Google Ads specialist. $30M+ managed across 50+ e-commerce and agency accounts in the US, UK and India. Book a free audit call
On this page
Short version: Audit in the order money leaks, not the order the interface presents. Conversion tracking first, because every other number depends on it. Then wasted spend, structure, budget allocation, feed, creative, and finally bidding. Most audits find between five and fifteen percent of spend going somewhere it should not, and almost all of it sits in the first three sections.
An audit is not a list of every setting in the account. It is a search for the specific places where money is leaving without coming back. This checklist is the one we run on accounts before taking them over, arranged so that the highest-value findings surface first.
Work through it in order. The sequence matters because a finding in section one invalidates the analysis in every section below it. There is no point tuning bids against conversion data that is counting purchases twice.
1. Conversion tracking: does the data mean anything?
Everything downstream depends on this. Smart Bidding optimises toward the conversions you report, so if the reporting is wrong, the algorithm is confidently steering toward the wrong outcome.
Is every conversion action counting something real? Open Goals, then Conversions. For each action ask what it represents and whether anything else already counts the same event.
Check for duplicates. The classic failure is a platform app and a manual tag both firing on the same thank-you page, doubling reported revenue. Compare a full week of Google Ads conversions against the back-end order count, segmenting Google Ads by conversion date rather than click date so you are comparing like with like.
Is only one action set to Primary? Purchases should be primary. Add to cart and begin checkout should be secondary, so they stay visible in reports without feeding the bidding.
Is conversion value dynamic? A hardcoded average order value destroys value-based bidding. The tag must pass real order totals.
Is the attribution window sane, and is the model consistent across actions you compare?
Enhanced conversions enabled? If not, you are likely under-reporting and the bidding is working with less signal than it could.
If tracking is broken, stop here. Fix it, let two weeks of clean data accumulate, then resume. Any other conclusion drawn from a broken account is guesswork wearing a spreadsheet. For the Shopify-specific version of this, see our guide to Shopify Google Ads conversion tracking.
Audit in this order. Findings later in the list cannot be trusted until the earlier ones are clean.
2. Wasted spend: where is money going for nothing?
This section usually pays for the audit on its own.
Search terms report, last 90 days. Filter to clicks greater than zero and conversions equal to zero, sort by cost descending. The top twenty rows are your immediate answer.
Is there a negative keyword list applied to every campaign? Check Tools, then Shared library. A shared list of universal junk applied everywhere is the baseline hygiene most accounts lack. Our guide on how to add negative keywords covers match types and levels.
Are you paying for job seekers, researchers, and freebie hunters? Look for "jobs", "salary", "how to", "diy", "free", and "cheap" in the search terms.
Are you paying for products you do not stock? The biggest e-commerce leak by some distance.
Placement report on Display and Performance Max. Look for mobile game apps and low-quality sites absorbing impressions. Exclude aggressively.
Location settings. Check that "Presence" is selected rather than "Presence or interest", unless you deliberately want people merely interested in your area.
Is brand traffic separated from non-brand? Mixed together, brand conversions flatter everything and hide non-brand failure.
3. Account structure: can this account be managed?
Structure determines whether anyone can read the account and act on what they see.
Do campaign names follow a convention that tells you type, funnel stage, and geography at a glance?
Is brand in its own campaign with its own budget, so it cannot absorb prospecting spend?
Are ad groups tightly themed? An ad group with forty unrelated keywords cannot have relevant ad copy for all of them.
Are campaigns competing with each other for the same searches? Check for the same term triggering multiple campaigns.
Is Performance Max swallowing brand traffic? Without brand exclusions, PMax will claim conversions that would have happened anyway and report a flattering ROAS. Our piece on what Performance Max really does with your budget covers this.
How many campaigns are actually active, and does each have enough budget to exit the learning phase? Ten underfunded campaigns perform worse than three funded ones.
4. Budget and bidding allocation
Where is the money actually going? Sort campaigns by cost. Does the ranking match your commercial priorities, or has spend drifted toward whatever the algorithm found easiest?
Which campaigns are budget-limited? A campaign hitting its cap with a strong ROAS is a growth opportunity sitting in plain sight.
