An ecommerce Google Ads audit should not start with campaign names, optimization scores or a list of settings someone dislikes. Start with the money trail.
Can you reconcile store orders to the purchases Google Ads is counting? Is the account bidding toward the right actions? Do the targets reflect margin, returns and new-customer value? Until those answers are clear, a tidy campaign structure can still optimize the wrong business outcome.
This Google Ads audit checklist is organized in commercial order: economics and measurement first, then product data, demand capture, campaign roles and landing pages. Each section ends with one of three outputs: keep, fix or investigate.
1. Reconcile store orders before judging ROAS
Choose a recent period with stable tracking and compare three records:
- completed ecommerce orders and revenue in the store platform;
- purchase events and value in analytics; and
- purchase conversions and conversion value attributed in Google Ads.
The totals will not match perfectly. Attribution windows, time zones, consent, refunds and reporting delays can create legitimate differences. What matters is whether the gap is understood and reasonably consistent.
Investigate large or erratic gaps before changing bids. Common causes include duplicate tags, thank-you pages that fire again on reload, imported analytics events sitting beside native Google Ads conversions, test orders, missing transaction IDs and conversions counted with the wrong value.
If the Ads account reports materially more purchases than the store, use BPM's duplicate ecommerce purchase conversion diagnostic before treating reported ROAS as spendable revenue.
Output
- Keep: the difference is explained, stable and documented.
- Fix: a known implementation error is inflating or suppressing purchases or value.
- Investigate: the teams cannot yet trace an order from checkout through the advertising report.
2. Check what bidding is actually optimizing
Open the conversion-goal settings at both the account and campaign levels. List every action that appears in the Conversions column and answer two questions: Is this a real business outcome? Should this campaign bid toward it?
Google distinguishes primary conversion actions, which can be used for bidding when the associated goal is selected, from secondary actions used mainly for observation. Google also notes an important exception: a secondary action included in a custom goal can still be used for bidding when that custom goal is applied to a campaign.
For an ecommerce account, purchases are usually the central primary action. Add-to-cart, begin-checkout, page-view and time-on-site events can be useful diagnostics, but promoting all of them into the bidding target can reward cheap activity instead of sales.
Record the source, counting method, attribution setting, value behavior and campaign coverage for every action. Do not delete a historical conversion merely because its name is messy. First determine whether it is used by active campaigns and whether another action depends on it.
3. Translate revenue targets into contribution economics
A target ROAS is not automatically a profitability target. It is a ratio inside the ad platform.
Before calling a campaign efficient or inefficient, document:
| Input |
|---|
| Average order value |
| Gross margin |
| Discounts and shipping subsidy |
| Return and cancellation rate |
| Payment and fulfillment costs |
| New-customer value |
| Acceptable acquisition cost |
An account can hit a platform ROAS target while leaning on low-margin products, branded demand or returning customers who would have purchased anyway. The audit should state what the target protects and what it ignores.
4. Audit new-customer definitions before paying a premium
If campaigns use new-customer acquisition settings or separate reporting, inspect the definition rather than accepting the label.
Ask how the platform distinguishes new from existing customers, which customer lists or tags are supplied, how long the customer history goes back, whether guest checkout is common and how repeat purchases are handled across devices and emails. Compare the Ads result with the store or CRM definition for the same period.
Do not value a “new customer” more highly until the business agrees on the definition and can test the classification. A setting is not proof of incrementality.
5. Review the product feed as advertising input
For Shopping and retail Performance Max, the product feed is part of targeting, creative and eligibility. Google says accurate, correctly formatted product data helps match products to relevant queries; missing or conflicting identifiers, variant attributes, images, price or availability can limit eligibility or cause disapprovals.
Audit both diagnostics and commercial usefulness:
- disapprovals, warnings and destination issues;
- titles that reflect how buyers describe the product;
- correct GTIN, brand, product type and variant grouping;
- price, sale price and availability matching the landing page;
- clear images without avoidable clutter;
- shipping and return information;
- custom labels that support margin, seasonality, inventory or business priority; and
- products spending despite weak availability, poor contribution or no strategic role.
Use BPM's Google Shopping feed optimization checklist for the field-level review. The audit's job is to decide which feed problems affect revenue first.
6. Separate product performance from campaign averages
Campaign-level ROAS can hide a portfolio problem. Break performance out by item ID, product type, brand, margin group and inventory status.
Look for products that consume spend without conversions, profitable items that rarely receive traffic, bestsellers that dominate the average and low-margin products that appear efficient only because the platform sees revenue rather than contribution.
