Google Ads Account Audit: Fix the Leaks Before More Budget

If you are about to raise budgets in one of your Google Ads accounts, stop for a moment. My working hypothesis is that most scale-ups fail because the account leaks before spend increases. That is not a proven finding; it is why I audit first and scale second. I do not start with bid strategy. I start with account hygiene, because a higher budget poured into broken tracking, wasted search terms, poor placements, thin exclusion lists, capped campaigns, or mismatched landing pages simply buys more of the same problem. In this tutorial I will walk through the exact sequence I run before any budget increase. We will cover why I audit before touching the budget, the audit sequence itself, then six checks: tracking integrity inside Google Ads accounts, search term and query waste, placements, networks and formats, negative keyword and exclusion lists, budget caps, pacing and the bid strategy reading them, and landing page and message match. I will also show where automation hides leaks, what to fix first when several checks fail, and why the audit comes before the budget, every time.

Why I Audit the Account Before I Touch the Budget

I treat this as an argument, not a finding: most scale-up failures I see are not bid strategy problems. They are accounts that were already leaking before the spend went up. I have no benchmark to cite for that, so I use it as my reason to audit first rather than as a statistic.

A leaking account scales the leak. Double the budget and you double the wasted search terms, the junk placements and the broken conversion data. Cost per acquisition gets worse while the spend line looks healthy, which is the worst combination because nobody panics.

Bid strategy changes get all the attention because they are visible in the Google Ads manager. Structure, exclusions and measurement are what decide whether the extra money actually lands well inside a Google Ads account. That is why How Google's Smart Bidding Changes Should Shift Your Budget Strategy matters, but only once the account itself can be trusted.

A pre-scale audit is cheaper than a post-scale clean up. If you add budget first, you pay for the waste the whole time you are diagnosing it, and you lose the clean before and after comparison that proves the increase worked.

What this piece is not: it is not a bid strategy guide and I am not promising a percentage uplift. It is the hygiene pass I run before any pound of new budget is added. So before the money moves, I work through an ordered set of checks.

The Audit Sequence I Run Before Any Budget Increase

The order is the method. I run these checks in strict sequence, and I do not skip ahead because a later stage looks more urgent.

Tracking integrity comes first. If conversions are wrong, every decision after that sits on bad numbers, and a bigger budget only makes the error more expensive. If this check fails, I stop there. Auditing search terms on top of broken conversion data tells me nothing I can act on.

Then demand-side waste: the search terms report inside the Google Ads account, placements, networks and formats, plus the negative keyword and exclusion lists that stop the same waste returning next month. A one-off clean without a list is not a fix, it is a subscription.

Then the money plumbing: campaign and shared budgets, pacing, bid strategy, ad schedule and geo settings. A capped campaign will not spend the new budget even when it deserves to, and no extra funding changes that.

Last, the receiving end: final URLs, landing page match and mobile load. Cleaning the traffic and sending it to a page that cannot convert just moves the loss downstream.

I keep this sequence separate from strategy debates. My notes on Google Ads in 2026 cover what is working right now; this is the hygiene pass that makes that spending measurable.

I only raise budget when every check passes, or when a failing check has a dated fix plan and a named owner. A passed check is a decision I can defend later, not a feeling.

Check 1: Tracking Integrity Inside the Google Ads Account

I start every audit with the number every other decision leans on. If tracking is wrong, a bigger budget buys a bigger lie.

I verify that each conversion action fires once per real conversion. Duplicate counting is one of the most common faults I find in an established Google Ads account: a tag manager rule plus a form plugin plus a thank you page that reloads on submission reports two or three conversions for one sale. Every cost per acquisition then looks better than it is.

I test with a real conversion path on my phone, in an incognito window, once with consent accepted and once with consent declined. Preview mode and test tags hide exactly the failures that only appear in a live session.

I check the conversion window and the count setting against the actual sales cycle. The 30 day default misreads a 90 day deal, and counting every conversion rather than one inflates a lead gen account that gets repeat form fills.

I reconcile Google Ads conversions against the system that owns the revenue, whether that is Shopify, Stripe, HubSpot or a CRM. If the platform says 40 sales and the account says 65, I do not raise budget until I know which one is lying. That reconciliation is the unglamorous half of why attribution breaks for most advertisers.

I also hunt for conversion actions switched off for bidding but still feeding reports, and for goals that count a page view, a scroll or a button click as a conversion. Those two do not throw an error. They just quietly distort every decision you make afterwards.

Only once the numbers hold do I move on to where the money is actually going.

Check 2: Search Term and Query Waste

Once tracking is clean, I move to where the money actually leaves the account. I read the search terms report at account level, then per campaign, filtering for terms with meaningful spend and no conversion over the window that matters to the business, not the default. Google's help page on the search terms report is my only citation, and only for where that data lives. It says nothing about how much waste to expect, so I benchmark each account against its own history.

I sort by cost, never impressions. A term with 4,000 impressions and 12 clicks is cheap. One term quietly eating 15 percent of a campaign budget is the problem, and impressions flatter it out of view.