Are bid strategies appropriate? Target ROAS needs conversion volume to work. On a campaign with three conversions a month it will thrash.
Are targets realistic? A tROAS set far above what the account has ever achieved will simply throttle delivery until the campaign stops spending.
Is the ROAS target built on margin? A 4.0 ROAS on a 60 percent margin product and a 4.0 on a 20 percent margin product are completely different outcomes. See what counts as a good ROAS.
Check the change history before concluding anything. If someone changed targets last week, current performance reflects the learning period, not the strategy.
5. Product feed, for e-commerce accounts
On an e-commerce account the feed is the campaign. Shopping and Performance Max cannot show a product the feed describes badly.
Merchant Center diagnostics. How many products are disapproved, and how many are approved but limited? Both are invisible in Google Ads.
Are titles front-loaded with the words buyers actually search, rather than starting with an internal SKU or the brand name alone?
Are GTINs, brand, and product attributes populated? Missing identifiers restrict eligibility.
Are images clean and compliant? The image is the ad in Shopping.
Is pricing and stock synced? Mismatches between feed and site cause silent disapprovals.
Are products segmented by margin or priority, or is the whole catalogue in one undifferentiated bucket bidding equally on everything?
Does every ad group have at least one responsive search ad with a reasonable strength rating and genuinely distinct headlines?
Are all relevant assets in place? Sitelinks, callouts, structured snippets, images. Missing assets cost you both click-through rate and the space your competitors occupy.
Do ads match the landing page promise? An ad advertising a discount that the page does not mention wastes the click.
Do landing pages load fast on mobile? Run PageSpeed Insights. LCP above 2.5 seconds and CLS above 0.1 are both costing you conversions and ad rank.
Is the mobile experience actually usable, or is the form unusable below the fold?
7. Audiences, geography, and schedule
Are remarketing lists active and populated? Lists below the size threshold silently do nothing.
Are customer match lists uploaded and refreshed? Stale lists decay.
Does performance vary meaningfully by device, location, or time? Check before assuming it does not. Most accounts have at least one segment quietly underperforming.
Are exclusions in place for locations you cannot ship to or serve?
How to turn the findings into a plan
An audit that produces a forty-item list helps nobody. Sort what you find into three groups:
Fix now. Anything broken: tracking errors, disapproved products, obvious waste. These are not judgement calls and carry no risk.
Change deliberately. Structure, targets, and budget shifts. One meaningful change at a time, with at least a week between them so you can attribute the effect. Changing five things at once means learning nothing from any of them.
Test. New ad copy, new campaign types, new audiences. Genuine experiments with a hypothesis and a defined read date.
Resist the urge to act on everything in week one. An account that has three changes made carefully will beat one that has thirty made at once, because the second is unreadable for a month afterwards.
The five findings that show up in almost every audit
After enough audits the same problems recur, and they are rarely exotic. Knowing what usually breaks lets you check the likely culprits first rather than working through an account front to back.
Conversions counted more than once. A purchase fires from the Google Ads tag, from the Google Analytics import, and sometimes from a third tag installed by a previous manager. The account reports double or triple the real conversions, the ROAS looks excellent, and every bidding decision downstream is made on inflated numbers. The check is simple: compare Google Ads conversions to actual orders in the store backend for the same period. If they do not roughly agree, nothing else in the account can be trusted.
Brand and non brand mixed in one campaign. Brand searches convert cheaply and abundantly. When they share a campaign with cold traffic, the brand performance masks the non brand waste, and the algorithm spends into the blended average. Separating them almost always reveals that non brand is performing far worse than anyone believed.
Target ROAS set from a number nobody calculated. A target of 400 percent gets chosen because it sounds healthy, not because anyone worked out the contribution margin underneath it. The result is either a target so high the campaign cannot spend, or one so low it buys unprofitable revenue. Our guide to break-even ROAS covers the arithmetic that should sit behind that number.
Search terms nobody has read in months. The report exists, it is one click away, and on most accounts the last negative keyword was added a long time ago. This is the fastest money in any audit. See how to add negative keywords for the method.
