Google Shopping Agency Audit Checklist: What Should Be Checked First

The short answer

A good Google Shopping audit does not start with bid changes. It starts by asking a colder question: is the account giving Google clean enough commercial information to spend money well?

Most weak Shopping accounts have the same pattern. The campaign looks active, the ROAS looks passable in places, but nobody can clearly explain which products are carrying the account, which products are quietly wasting budget, or whether the feed is helping Google match products to the right searches.

That is what an agency should check before it talks about scaling.

What should be checked first

Start with the areas that can distort every later decision.

Audit area	Why it matters
Conversion tracking	Bad revenue data makes every ROAS decision unreliable
Merchant Center health	Disapprovals and warnings suppress products before campaigns can work
Product feed quality	Titles, attributes, and product types influence matching and eligibility
Product segmentation	Blended reporting hides winners, losers, margin issues, and stock problems
Campaign structure	PMax, Standard Shopping, and listing groups need a reason to exist
Landing pages	Shopping traffic is high intent, but weak product pages still leak money

If tracking is wrong or Merchant Center is messy, everything downstream becomes guesswork.

Feed quality

The feed is where many audits become useful. Not because the feed is glamorous, but because it is the product language Google has to work with.

An agency should check product titles, product type, Google product category, descriptions, images, GTINs, availability, pricing, variants, shipping, sale price, and custom labels. The audit should not just say "optimise titles". It should show which product groups need better wording and why.

Example: if a furniture brand sells "Oxford 3 Drawer Oak Bedside Table" but the title only says "Oxford Table", the feed is hiding important buying signals. The fix is not poetic copywriting. It is clearer product data.

Google's own product data specification is worth reading because it explains the unglamorous truth: Google uses your product data to understand and match products. That means feed work is not an optional SEO-style layer. It is part of the ad system.

In a proper audit, I would normally separate feed issues into three buckets:

• Eligibility problems that stop or limit products
• Matching problems that make products show for weaker intent
• Commercial-control problems that make products hard to segment or report on

Those buckets matter because they lead to different work. An eligibility issue might need Merchant Center or policy fixes. A matching issue might need product-title and product-type work. A commercial-control issue might need custom labels, category cleanup, and campaign restructuring.

Merchant Center problems

Merchant Center issues are easy to ignore because the account can still spend while products are limited, disapproved, or underperforming.

Check:

• Disapproved products
• Limited products
• Price mismatches
• Availability mismatches
• Shipping and returns warnings
• Missing identifiers
• Image policy issues
• Diagnostics by product category

The audit should prioritise by commercial impact. A disapproval on one low-margin accessory is not the same as a warning across a best-selling category.

Product waste

Product-level waste is where a Shopping audit often pays for itself.

Look for products that:

• Spend but do not convert
• Convert but at weak margin
• Drive clicks with poor product-page conversion rate
• Receive budget because they sit in the same campaign as stronger products
• Are out of stock often
• Have weak images or uncompetitive pricing

A good Google Shopping agency should not only say which products are wasting spend. It should explain whether the fix is feed work, exclusion, different structure, pricing, stock, landing page work, or a different bid target.

This is also where blended ROAS becomes dangerous. A campaign can show a respectable return while one group of products quietly subsidises another. If a best-selling category is carrying a long tail of products that never convert, the account can look stable while profit leaks underneath.

A better audit looks at the uncomfortable middle layer:

• Products with high spend and no sales
• Products with sales but poor margin
• Products with clicks but low add-to-cart rate
• Products with good conversion rate but limited impression share
• Products that sell only when heavily discounted
• Products that deserve separate targets because their economics are different

The point is not to cut every weak product instantly. Sometimes a product is a gateway product. Sometimes a new product needs data. Sometimes a low-ROAS product has a high repeat-purchase value. But the account should know the difference between a strategic exception and lazy budget leakage.

PMax versus Standard Shopping

Many audits need to answer a simple question: should these products be in Performance Max, Standard Shopping, or both?

PMax can be excellent when the account has strong tracking, good feed data, useful creative, and enough conversion volume. Standard Shopping can be better when you need clearer search-term control, product isolation, or cleaner testing.