Then assign one action to each meaningful group:
- Keep: economics and traffic are healthy enough to continue.
- Fix: the product, feed, landing page or measurement has a specific repair.
- Investigate: there is too little clean data or the commercial role is unclear.
- Reduce or exclude: the business case does not support continued spend right now.
- Invest: the product has proven economics and a clear opportunity for more qualified demand.
7. Separate branded demand from demand creation
Blend branded and nonbranded search together and the account can look healthier than it is. Brand queries often represent people already looking for the company or product. Nonbrand campaigns must compete for less-certain demand.
Review campaign segmentation, search terms, brand exclusions and budget allocation. Google's search-terms report shows the significant searches that triggered ads, but some low-volume terms are omitted for privacy. Treat the report as useful evidence, not a complete transcript of demand.
For Performance Max, Google provides brand exclusions that can apply across Search and Shopping inventory, with options for retail advertisers to preserve branded Shopping traffic while excluding brand from Search text ads. The right setup depends on the role assigned to each campaign; the audit should make that role explicit.
Use BPM's branded versus nonbranded search ROAS guide to build a cleaner comparison.
8. Give every campaign a job
An ecommerce account does not become strategic because it has many campaign types. Write one sentence describing each active campaign's job.
Examples:
- protect high-intent branded searches;
- capture category demand with Search;
- sell an approved product group through Shopping inventory;
- find incremental demand across Performance Max inventory;
- remarket to an eligible audience; or
- test a new offer with a defined budget and learning question.
Flag campaigns whose role duplicates another campaign, whose budget is too small to answer the stated question, or whose success depends on conversions the business does not trust. For a deeper partner-scope review, use BPM's Performance Max agency evaluation guide.
9. Read search terms for intent and waste
Review meaningful search terms by campaign and product group. Label them as brand, category, competitor, informational, irrelevant or unclear. Check whether the landing page and product assortment satisfy the query. The SKU-level ecommerce PPC agency scope shows how product economics, feed status and ownership should carry into that review.
Do not add negatives mechanically because a query did not convert in a short window. Consider spend, clicks, commercial relevance, product availability and the amount of reliable conversion data. One expensive irrelevant term is different from a relevant category query that needs a better landing page.
The output should be a short set of actions: exclusions, new keyword or theme tests, product-title improvements, landing-page changes and questions requiring more data.
10. Audit the landing page after the traffic diagnosis
If qualified clicks arrive but purchases do not, inspect the destination before demanding another campaign rebuild.
Check mobile speed and usability, price and availability consistency, product detail, shipping and returns, reviews or proof, variant selection, checkout friction and whether the page fulfills the promise made in the ad or feed.
BPM's ads-versus-landing-page diagnostic helps separate a traffic-quality problem from a conversion-path problem.
Copy this ecommerce Google Ads audit worksheet
| Audit area | Evidence | Decision | Owner |
|---|---|---|---|
| Store orders vs. Ads purchases | Store, analytics and Ads comparison | Keep / Fix / Investigate | |
| Primary conversion goals | Goal and action export | Keep / Fix / Investigate | |
| Margin-based target | Contribution inputs and assumptions | Keep / Fix / Investigate | |
| New-customer definition | Ads, store and CRM comparison | Keep / Fix / Investigate | |
| Feed eligibility and quality | Merchant Center diagnostics and sample products | Keep / Fix / Investigate | |
| Product-level economics | Item/category report with margin and inventory | Keep / Fix / Investigate | |
| Brand vs. nonbrand | Campaign roles, search terms and exclusions | Keep / Fix / Investigate | |
| Campaign roles | One-sentence job for every active campaign | Keep / Fix / Investigate | |
| Search intent | Search-term sample and resulting actions | Keep / Fix / Investigate | |
| Landing-page fit | Mobile purchase-path review | Keep / Fix / Investigate |
What should happen after the audit?
Do not end with 40 observations sorted by campaign. Prioritize the few findings that change the reliability of the data or the economics of the account.
The usual order is:
- repair broken purchase measurement;
- correct the goals used for bidding;
- document economics and acquisition definitions;
- resolve feed and product-level blockers;
- clarify brand, nonbrand and campaign roles; and
- improve search-term controls and landing pages.
Some findings deserve an immediate fix. Others need a controlled test. A good audit says which is which, who owns the work and what evidence will decide whether the change stays.
If you want a second set of eyes on measurement, product priorities and campaign roles, explore BPM's paid search management or schedule a conversation.



