How to Audit a Google Ads Account (Ecommerce Checklist)

Most Google Ads audits start in the wrong place. They open the campaigns, look at the bidding, and produce a list of tidy structural suggestions. Meanwhile the account is reporting conversions that never happened, paying for its own brand name, and sending traffic to a product page that takes six seconds to load.

This is the order I actually work in when I take over an ecommerce account, which is roughly the order of how much money each thing tends to be costing. It is written for a Shopify store spending somewhere between $5,000 and $100,000 a month.

Start with the tracking, not the campaigns

Everything downstream is a guess if the measurement is wrong, and in most accounts I inherit it is wrong in at least one way. Smart Bidding does exactly what the conversion data tells it to, so bad data does not produce mild underperformance — it produces confident, well-optimised spending in the wrong direction.

• Count the conversion actions. Is more than one of them set to “Primary”? A purchase and an add-to-cart both counted as primary means bidding is optimising toward carts, and carts are cheap.
• Compare Google Ads conversions to Shopify orders for the same period. They will never match exactly, because of attribution windows and view-through. If they are out by more than roughly 20–30%, something is broken rather than merely different.
• Check for double counting. A conversion sitting in both a Google Ads tag and an imported GA4 key event will be counted twice.
• Look at the conversion window. A 90-day window on a product people buy the same day flatters every campaign in the account.
• Check enhanced conversions are on and actually receiving data, not just enabled.

What should you look at first in a Google Ads audit?

The search terms report is the only place the account tells you the unvarnished truth about what you are buying. Set the date range to the last 90 days, sort by cost, and read down the list.

You are looking for three things: terms that spend with no conversions, terms that are obviously the wrong intent (“free”, “cheap”, “DIY”, “how to make”, a competitor’s brand you did not mean to bid on), and terms converting well that are not yet keywords in their own right.

That third one is where the money is, and almost every audit checklist omits it because cutting waste makes a better screenshot than harvesting winners.

How much should you pay for your own brand name?

Branded search is the most reliably misread line in an ecommerce account. It always looks magnificent — high conversion rate, low cost per acquisition, enormous ROAS — because the people searching your brand name had already decided to buy.

The question is not whether branded is profitable. It is how much of that revenue you would have got for free from the organic result immediately below your ad.

• Split branded into its own campaign if it is not already. You cannot manage what you cannot see.
• Work out what share of total account conversions it is responsible for. Above about 40% and your headline ROAS is mostly measuring your existing customers.
• Check whether competitors are actually bidding on your brand. If nobody is, your cost per click should be pennies — and if it is not, you may be bidding against yourself.

Open up Performance Max

Performance Max is where budget goes to become unaccountable. It will happily absorb branded search, existing customers and remarketing, then report all of it as new performance.

• Add brand exclusions unless you have deliberately decided otherwise. Without them PMax will buy your brand name and take the credit.
• Check the asset group report for search themes and audience signals that have drifted from what you sell.
• Look at the channel split if you can get at it. A PMax campaign spending most of its budget on Display is usually buying cheap impressions, not customers.
• Compare it against a standard Shopping campaign on the same products before concluding it is winning.

Check the feed, not just the ads

For most ecommerce accounts Shopping does more of the work than text ads, which makes the product feed a bigger lever than the campaign structure. It is also the thing audits skip most often, because it lives in Merchant Center rather than Google Ads.

• Disapprovals and warnings, especially on your highest-margin products
• Titles: does the title contain the words people actually search, in the order they search them?
• GTINs, brand and product type filled in, because they feed the matching
• Price and availability mismatches between the feed and the live store
• Out-of-stock products still being advertised

Why is a profitable campaign not spending its budget?

Now, and only now, the campaign mechanics.

• Budget-limited campaigns that are profitable. This is free money and it is surprisingly common.
• Target ROAS set from a number nobody remembers choosing. If it was set against broken conversion data, it is arbitrary.
• Campaigns still in learning because someone edits them every few days.
• Ad schedule and location settings, particularly “Presence or interest” targeting quietly advertising you to countries you do not ship to.

Look at the pages the clicks land on

An audit that stops at the platform boundary has skipped the half of the problem you can most cheaply fix. The ad bought a visit; the page decides what it was worth.

• Load the top five landing pages by spend on a phone, on a normal connection, and time them — speed is usually a Shopify build problem, not an ads one
• Check the offer on the page matches the promise in the ad
• Look at the mobile add-to-cart: is it visible without scrolling?
• Compare conversion rate by landing page — the gap between best and worst is usually larger than any bidding change you could make

What to fix first

A finished audit is not a list of everything wrong. It is an ordered list of what to do on Monday. This is roughly how I rank findings:

Notice that restructuring — the thing most audits lead with — is last. It is the most visible work and rarely the most valuable.

Questions people ask