How to Audit a Google Ads Account Before Hiring an Agency

Before you open the account: set up the review properly

Three practical things first. One, make sure you have Admin or Standard access to the account itself — if your current or previous agency owns the login, that's already a finding, and it needs resolving before anything else (account ownership is covered again in step twelve). Two, pull the last 90 days as your review window; 30 days is too noisy and 12 months hides what changed recently. Three, export or screenshot as you go. The written record of what you found is what becomes the brief later.

One honest caveat: a self-audit won't catch everything an experienced outsider catches, because you're reviewing assumptions you or someone you hired may have made. It will catch most of the expensive stuff, which is what the hiring decision actually turns on.

Step 1: Conversion tracking — because every later check depends on it

Start here, always. If tracking is wrong, the numbers you're about to review in every other step are wrong with them, and an agency pitching on your dashboard numbers inherits the same blindness.

Open Tools → Conversions and check four things. Does every conversion action map to a real business outcome — a submitted form, a phone call, a purchase — rather than a page view or a button click? Is the count setting on "One" for lead actions ("Every" double-counts refreshes and repeat submits, and is the single most common source of inflated numbers I find in audits)? Do the tags actually fire on the real thank-you or confirmation event, verified in Tag Assistant or GTM preview rather than assumed? And do platform-reported conversions roughly reconcile with what your inbox or CRM actually received — within about 10%?

If tracking fails these checks, stop and fix it or hire someone to fix it before anything else. Every agency proposal you evaluate afterwards should start by verifying tracking too; if a pitch quotes your current cost per lead without asking whether the tracking is trustworthy, that tells you how they'll run the account.

Step 2: Campaign structure — can a stranger understand it in five minutes?

Open the campaign list and ask one question: could someone who knows PPC but nothing about your business explain why each campaign exists within five minutes? Structure should map to distinct service lines, objectives or geographies. A campaign per service (or per service per priority city, for local businesses), brand searches separated from everything else, and nothing that exists only because it was created one Tuesday in 2023 and nobody dared delete it.

Specific things to look for: brand and non-brand mixed in one campaign, which makes overall numbers look better than the non-brand reality; one giant campaign holding every keyword the business has ever bid on; and Display or Search Partners traffic switched on by default rather than chosen. Each of these makes Smart Bidding learn slower and makes the account harder for whoever you hire next to assess quickly.

What it means for hiring: structural problems are a rebuild, not a tune-up. That's the kind of finding worth paying for once as a project, before signing any monthly retainer.

Step 3: Keywords and match types

Look at the keyword report over the 90-day window and sort by spend. Three checks. First, are the top spenders phrases a genuine buyer would type — "emergency plumber austin" — or generic category words — "plumbing" — that a small budget can never compete on? Second, how much sits in keywords with zero conversions over the window? A keyword that has spent steadily for 90 days without producing a single conversion is spending your budget to prove it doesn't work. Third, what's the match type spread? Broad match with no conversion history and no negative list is the most expensive way to learn what people type.

Match type should match the account's data volume: exact and phrase for accounts under a few hundred conversions a month, broad only where there's enough conversion data for Smart Bidding to steer it.

What it means for hiring: a keyword list full of zero-conversion spenders is a one-day fix — pause, add negatives, retighten. If that's all you find, you may not need an agency at all yet.

Step 4: Search terms and negative keywords

The search terms report is where wasted spend hides most visibly. Open it for the full 90 days, sort by cost, and apply one test: would you deliberately buy this search term again today, knowing what you know? Add up the spend on everything that fails the test. Above roughly 15% of total spend, that's a clear, quotable finding — and it's almost always fixable in a day with negative keyword additions.

Then check the negative lists themselves. Are there any? Were they built once at launch and never touched, or maintained weekly? A negative keyword list frozen at launch is one of the clearest signs an account has been managed on autopilot — it's the first thing I check when someone tells me their previous agency "didn't really do anything".

What it means for hiring: whoever manages the account next should be doing this weekly. Ask every candidate agency directly: who reviews search terms, how often, and can you show me an example of negatives added for another client? A vague answer here predicts autopilot.

Step 5: Location targeting

Open the location settings on each campaign and check the option Google defaults to "Presence or interest". That default happily serves your ads to people in other countries searching for information about your area — a well-known source of spend that never converts, particularly on US campaigns targeted at specific cities or states. Switch it to "Presence" and watch the phantom traffic disappear.

Then check the actual locations report: how much spend has gone to places you don't serve? For local service businesses this check alone regularly surfaces 10–20% of spend buying clicks from searchers who could never become customers.