The audit should check whether PMax is carrying everything because it is genuinely best, or because nobody wanted to build a more deliberate structure.

Landing page checks

Shopping traffic lands deep in the site, usually on product pages. The audit should therefore look at the page receiving the spend.

Check:

• Product image quality
• Price and offer clarity
• Delivery information
• Returns information
• Reviews and trust signals
• Variant selection
• Stock status
• Page speed
• Mobile checkout friction

Sometimes the Shopping account is fine and the product page is doing the damage. An honest agency will say that.

For ecommerce, I like to check the product page as if I had just clicked the ad with mild scepticism. Can I see the product clearly? Do I know the delivery cost? Is the returns policy obvious? Are reviews near the decision point? Are variants easy to choose on mobile? Does the page answer the questions that would stop someone buying today?

This matters because Shopping traffic is not magic. It is often high intent, but high intent can still leave if the page creates friction. A campaign manager who never looks at the product page is only managing half the problem.

What the audit should produce

The output should not be a 40-page deck of screenshots.

It should produce:

• The biggest tracking risks
• The Merchant Center issues that matter commercially
• Feed changes by category or product group
• Products to exclude, isolate, or improve
• Campaign structure recommendations
• Landing page problems affecting spend
• A first 30-day action plan

The best audit feels slightly uncomfortable because it names the trade-offs. It should make the next decision easier, not just make the account look busy.

A useful first-hour audit flow

If I had one hour inside a Shopping account, I would not start by clicking through every campaign setting.

I would usually check:

1. Whether purchase tracking and revenue look believable
2. Merchant Center diagnostics for commercially important product issues
3. Top spending products over the last 30 to 90 days
4. Products with spend and no conversion value
5. Product categories carrying most revenue and most wasted spend
6. Whether PMax and Shopping are overlapping sensibly
7. Whether product pages match the products receiving spend
8. Whether the next action is feed, structure, bid target, product exclusion, or landing page work

That first pass is enough to find the shape of the problem. The full audit then adds detail, evidence, and priorities.

How to prioritise the fixes

The hardest part of a Shopping audit is rarely finding issues. Most accounts have plenty. The hard part is deciding which issues deserve attention first.

I would usually prioritise in this order:

1. Tracking problems that make performance data unreliable
2. Merchant Center issues affecting important products
3. Feed problems across high-spend or high-potential categories
4. Product groups wasting meaningful budget
5. Structure issues that hide performance differences
6. Landing page problems affecting products already receiving traffic
7. Smaller feed cleanups that improve hygiene but do not change decisions yet

This order matters because it prevents cosmetic optimisation. It is easy to spend days tidying product descriptions while a broken purchase tag, a disapproved hero category, or a bad PMax structure continues to waste money.

What to ask the ecommerce team

A good audit should not happen entirely inside Google Ads. The ecommerce team usually knows things the platform cannot show clearly.

Useful questions include:

• Which products have the best margin?
• Which categories are strategic this quarter?
• Which products are overstocked?
• Which products have supply issues?
• Which products drive repeat purchases?
• Which products look good on revenue but bad after returns?
• Which products are price-sensitive?
• Which product pages have recently changed?

These answers can completely change the recommendation. A product that looks weak in Google Ads might be worth keeping if it drives repeat customers. A product that looks strong on ROAS might be less attractive if the margin is poor or returns are high.

The audit should bring platform data and business context together. Otherwise, the recommendations will be mathematically tidy but commercially thin.

A 30, 60, 90 day improvement plan

The audit should lead to a practical sequence of work.

In the first 30 days, fix measurement, Merchant Center issues, obvious product waste, and the biggest feed problems affecting important products. This is the stabilisation phase.

By 60 days, the account should have better product segmentation, clearer custom labels, cleaner reporting, and a stronger view of which products deserve more or less budget.

By 90 days, the focus should shift toward scaling the stronger product groups, testing structure, improving product pages, and building a repeatable feed-maintenance process.

That timeline keeps expectations realistic. Shopping improvements can be quick when the issue is obvious, but stronger product data, cleaner segmentation, and better commercial reporting usually compound over a few cycles.