How to audit a Google Ads account that is already spending

The first thing I do when I inherit a Google Ads account is not look at performance. It is check whether the performance is real. In the accounts I have taken over, the single most common finding is not a bad keyword or a lazy ad. It is that the reported conversion number is wrong, usually in the advertiser’s favor, and every decision made in that account for the past year was made on a number that did not exist. Audit in the wrong order and you spend a week optimizing toward a fantasy.

So the order below is not a preference. It is a dependency chain. Each step is only meaningful if the step above it checked out.

Why tracking has to come first

Every automated bidding strategy in Google Ads optimizes toward the conversions you report. Target CPA, Target ROAS, Maximize conversions, the bidding inside Performance Max and Demand Gen, all of it. The system does not know what a real customer is. It knows what you told it a conversion is, and it will buy more of exactly that.

So a tracking error is not a reporting error. It is a spending error, compounded daily. If the conversion tag fires on a page refresh, the algorithm learns to buy traffic from people who refresh pages. If a newsletter signup is primary alongside a purchase, the algorithm chases the cheap one, because that is how it hits its target. If half your European traffic has no consent signal, those conversions never make it back and the algorithm bids down the segments that were working. Nothing else in the audit means anything until you know which numbers you can trust.

What to check, specifically

Which conversion actions are marked primary. Open the conversions view and sort by whether each action is included in the Account default goal. I regularly find five or six primary actions, including things like phone number clicks, PDF downloads and time on page. Anything that is not the actual money event should be secondary.
Duplicate counting. Look for a conversion action defined both in Google Ads and imported from Analytics, both counted. This inflates conversions by roughly double and is very common in accounts that changed hands.
The count setting. One conversion per click for lead generation, every conversion for ecommerce. A lead form set to every will count the same person filling the form three times as three leads.
Conversion values. If there are no values, the account cannot use value based bidding at all. If every lead is given the same arbitrary value, that is fine as a starting point but it means Target ROAS is really just Target CPA in disguise.
Consent signals. For any account with European traffic, confirm Consent Mode v2 is implemented and the advertising consent parameters are actually being passed, not just that a cookie banner exists. A banner with no consent signal wired to the tag gives you the compliance overhead and none of the modeling benefit. Without it you also lose remarketing audiences and enhanced conversions for that traffic.
Enhanced conversions. Check whether it is on and reporting a healthy match rate. If it was switched on last week, remember it takes weeks before the effect shows up, so do not read early numbers as failure.
Attribution and lookback windows. Note them, because they explain a lot of apparent performance changes that have nothing to do with the campaigns.

Then do the thing nobody wants to do. Compare a month of reported Google Ads conversions against the actual leads or orders in the client’s CRM or inbox. Not the total. The ones attributed to paid search. If the two numbers are within a reasonable margin, you can proceed. If Google reports two hundred leads and the sales team remembers forty, stop the audit and fix the tracking, because everything below is built on sand.

Step two: search terms and where the money actually went

Once you trust the numbers, find the waste. This is the fastest source of real savings in almost every account I have opened.

Pull the search terms report for the last ninety days, sorted by cost, and read it. Not skim it. Read it. You are looking for three things: wrong intent, things you do not sell, and people looking for free or DIY versions of your service. In most accounts a meaningful share of spend sits in terms the owner would reject on sight. Where the waste comes from is usually predictable:

Broad match with no negative list. Broad match plus Smart Bidding can be good, but only with a mature conversion signal and a maintained negative list. Broad match on a thin account is an expensive way to discover your category adjacencies.
Search partners and display expansion left on. These are on by default in several setup paths. Segment by network and look at cost and conversions per network. In most accounts I check, the partner and display portion converts far worse than search itself, and turning it off costs nothing in real volume.
Location targeting set to presence or interest. The default includes people merely interested in your location. A plumber in Varna does not need clicks from people researching Varna from another continent. Set it to presence.
Performance Max soaking up brand traffic. If a Performance Max campaign is running alongside search, check whether it is quietly serving on your own brand terms and taking credit for conversions your brand campaign would have won at a fraction of the cost. Since campaign level negative keywords and channel level reporting became available, you can now see and control this properly. Use it. Also check the channel breakdown to see how much budget is going to Display and Discover inventory rather than Search and Shopping.
Automatically applied recommendations. Check whether these are enabled. Some of them, particularly automatic keyword expansion and broad match upgrades, will change the account without asking. I turn most of them off.

Quantify the waste as a percentage of total spend and write it down. It is the clearest thing you will show the client.

Step three: structure problems

Structure is only worth fixing where it prevents you from controlling budget or reading results. Do not restructure for elegance.

The problems that genuinely matter:

Brand and non brand mixed together. This is the most damaging structural fault, because brand search converts cheaply and hides the real cost of everything else. If they share a campaign, the account looks profitable when the acquisition half is not. Separate them before you judge anything.
Fragmentation. Twenty campaigns each getting two conversions a month. Every bidding strategy is stuck in learning forever and none of them can optimize. Consolidate. Modern bidding wants conversion volume concentrated, not sliced.
The opposite problem. One campaign holding three unrelated services with one shared budget, so the highest volume service eats everything and the highest margin one never serves. Split only where you need separate budget control.
Ad group bloat. Fifty keywords in one ad group with one set of ads that cannot be relevant to all of them. Group by intent and shared meaning.
Geography and language mixed. Countries with very different costs and conversion rates sharing one budget will always be misallocated.

