Guides / Advisory / 9 min read / Updated 1 August 2026

How to audit a Google Ads account in an hour

Most audits start with structure because structure is easy to have opinions about. Structure is the fourth thing. If the conversion data is wrong, every opinion after it is decoration.

Step one: is the conversion data real?

Before looking at a single campaign, open the conversions section and answer four questions.

• What is actually counted as a conversion? If page views, newsletter sign-ups and phone-number clicks are all in there alongside purchases, the account has been optimising towards a blend of things with wildly different values.
• Which of them are set to "primary"? Only the actions you genuinely want more of should be feeding bidding. Soft actions belong in reporting, not in the bid model.
• Is anything counted more than once per sale? Check the counting setting on each action.
• Do the totals reconcile with the back office? Take one settled day and compare. If they do not, that is the job, and everything else waits. The tracking guide covers the failure modes.

Step two: where did the money go?

Set the date range to the last 30 or 90 days and sort by cost, descending, at every level. Not by conversions, not by click-through rate. Cost.

1. Campaigns by cost. Usually a small number of them are most of the spend. Those are the only ones worth an hour today.
2. Search terms by cost. This is the highest-yield screen in the entire product. You are looking for money spent on queries that were never going to buy: research phrasing, job seekers, competitor names you did not intend, free-and-cheap modifiers, and anything about a product you do not sell.
3. Devices, locations and hours by cost. Money going to a country you do not ship to, or to hours when nobody answers the phone in a business that depends on answering the phone.

Step three: the settings that quietly cost money

• Search Partners and Display expansion on Search campaigns. Frequently on by default, frequently spending on placements nobody chose. Segment by network before deciding, then decide.
• Location targeting set to presence-or-interest. The default includes people merely showing interest in your area, which for a local service business means paying for the wrong country. Set it to presence for local businesses unless you have a specific reason.
• Broad match without a negative list. Broad match is not the problem people say it is when it is fed good conversion data and fenced with negatives. Without either, it is a budget disposal system.
• Auto-applied recommendations. Check whether they are on. Some are harmless, some rewrite your keywords and your ads, and all of them optimise for Google's definition of better.
• Ad rotation and bidding strategy mismatches, and any campaign still on a strategy that made sense at a conversion volume the account no longer has.

Step four: structure, finally

Now the argument people wanted to have at the start is worth having, because you know which campaigns matter and what the data underneath them is worth.

• Fragmentation. Twenty campaigns each with three conversions a month means no bidding strategy anywhere has enough signal. Consolidation is usually the fix, and it is usually resisted.
• Brand mixed in with everything else. Brand traffic converts at a rate that flatters whatever it is combined with. Separate it or you cannot see what non-brand is actually doing.
• Budget against opportunity. Campaigns limited by budget while others underspend is the easiest reallocation in the account.
• Overlapping keywords across campaigns competing with each other for the same query.

Step five: the ads themselves

Last, because they are the most fun and the least likely to be the binding constraint.

Check that every ad group has enough ads to test with, that headlines are not all saying the same thing in different words, that the assets are actually being served, and that the ad's promise matches the page it points at. That last one is where the leaks usually are — an ad promising free shipping pointing at a page that charges for it does not have an ad problem.

What not to change on day one

The temptation after an audit is to fix everything at once. Resist it, for a reason that has nothing to do with caution: if you change bidding, structure, budgets and ads in the same week, you will not be able to attribute the result to any of them, and you will have destroyed the account's history as a baseline.

Do the ones that are unambiguously wrong immediately — broken tracking, spend in the wrong country, obviously irrelevant search terms. Queue the rest and change them one at a time, with enough space between to read the result.

And write down what you changed and when. An account with no change log is an account where nobody can explain last month.