Insights
Google Ads audit checklist: 25 checks, ranked by profit impact
Google Ads audit checklist for 2026: tracking, money flow, bidding targets, Performance Max and landing pages, ranked by how much spend they save.
Most Google Ads audits are long lists of best practices. That makes them hard to act on. This checklist is ordered differently: by how much money each problem usually costs. Start at the top and stop when you run out of time. You will have fixed the expensive problems first.
1. Measurement: is the data true?
Smart bidding optimises toward whatever you report as a conversion. If that signal is wrong, everything downstream is wrong too.
- Primary conversions are real business outcomes. Purchases, qualified leads and booked calls, not page views, scrolls or “contact page visits”. Secondary actions should be set to secondary so they don’t steer bidding.
- No double counting. Check for duplicate purchase tags (for example, a GA4 import and a Google Ads tag both counting purchases) and for leads counted “every” time instead of “one” per click.
- Conversion values reflect profit. For e-commerce, consider sending gross profit rather than revenue, or at least excluding tax and shipping. For lead gen, assign different values to lead types.
- Consent Mode v2 is live for EU traffic. Without it, Google loses conversion and audience data from users in the EEA.
- Enhanced conversions are enabled. They recover conversions that cookies miss by matching hashed first-party data.
- Offline conversions flow back from the CRM. If you sell through a sales team, import qualified lead, opportunity and won stages. Otherwise the algorithm optimises for form fills.
- Ads, GA4 and the backend roughly agree. A gap of 10–20% is normal. A gap of 50% means something is broken.
2. Money flow: where does the budget actually go?
- Brand vs non-brand split. What share of spend and conversions comes from people already searching your brand name? Brand traffic is cheap and converts well, which flatters account-level ROAS.
- Search terms from the last 90 days. Sort by cost. Look for irrelevant terms, job seekers, “free”, “DIY”, competitors you didn’t intend to target and support queries.
- Locations and location options. “Presence or interest” can show ads to people who are interested in a location but not in it. For local businesses, use presence.
- Devices, days and hours. Check for segments with spend but no conversions over a meaningful period.
- Display and partner network leakage. Search campaigns with Display expansion or search partners enabled often spend quietly on low-intent inventory.
3. Bidding: are the targets calculated or guessed?
- Targets come from your economics. Target ROAS should start from your break-even ROAS (1 ÷ gross margin). Target CPA should start from what a customer is worth. If nobody can explain where a target came from, it was guessed.
- Each campaign has enough conversion data. Smart bidding needs volume. A campaign with a handful of conversions a month is guessing. Consolidate, or optimise towards a higher-volume action first.
- Different margins, different targets. One ROAS target across products with 15% and 60% margins loses money on the first group and under-invests in the second.
4. Performance Max: what is it really buying?
- Brand exclusions are set. Otherwise PMax often harvests brand searches and reports them as new wins.
- New vs returning customers. If most PMax conversions are repeat buyers or remarketing, the campaign may not be growing the business.
- Search-term and channel reports are reviewed. Check where PMax spends (Search, Shopping, YouTube, Display) and which queries it matches. Add account-level negatives where needed.
- Feed segmentation. Best-sellers, low-margin products and products with no impressions should not all sit in one asset group.
5. Structure and ads
- Match types are deliberate. Broad match can work with strong conversion data and smart bidding; with weak data it spends on loosely related searches.
- No overlapping campaigns competing for the same queries with different budgets and targets.
- Ads are tested against hypotheses, not rotated at random. “Ad strength” is a guide to asset variety, not a predictor of profit.
6. Feed and landing pages
- Merchant Center is clean. Check disapprovals, price and availability mismatches, missing GTINs and policy warnings.
- Titles match how people search. Brand + product type + key attribute (size, material, model).
- Landing pages load fast and match the ad. If the ad promises a price, offer or product, the page should show it immediately.
How to prioritise the fixes
Score each issue on two axes, profit impact and effort, and do the high-impact, low-effort fixes first. In our experience the biggest wins almost always come from measurement (1–7) and bidding targets (13–15), not from new keywords or ad copy.
If tracking is broken, every other optimisation is a guess. Fix the signal, then let the algorithm work.
Want a second pair of eyes? Our Google Ads audit works through this checklist on your account and ranks the fixes by profit impact.