Then I split what I find into three buckets, because each needs a different fix. Irrelevant intent gets broad negatives that block a theme. Competitor or brand confusion gets careful phrase and exact negatives, since I may still want some of that traffic. Job seeker or freebie traffic, the free templates and salary queries, gets its own list so it never mixes with the other two.

I look hard for close variants and broad match expansion. Queries that never matched my keywords literally are where budgets disappear, and the list grows every week without anyone adding a keyword. My 2026 read on Google Ads in 2026: What the Data Actually Means for Paid Search covers why that drift matters more now.

I add the worst offenders as negatives the same day, then write down the pattern rather than the term. Patterns belong in a shared list applied across the account, because single keywords only ever protect one campaign, and that is how the leak returns.

Next I check where those clicks actually landed.

Check 3: Placements, Networks and Formats

Search terms tell me who the money chased. Placements tell me where it landed.

I open the placements report and look at where Display and Video spend actually went. Mobile game apps, parked domains and Made for Advertising sites are normal in a Google Ads account that has never been cleaned, and they can absorb budget for months before anyone notices.

My rule is simple: I turn off what I cannot verify. If a placement shows no conversion in the window I use and I cannot explain why it is there, it does not get a second month of budget. I do not need to prove it is worthless. The absence of a reason to keep it is enough.

Network settings come next. Search Partners and Display are easy to leave switched on, and they widen reach well beyond what the original plan assumed. Reach is not the same thing as demand. I check them campaign by campaign, because the account level view hides where the difference actually sits.

When the pattern is structural I exclude whole categories rather than blocking sites one by one. Chasing individual placements is a job that never ends and never keeps up with inventory growth. I apply the same logic when I reverse-engineer competitor strategy from the Meta Ads Library: patterns beat one-offs.

I check audience and exclusion overlap too. Two similar audiences stacked in one ad group compete for the same impression, and the narrower one can starve. The report then looks like a bidding problem when it is a targeting problem.

Every category I cut here needs somewhere to live, so the exclusion lists are the next check.

Check 4: Negative Keyword and Exclusion Lists

Once I have stopped the worst placements bleeding money, I turn to the lists that stop the same waste coming back. One off negatives rot. Every term I add manually to a single campaign protects that campaign and nothing else, so the next launch starts from zero. I move recurring waste into shared negative keyword lists applied across the Google Ads account, so a new campaign inherits the hygiene rather than relearning it.

Removing a term from a shared list affects every campaign attached to it, so anything that should only be blocked in one place moves to campaign level instead.

I keep brand, competitor and generic lists separate. Drop a broad brand term into a shared generic list and you can cut legitimate traffic from a launch campaign months later without ever connecting the two events.

Then I audit for negative conflicts: negatives that block keywords I am actively bidding on. Google will let the two coexist quietly, and you lose traffic with no error message. I check every level in the Google Ads manager, campaign, ad group and shared list, before I trust any keyword's performance.

Match types get chosen deliberately. Broad negatives block over a theme, phrase negatives trim around an intent, exact negatives remove a single query surgically. The wrong one either leaks waste through or over trims good traffic.

Finally, I put a review cadence on the lists so they grow with the account. A list written eighteen months ago becomes a stale lock that blocks new opportunities, especially after a product or pricing change. I review quarterly at minimum, and I date every addition.

With exclusions clean, the next check is the money plumbing: budget caps, pacing and the bid strategy reading them.

Check 5: Budget Caps, Pacing and the Bid Strategy Reading Them

The lists only hold the line if the money behind them is set right, so my next check is the plumbing.

Lost to budget or lost to rank. A budget increase on a campaign already limited by budget does not always spend more. Before I add a pound, I compare lost impression share due to budget against lost impression share due to rank. Those two point at completely different fixes: one says the cap is the constraint, the other says bids or quality are.

Shared budgets and pacing. I look for shared budgets that starve one campaign to feed another, because the winner never gets full funding while the loser keeps spending. I also watch daily caps throttling the last hours of the day, which is when I see the strongest converting traffic arrive in several accounts. If evening traffic converts and the cap is gone by 3pm, that is a pacing fix, not a budget one.

Bid strategy on thin data. I confirm the strategy has enough conversion volume to work. Moving to Target CPA or Maximise Conversions on thin data can spend more for worse returns while the account blames the budget. Google is relabelling Maximise Conversions with a Target CPA as Target CPA, and nothing about the bidding changes.

The boring settings. Campaign end dates, ad schedules and geo targets. A campaign that stops at 5pm, or excludes a region that used to be profitable, will never use the new money however high I set the cap.

What the extra pound is for. More volume on a proven term and a deliberate test are different asks with different targets, so I set the target before the spend. If that spend carries a seasonal push, the same discipline applies to How to Run a January Sale Campaign That Actually Converts.

Check 6: Landing Page and Message Match

Budget mechanics clean up the supply side of the account. The demand still has to land somewhere, so the last check is the page I am actually paying for.

I open every final URL in the account and check three things: it resolves, it loads quickly on a 4G phone, and it matches the promise in the ad. A mismatch between query, ad and page is a tax I pay on every click.

