How to audit your own Google Ads account

When results slip, most advertisers reach for one of two explanations: the budget is too small, or the market is too competitive. Sometimes that is the real story. More often the account has three or four fixable problems sitting in plain sight, and nobody has opened the hood in months.

You do not need an agency to find them. You need about an hour and a willingness to click into the reports most people skip. Below is the walk-through we run when we take over an account, in the order we actually do it. Nothing here requires a tool you do not already have.

1. Read the search terms report before anything else

There are two reports that sound the same and are not. The keywords report shows the words you told Google to bid on. The search terms report shows what people actually typed to make your ad appear. On broad and phrase match, those two lists drift apart quickly, and the gap is where money leaks.

Open Insights and reports, then Search terms, and set the date range to the last 30 to 90 days. Read down the list and ask one question of each row: would I have paid for that click on purpose? A plumber will find queries like "how to fix a leaky faucet" and "plumber salary." A law firm will find "free legal aid" and the name of a competitor's cheaper service. None of those people are about to become customers, and every one of them cost you a click.

Add the obvious offenders as negative keywords right there. If you see the same junk theme over and over, add the root word as a negative so you stop paying for the whole family of it. This one report, cleaned up monthly, does more for a wasteful account than almost anything else.

2. Check what your conversions are actually counting

A conversion number is only as honest as the thing behind it. Before you trust a single performance decision, find out what the account is calling a conversion. Go to Goals, then Conversions, and look at each conversion action and how it counts.

Here is what tends to be wrong. Some accounts count every phone call over 15 seconds, so a wrong number and a robocall both register as leads. Some count a form submission and a call from the same person as two conversions, which quietly doubles the reported results. Some are set to count "every" conversion per click instead of "one," which inflates e-commerce-style tracking on a business that books one job per customer.

The fastest sanity check is to compare the platform's conversion count for last month against what your CRM, booking software, or call log says actually happened. If Google reports 140 conversions and you booked 38 jobs, the algorithm has been chasing a number that has little to do with your revenue. That gap is worth more attention than any bid setting.

3. Find where the money actually goes

Averages hide problems. An account can show a decent overall cost per lead while one campaign quietly eats 40 percent of the budget and produces almost nothing. Sort your campaigns by cost, highest first, then look across at conversions and cost per conversion for the same rows.

You are looking for the campaign or ad group that spends a lot and returns little, and for the opposite: the small line that converts well and is starved for budget. Both are common, and both are easy to fix once you can see them. Do the same one level down, by keyword, inside your biggest-spending campaign. It is normal to find that a handful of keywords carry the results and a long tail of others just burns money.

4. Look at your match types and how broad the account runs

Broad match can work, but only with strong conversion data and tight negatives guiding it. Without those guardrails it wanders, and you end up paying for loosely related searches. Check how much of your spend sits on broad match keywords, and cross-reference that against what you found in the search terms report. If broad match is spending freely and the search terms list is full of junk, the two problems are connected.

This is also the moment to check whether the account is leaning on Performance Max or broad match to "find new customers" without anyone watching what it finds. Automation is fine. Automation with no one reading the search terms behind it is how budgets disappear.

5. Check the boring settings that waste money silently

A few settings cost people money for months because no one thinks to look at them.

• Location targeting set to "presence or interest" instead of "presence." The default can show your ads to people who are merely searching about your area from three states away. For a local business, "presence" is almost always what you want.
• Search Partners or the Display Network left on inside a Search campaign. These often deliver cheap, low-intent clicks that flatter your volume and rarely book.
• Ad schedule and device with no adjustments, when your booking data clearly shows some hours and devices convert far better than others.

None of these are dramatic. Together they are often 10 to 20 percent of a budget spent on impressions that were never going to turn into work.

6. Read three ads and one landing page like a customer

Close the reports and look at the account the way a person searching would. Pull up your top ad and ask whether it says something specific and true, or whether it reads like every other ad on the page. Then click through to the landing page on your phone. Does it load fast, say the same thing the ad promised, and make it obvious how to call or book? A surprising amount of "our ads do not work" is really "our ads work and our page loses the lead."

When doing this by hand gets old

The checks above are not hard. They are tedious, and they get more tedious the larger the account and the more platforms you run. Doing this properly across Google, Microsoft, and Meta every month is a real job, which is exactly why most owners do it once, feel better, and never come back to it.

That is the reason we built TrailMap. You export your report from the ad platform, upload it, and TrailMap runs this style of audit across the whole account in minutes, then writes back a plain-English action plan sorted by what will move the needle first. The math is all computed, not guessed, and the recommendations read like a senior analyst wrote them, because the tool was built from how our analysts actually work. It will not replace an hour of paying attention, but it turns that hour into a few minutes and catches the patterns a human eye skims past.

Whether you run the audit yourself or hand it to software, the point is the same: an ad account is not a set-and-forget machine. The ones that perform are the ones somebody actually reads.