-lens · 7-day trial · no credit card · read-only
// guide · google ads audit
Google Ads Audit: A 2026 Field Guide + 12-Step Checklist.
A real audit — one that actually reads your account — typically surfaces 15-30% of spend silently leaking to waste sources you can fix in a week. This guide defines what an audit is, how often to run it, manual vs AI-assisted approaches, and a 12-step checklist you can apply from scratch. At the end, an autonomous agent option (d-lens) that runs the same scan in 90 seconds.
// short answer
A Google Ads audit is a systematic review of an account's structure, settings, performance, and cost — finding hidden waste and missed conversion opportunities. 12-module summary: search terms, quality score, match types, negatives, audience overlap, conversion tracking health, automated bidding data, campaign structure, creative rotation, Performance Max data leakage, attribution model, device + location performance. Manual audit takes 4-8 hours; AI-assisted (e.g. d-lens) averages 90 seconds.
// 01What is a Google Ads audit?
A Google Ads audit is a systematic review of a paid advertising account across structure, settings, performance, and cost. Three goals:
- Find hidden waste. Which keywords, campaigns, audiences or devices burn budget without producing conversions?
- Spot missed opportunities. Which queries hit your CPA target but receive too little investment? Which campaign types are under-tested?
- Bring the account closer to a healthy reference. Is conversion tracking firing correctly? Is the automated-bidding algorithm fed enough data?
Audits aren't one-off inspections. Campaigns are live systems; budget flow, competition, seasonality, creative fatigue and platform algorithms shift constantly. A good audit discipline is therefore periodic and trigger-driven — when CPA breaches a threshold, when conversion volume drops.
// important distinction
Audit ≠ optimization. Audit is diagnosis; optimization is treatment. During an audit, the account is read-only — you produce the table. The action decision comes after reading the audit report.
// 02Why audit at all?
Even teams running Google Ads for years tend to have 15-30% of spend bleeding into waste areas when an outside eye (or a good tool) looks. This isn't an inflated figure — typical 2024-2026 industry baselines:
| Waste area | Typical impact | Detection difficulty |
|---|---|---|
| Irrelevant search terms (search waste) | 8-15% of budget | Low (reportable) |
| Audience overlap | 4-10% of budget | Medium |
| Wrong / missing conversion tracking | impacts entire budget | Medium-high |
| Performance Max data leakage | 5-12% of budget | High |
| Wrong device/location bidding | 3-8% of budget | Low |
The takeaway: if you run a $100K/month account, an audit typically surfaces $15-30K in recoverable inefficiency. The audit is the cheapest, fastest first step to recovering that money.
// fast path
Scan My Account →
// 03Manual vs AI-assisted audit
Manual audit — analyst's eye
A good analyst can audit an account in 4-8 hours. Strength: business-context interpretation — "why is this campaign capping?", "is brand-protection spend justified here?". Weakness: comprehensive scanning exceeds human attention span; cross-checking quality score against device-bid curve simultaneously is exhausting and gets skipped.
AI-assisted audit — agent approach
An autonomous audit agent (e.g. d-lens) scans across 46+ modules in parallel. Strength: speed (seconds) and breadth (dimensions humans skip). Weakness: business context interpretation is shallower — the "why" question still needs an analyst.
Practical recommendation: combine the two
Professional teams combine: agent runs 46+ modules in parallel, analyst interprets the resulting table against business goals. This way, every finding has a concrete next-action attached ("pause this campaign", "switch this match type"), and the action's business meaning is human-validated.
// 0412-step audit checklist
The 12-step audit, applied from scratch in order. For each step: what to check, which report to open, which threshold to watch.
Google Ads Audit Checklist
- Conversion tracking health — Is each conversion action collecting data over the last 30 days? Are tag "Recording" indicators green? Is the gap between GA4 and Google Ads conversion counts under 10%? tools → conversions
- Search terms report — What % of last-30-days queries got clicks but zero conversions? How many of those are irrelevant (e.g. "free", "jobs")? reports → search terms
- Match type distribution — Is broad match's share of spend over 60%? Are negative keyword counts adequate (50+ per campaign)?
- Quality score — landing page experience — What's the spend share of keywords with QS <5? Are those keyword landing pages relevant?
- Automated bidding data sufficiency — Do tCPA / tROAS campaigns have under 30 conversions in the last 30 days? (Algorithm runs poorly with insufficient data.)
- Campaign structure — Does each campaign serve a single business goal? Are the same keywords competing across multiple campaigns? (i.e. internal cannibalization)
- Audience overlap — Are remarketing + similar audiences + interest layers showing the same impression to the same user? What's the overlap percentage in Audience Insights?
- Performance Max data leakage — Is PMax cannibalizing brand queries? Is a brand-exclusion list in place? Are asset-group search-theme insights being reviewed?
- Device bid adjustment — Does mobile vs desktop conversion rate diverge by >30%? Is a device bid adjustment applied?
- Location performance — Which cities consume X% of budget but contribute a much smaller share of conversions? Are geo-negatives in place?
