The Google Ads audit checklist I run on every new account

Every account I open gets the same first pass, in the same order. The order isn't cosmetic. Each check depends on the one before it, and the most common audit mistake I see is judging performance in an account where the measurement is broken. You end up optimizing a scoreboard that was lying the whole time.

This is the sequence. It's also, not coincidentally, how the free 3-Changes Audit gets produced: I run this list against your account with read-only access and send back the first three changes I'd make, in writing.

1. Verify the tracking before believing anything

Nothing else in the audit means anything until this check passes, so it goes first.

I trigger a test conversion myself and watch it land. Then I look for double counting, which usually comes from the same action being tracked twice, once by a Google tag and once through an imported analytics goal. An account that counts every lead twice looks brilliant and spends like it.

For call-based businesses, I listen to what actually counts as a conversion. If a wrong number or a 12-second hangup registers the same as a booked job, the account has been optimizing toward phone activity, not customers. The report looks better than the schedule does, and the gap between those two is where the budget went.

Auto-tagging gets checked here too, because without it the account can't connect clicks to outcomes later. This is the plumbing everything downstream drinks from.

2. Read the conversion hierarchy

Google lets you mark each conversion action as primary, meaning bidding optimizes toward it, or secondary, meaning it's observed but not chased. Most accounts I open have never made that distinction deliberately.

The failure mode is predictable. A newsletter signup, a PDF download, and a booked appointment all sit as primary actions, so the system treats them as equally valuable and drifts toward whichever is cheapest to produce. That's almost never the one that pays the bills.

The check is simple to describe: does the list of primary actions match the list of things that make this business money, with values that reflect the difference between them? A consult request and a five-figure job should not carry the same weight. When they do, the budget quietly follows the cheap one.

3. Look for the account competing with itself

Structure comes third, and I'm looking for two diseases.

The first is overlap: multiple campaigns or ad groups eligible for the same searches, which means the account bids against itself and pays the markup. The search terms data shows this quickly, the same query being served from two places at two different costs.

The second is fragmentation. Accounts accumulate campaigns the way garages accumulate boxes, and every split divides the conversion data into thinner slices. Automated bidding learns from conversions per campaign, so ten campaigns sharing the data that two campaigns would concentrate means everything learns slowly and nothing learns well. Consolidation is usually the recommendation, and it's usually resisted, because the fragmented structure represents someone's past decisions. The data doesn't care.

4. Pull the map

Geography is where I find the quietest waste.

Google's location options include a default that can show ads to people who were merely interested in your area, not physically in it. Their location targeting documentation covers the settings. For a service business with trucks and drive time, that default can mean paying for clicks from people your crew will never reach.

So the check is: pull the geographic report and compare where the money went against where the business actually serves. Then look at the radius itself. A service area drawn as a neat circle around the shop rarely matches real drive times, and the zip codes at the expensive edge of that circle are often the first thing I'd cut.

5. Follow the budget

Last, because it only makes sense once the first four are known: where does the money actually go?

I break spend down three ways. By search term, to see how much goes to queries that have never once converted. By brand versus non-brand, because branded clicks from people already searching your name can flatter the account's averages and hide what acquisition really costs. And by campaign against its results, to find the mismatches: campaigns that are budget-capped while producing, and campaigns spending freely while producing nothing.

The pattern I find most often is that spend follows history rather than performance. Budgets were set at some point, for some reason, and nobody has moved them since. Reallocating what's already being spent is frequently the biggest available win, and it costs nothing.

What happens with the findings

Five checks, one afternoon, no changes made to anything while I look. The output is a short list, ordered by impact: what's broken, what it's costing, what I'd change first.

If you want this run against your own account, that's the free 3-Changes Audit. You grant read-only viewer access, and within 3 business days you get the first 3 changes I'd make, in writing. Whether you hire me to make them, through paid ads management or on your own, is a separate decision, and the list is yours either way.