The converting element has to sit above the fold on mobile, because most paid clicks in a Google Ads account arrive on a phone. Desktop performance tells me nothing about what happened on that device. To make that concrete: a page that repeats the ad headline but buries the form three scrolls down on mobile fails this check, however strong the ad is.

For my top spending terms I check query to page match specifically. The headline, the ad copy and the search should use the same words, or the visitor does translation work before they decide. This is where Google Ads and CRO: Why the Click Is Only Half the Equation becomes relevant: the click is only half the job.

I also confirm tracking fires on the landing page path I am actually sending traffic to, not on a legacy version someone forgot to update. I have seen an entire conversion action stranded on an old URL.

I fix the page before I raise budget on it. Sending double the traffic to a page that converts at half the rate just scales the loss with better looking reports.

Where Automation Makes the Leaks Harder to See

Landing pages are the last check I can make by eye. Automation is where my eyesight gets worse, and I want to be clear that what follows is my argument rather than documented fact.

The automated formats now sitting in the Google Ads manager, Performance Max, AI Max for Search, AI Max for Shopping and Smart, concentrate control in the system and thin out the visibility I rely on when I audit by hand. AI Max for Search is a current development rather than a legacy setting, covered in a February 2026 interview on the shift to AI Max, so I do not expect this to quietly age out of the problem.

Search terms and placements still exist inside those campaigns. They surface later, aggregated, and with far less control than a standard search campaign gives me. So instead of one clean table I lean on asset group, audience and segment level reporting, and I accept that I am reading a summary rather than a raw list.

That delay has a price. I export search term and placement data from automated campaigns on a fixed schedule, weekly at minimum, and I treat the export as the audit rather than a nice to have. Waiting for the interface to hand me the waste means I pay for that waste every week I wait.

Per campaign controls are limited in automated formats, so account level negatives and brand exclusions become the guardrail. Shared negative lists applied across the Google Ads account are the only reliable place to hold the line, because a new automated campaign inherits them without anyone remembering to configure it.

I also judge automation on incremental conversions against a holdout. Platform reported conversions can include traffic I was already paying for somewhere else in the account, which flatters the campaign and hides the overlap.

What to Fix First When Several Checks Fail

When automation is hiding the waste, fix order matters more, not less. I work by leverage, not by what is quickest to click.

Tracking first

Tracking comes first, because a broken conversion number invalidates every decision I make with the new budget. The measurement landscape is fragmenting and buyers increasingly want incrementality proved rather than reach reported (IAB/PwC Internet Ad Revenue Report), so a trustworthy conversion count is the foundation. If Shopify says 40 sales and the account says 65, I stop there.

Stop the bleeding

The largest wasted search terms and placements by cost get fixed the same day. That money is already leaving, needs no analysis, and the fix is one negative or exclusion.

Structural hygiene

Then shared negatives, network settings and audience exclusions. These stop the leak recurring instead of plugging one hole at a time.

Money plumbing

Budgets, pacing and bid strategy come after demand is clean. Raising a cap on dirty demand just buys more of the same.

Landing pages

Landing page work comes last but lands before the increase. It ships slowly, so I start it in parallel and refuse to let it hold a clean account hostage. Most shoppers abandon before buying, and the page fixes that recover them compound across the funnel (The Sales Mods That Actually Move Revenue). Throughout, I log which checks passed, which failed and the date of each fix, so the increase is a decision with evidence behind it.

The Audit Comes Before the Budget, Every Time

Work the checks in that order: tracking integrity, search term waste, placements and networks, exclusion lists, then budget mechanics and landing page match. Fixing each one before the spend goes up is cheap. Fixing it after means you pay for the leak for the entire time you are diagnosing it.

So I never raise budget to solve a leak. Fix what fails, date the fix, then add the pound to a clean Google Ads account. That is the only version of the increase I can actually measure, because I know what the account looked like before the money arrived.

I keep a written record of every passing check, with dates, so the next budget conversation rests on evidence rather than a hunch. I also put the whole audit on a repeating cadence, monthly or quarterly depending on spend, rather than running it once as a project. A clean account starts drifting the following week.

Expect that drift. Match types expand, placements accumulate, new campaigns inherit settings nobody re-reads, and landing pages change without telling the ads team. A single automated campaign can reach Search, Display, YouTube, Discover, Gmail and Maps at once, which is why I verify where money landed rather than trusting the setup I built (Google's Performance Max documentation).

Run the audit as the last gate before every budget decision, and the increase stops being a gamble.

Conclusion

An audit before a budget increase is not bureaucracy. It is the cheapest insurance the account can buy. Work the checks in order: tracking integrity, search term waste, placements and networks, exclusion lists, then budget mechanics and landing page match. Fix what fails, date the fix, and record every passing check.

My next step with any client account is the same one I take every month. I run the full audit as the final gate before I touch the budget, then put it on a repeating cadence so drift gets caught early. Search, Display, YouTube, Discover, Gmail and Maps all pull spend in different directions, and only a verified account shows where the money actually landed.

Do the audit first. Then raise the budget, and let the evidence carry the decision.