A product feed left at whatever the platform exported. Titles truncated, no category attributes, missing GTINs, no custom labels for margin segmentation. For any Shopping or Performance Max account the feed is the campaign, and an unoptimised feed caps performance no matter how well the campaign is structured.
What an audit cannot tell you
An honest audit names its own limits. Looking at an account from the outside, there are things the data will not reveal, and pretending otherwise produces confident recommendations that turn out to be wrong.
You cannot see true profitability without margin data. ROAS is revenue divided by spend, and revenue is not profit. An account at 300 percent ROAS on 70 percent margin products is thriving; the same number on 25 percent margin products is losing money. Without the margin figures from the business, an auditor is guessing.
You cannot see the effect of anything that happened off platform. A stock outage, a price change, a competitor's promotion, a PR moment, a broken checkout. Sharp performance changes frequently have causes that live nowhere in the Google Ads interface, and an auditor who attributes every dip to campaign settings will misdiagnose.
You also cannot judge creative quality from performance data alone in a low volume account. If a set of ads has run 400 impressions, the difference between them is noise. Statistical patience is part of a competent audit, and calling a winner early is one of the more common ways audits do damage.
Doing it yourself versus paying someone
This checklist is genuinely runnable by anyone with access to the account and a few hours. Working through conversion tracking, search terms and structure will surface most of the obvious waste, and doing it yourself builds an understanding of the account that a delivered report never quite transfers.
The case for paying someone is pattern recognition rather than access. Someone who has audited a few hundred accounts recognises a structural problem in the shape of the data in a way that a checklist cannot encode. They also have no attachment to the decisions already made, which matters more than it sounds: auditing your own work honestly is difficult, and most people quietly justify the setup they built.
The practical middle path is to run the checklist yourself first, then get a second read on the two or three findings you are least sure about. That costs little and avoids paying someone to tell you things you already knew.
Frequently asked questions
How long does a Google Ads audit take?
A thorough audit on a mid sized e-commerce account takes three to six hours of focused work. Larger accounts with many campaigns and a complex feed take longer. Anything delivered in under an hour is a template with your numbers dropped in.
What access does an auditor need?
Read only access to the Google Ads account is enough for most of it. Read access to Google Analytics and Merchant Center makes the tracking and feed sections far more reliable. An auditor asking for admin access before agreeing scope is worth questioning.
Should I audit before or after changing agencies?
Before, if you can. An audit run while the incumbent is still in place gives you a factual basis for the conversation and tells you whether the problem is the manager or the account. Auditing after the handover mixes the previous manager's decisions with the transition itself.
Is a free audit worth anything?
It depends entirely on whether it is a genuine review or a sales document. A real free audit names specific things in your account and includes findings that are inconvenient for the person delivering it. If every finding leads to the same conclusion that you should hire them, it is a pitch.
How much of an audit can be automated?
The data gathering, largely. Scripts can pull search terms, flag disapproved products, surface budget constrained campaigns and check tracking coverage. The judgement of what matters and in what order is the part that is not automatable, and it is the part worth paying for.
How often to audit
A full audit twice a year is enough for a stable account, plus one whenever something changes materially: a new manager, a platform migration, a website replatform, or a sudden performance shift nobody can explain. Between full audits, a monthly pass on sections two and four catches most drift in about thirty minutes.
If you are an agency auditing accounts for clients rather than running your own, the same sequence works, and the first two sections are where you will find the credibility-building wins. We cover delivery in white label Google Ads for e-commerce agencies.
Where to start
If you only have an hour, do sections one and two. Verify that conversion tracking counts each sale exactly once, then pull the 90-day search terms report and filter to spend with no conversions. Those two checks find the majority of what a full audit would surface, and they require no changes to the account to complete.
If you want an independent read on your account, book a free audit call and we will run this checklist against it and send you what we find, whether or not you decide to work with us.
About the author
This guide is written by Vasant Chaudhary, a Google Ads specialist with more than five years of experience managing over 50 e-commerce accounts across the US, UK, and India. He focuses on Google Shopping, Performance Max, and product feed management. Get in touch or start with a free audit call.