Also check the account level basics that people forget: shared negative lists applied, conversion goals set at the right level, ad schedules, and whether anything is paused that should not be.

Step four: bidding and budget misallocation

Now look at where money is going relative to where it works. Sort campaigns by spend, then by cost per acquisition. The pattern I see most often is that the biggest spender is not the best performer, and nobody has moved money because nobody had a trustworthy number to move it on.

Things to look at, in order:

Impression share lost to budget. If a campaign is hitting its target cost per acquisition and losing significant impression share to budget, that is free money sitting on the table. Increase the budget before you touch anything else.
Impression share lost to rank on your best converting terms. That is a bid, quality or relevance problem, and it usually means the ad or the landing page is weaker than the competition rather than the bid being too low.
Targets that were set once and never revisited. A Target CPA set two years ago at a number the account has beaten every month since is throttling volume. A target set far below what the account has ever achieved will strangle delivery until the campaign barely serves. Move targets in steps of ten to fifteen percent, not in jumps.
Budget limited campaigns using target based bidding. Since Google changed this behavior in 2026, budget limited Target CPA and Target ROAS campaigns now spread spend across a wider set of auctions and report closer to the set target, rather than appearing to beat it. If someone in this account was reading a Target CPA campaign that seemed to be outperforming its target, that reading is out of date. Check whether reported cost per acquisition moved recently for no obvious campaign reason.
Strategies applied to campaigns with no data. Target ROAS on a campaign with four conversions a month is not a bidding strategy, it is a random number generator. Move it back to maximize conversions until volume supports a target.

Step five: ads and landing pages

By now you know what is true, what is wasted and where money should go. Only now is it worth judging creative, because a good ad on a broken account still loses.

On ads, check that each ad group has responsive search ads with genuinely distinct headlines rather than ten paraphrases of one sentence. Check pinning. Heavy pinning is the most common self inflicted wound I see, because it removes the format’s only real mechanism. Treat ad strength as a rough signal, not a target. Then check whether sitelinks, callouts and structured snippets exist at all. Missing assets are one of the cheapest fixes available.

On landing pages, click every ad yourself on a phone. Look for a mismatch between the promise in the ad and the first line on the page, and for homepage traffic that should go to a service page. Time the load. Count the form fields. Check the phone number is tappable. A slow, generic page with a nine field form caps the account no matter how well the campaigns are managed, and no amount of bid tuning fixes it.

Fixable account, or rebuild?

This is the judgment call the client actually needs from you, and most audits avoid making it. My rough rule:

Fixable if the tracking is either correct or correctable, the account has a meaningful history of real conversions, and the problems are waste, misallocation and neglect. History has value. A campaign with two years of clean conversion data is an asset you should not throw away, because rebuilding resets the learning and you will pay for that in a worse first month.

Rebuild if any of the following are true: the conversion data is fundamentally wrong and always has been, so the entire optimization history was trained on false signals; the structure mixes brand, geography and services so thoroughly that you cannot separate them without touching everything anyway; or the account was built for a business model, product range or market the client no longer has.

When you do rebuild, do it in parallel rather than by deletion. Build the new campaigns, run them alongside at a small budget until they are producing, then shift the money over. And keep the old account, do not delete anything, because the historical data is still the best benchmark you have for whether the rebuild worked.

The audit checklist, in order

List every conversion action and confirm which are primary. Demote everything that is not the money event.
Look for duplicate conversion actions counted twice.
Check count settings: one per click for leads, every for ecommerce.
Check conversion values and whether value based bidding is even possible.
Confirm Consent Mode v2 is implemented and the advertising consent signals actually fire.
Check enhanced conversions status and match rate.
Reconcile a month of reported conversions against the client’s CRM or inbox. Stop here if they do not match.
Read ninety days of search terms sorted by cost. Quantify the waste.
Segment by network. Turn off search partners and display expansion if they underperform.
Set location targeting to presence only.
Check whether Performance Max is absorbing brand traffic. Review its channel breakdown and add campaign level negatives.
Turn off automatically applied recommendations you did not choose.
Separate brand from non brand if they are mixed.
Consolidate fragmented campaigns that cannot gather enough conversions to learn.
Split campaigns where one service or country is eating a shared budget.
Review impression share lost to budget and to rank, campaign by campaign.
Review every bid target against what the account has actually achieved in the last ninety days.
Move campaigns with too little volume off target based strategies.
Check responsive search ads for distinct headlines, minimal pinning, and complete assets.
Click every ad on a phone and grade the landing page for match, speed, form length and clarity.
Write one page: what is wasted, what is misallocated, what is broken, and fix or rebuild.

What to do next

If you are reviewing your own account today, do steps one to seven and nothing else. Do not touch a bid, do not pause a keyword, do not rewrite an ad. Just find out whether the conversion number in your dashboard matches the number of real customers you got. That single check tells you more than the rest of the audit combined, and it takes an afternoon.

If the numbers do not reconcile, fix tracking before you change anything else, and accept that you are starting the account’s learning over. That is not a loss. You were never getting the performance the dashboard claimed anyway. If you want someone to run the reconciliation with you and tell you honestly whether the account is worth fixing or rebuilding, you can book a free consultation.