- Creative rotation (ad rotation) — Does each active ad group have at least 3 RSAs? How many "low" assets are flagged in asset performance?
- Attribution model + conversion window — Is the model still "last click"? Does the conversion window match your purchase cycle? (Short windows starve data-driven attribution.)
One analyst applying the 12 steps from scratch takes 4-8 hours. To shorten that — and most teams want to — there are tools that automate the audit (including d-lens, covered next).
// 05Common findings + how to fix them
Five findings that show up after hundreds of Google Ads audits, and the corresponding action:
Finding 1: Search waste 15%+
Typical cause: broad match overuse + insufficient negative list. Action: from the last 90 days' search terms report, bulk-add zero-conversion queries with 10+ clicks to the negative list. Consider revising broad-match-heavy campaigns to "exact + phrase" structure.
Finding 2: Conversion tracking 10%+ mismatch
Typical cause: GTM trigger double-counting or GA4 enhanced ecommerce misconfigured. Action: use Tag Assistant to verify how many times the conversion tag fires per page. If using server-side GTM, review request logs.
Finding 3: Performance Max cannibalizing brand
Typical cause: PMax has no brand-exclusion list and is pulling budget from the brand search campaign. Action: add the brand name under PMax campaign's "Brand exclusions". Track brand campaign impression share week-over-week.
Finding 4: tCPA bidding running on insufficient data
Typical cause: campaign has 15 conversions in last 30 days but tCPA target. Action: temporarily switch to "Maximize Conversions"; return to tCPA once conversions stabilize at 30+/30 days.
Finding 5: Audience overlap 20%+
Typical cause: same user is being targeted across multiple audience layers (remarketing + similar audience + customer match). Action: measure overlap in Audience Insights. Remove the lower-performing audience or add it as a negative.
// 06How often to audit?
No single right answer, but a practical operating rhythm:
- Daily — 5 minutes: anomaly check (CPA, conversion volume, delivery errors).
- Weekly — 30 minutes: search terms, creative performance, budget pacing.
- Monthly — 2 hours: 6-8 of the 12 checklist steps.
- Quarterly — 4-8 hours: all 12 steps, attribution + audit report.
- Trigger-based — immediately on anomaly: budget jumped 30%, conversion volume dropped 25% week-on-week.
This rhythm is far more sustainable than the "audit everything every day" approach. The point is to make audits regular and trigger-aware — not heroic one-offs.
// 07d-lens — autonomous audit agent
Applying the 12-step checklist manually every month — especially across multiple accounts — is inefficient. That's why we built d-lens, an autonomous ad audit and action agent that starts from those 12 steps and expands to 46+ modules.
How it works:
- Connect — read-only OAuth to Google Ads, Meta, TikTok and GA4 (revocable any time).
- Scan — 46+ modules run in parallel; average scan time 90 seconds.
- Action recommendations — not just "campaign X has 18% search waste"; concrete steps: "add these 23 queries to the negative list".
- Validate — you or your analyst review, approve, and apply changes in the account. d-lens does not write to the account; the decision always stays with humans.
Pro plan is $199/month ($179/month on annual billing); 7-day free trial, no credit card. Multi-account packaging available.
// free trial
Scan My Account →
Try d-lens for 7 days.
read-only · no credit card · 90-second average scan
// 08FAQ
Is a Google Ads audit free?
d-lens offers a 7-day free trial (no credit card). During the trial you can scan Google Ads, Meta, TikTok and GA4 across 46+ modules. For manual audits, Google Ads' own "Recommendations" tab is a free starting point but misses critical context (tCPA data sufficiency, audience overlap, etc.).
Can I audit someone else's account without account ownership?
No. The account owner has to grant access. Permissions are granted via MCC or direct OAuth. d-lens requires read-only — it reads but never modifies the account.
Can Performance Max campaigns be audited?
Yes, but with limits. Google reports PMax detail under "asset group insights" only — visibility into which queries the campaign is investing in is restricted. d-lens runs five PMax-specific modules: brand exclusion, asset group performance, search-theme overlap, conversion attribution health, negative-keyword leakage.
How much can performance improve after an audit?
Depends on the account baseline and how many actions you apply. Typical range: an account with 10%+ search waste can drop CPA 15-25% within 30 days. At-scale campaigns see smaller percentage gains but still meaningful absolute spend recovery.
Does d-lens share customer data or train AI on it?
No. Per Google API Services User Data Policy "Limited Use" rules, d-lens does not use customer data to train AI models. Aggregations produce only fully anonymous, account-independent industry baselines.
For more guides on agentic AI marketing and ad operations, visit /en/guides/. Written by Mesut Şefizade, founder of d-dat.
// relatedRelated glossary terms.
Quick definitions for the concepts referenced in this guide:
// next step
Scan your account in 90 seconds.
d-lens reads your ad accounts across 46+ modules and writes a "do this" action for every finding. 7-day free trial, no credit card, read-only access — revocable any time.