What it means for hiring: this is a ten-minute fix you can do yourself today. But if it's been wrong for a year under paid management, it's also evidence about how the account was being looked after — worth weighing in who you hire next.

Step 6: Budget allocation — where the money actually goes

Two checks here. First, is budget allocated by evidence or by habit? Sort campaigns by spend and compare against which campaigns produce leads. Even spreads across campaigns "to keep things fair" are a habit, not a strategy — budget should flow toward what produces enquiries, not be spread evenly so no campaign looks neglected.

Second, look for campaigns hitting their daily budget caps early in the day (the "Limited by budget" status, or Lost impression share due to budget). A campaign capped by midday is a campaign that's turned off during your best evening and overnight searches — and the fix is either more budget to that campaign or less to others, which is an allocation decision, not a spending decision.

What it means for hiring: if budget is capped and misallocated, you have a genuine optimization problem on your hands — the kind of ongoing work monthly management is actually for. This is one of the stronger signals that management, not just a one-off fix, would earn its fee.

Step 7: Ad quality — what buyers actually see

Open a few of the highest-spending ad groups and read the ads as a customer would. Are there at least two genuinely differentiated responsive search ads per ad group, or duplicates wearing different hats? Are sitelinks, callouts and call assets populated and current, or empty and leaving free ad space unused? Is the offer and any pricing claim in the copy still accurate? And does the ad's promise match what the landing page delivers — the mismatch between "free quote in 60 seconds" and a form that asks eleven questions is a conversion killer no bid strategy can rescue.

Ad Strength itself is a diagnostic, not a goal — chase message clarity, not a green badge. But assets left empty and ads untouched since launch are signs of an account running on whatever was built at the start.

What it means for hiring: weak ads and empty assets are cheap to fix and quick to test. If this is your main finding, one focused project beats an open-ended retainer.

Step 8: Landing pages

Check where each major ad group actually sends people. The homepage is the default failure: it converts at a fraction of the rate of a page built around one service and one action. A purpose-built landing page repeats the ad's promise, names the service and area, shows proof, and asks for exactly one action — form, call, or booking — with the phone number visible on mobile.

Then check speed on a phone, not a laptop — most paid clicks arrive on mobile and a five-second load has lost a share of them before the page renders. Google Ads' landing page report will show you which pages convert below par; for service businesses, any page under roughly 5% conversion is usually the page's problem, not the traffic's.

What it means for hiring: if the account sends everything to a homepage, the highest-leverage work isn't in the ads account at all — it's building one or two proper pages. Some agencies do this; many don't. Ask explicitly, because it's the piece most often quietly out of scope.

Step 9: Lead quality — the check that decides everything

Cost per lead is easy to game and easy to misread. A campaign can halve its cost per lead by widening targeting to people who will never buy, and the dashboard will look like a triumph while the calendar stays empty. The only number that matters is cost per customer: total media spend divided by customers actually won.

Getting to it requires knowing what happened to your leads. If you don't already, ask every enquiry how they found you, or tag leads in your CRM by campaign. Then compare: are the leads from the cheapest campaigns the ones that close, or is the expensive campaign quietly producing all the real customers? For B2B, feeding closed deals back into Google Ads via offline conversion tracking is the single biggest measurement upgrade available — our guide to Google Ads offline conversion imports explains the mechanics.

What it means for hiring: if nobody currently knows which campaigns produce customers, that's the gap a good manager closes first. Ask every candidate agency how they measure lead quality for other clients — not cost per lead, lead quality. The distinction in their answer tells you most of what you need to know.

Step 10: Attribution — what the account claims vs what actually happened

Last, because it's the check that keeps all the others honest. Google Ads reports conversions using its own last-click-leaning model and its own attribution windows; GA4 counts differently; your CRM counts what actually arrived. These three numbers will never match exactly — the question is whether they're within a plausible distance of each other.

Pull conversions by month from Google Ads and compare against your own lead records for the same months. Big gaps mean either tracking problems (back to step 1) or attribution settings doing work you didn't authorize. Also check the model itself: if the account credits conversions through data-driven attribution with a 30-day click window, a campaign can look productive on leads that arrived weeks after the click and may not have come from the ad at all. Neither model is a lie, but proposals built on one platform's self-reported numbers deserve your skepticism — ask any prospective agency how they reconcile platform-reported conversions against business reality.

One more attribution trap worth naming: results claimed during the first two or three weeks of any new management arrangement are mostly the previous work's momentum. Judge any engagement on 60–90 days, not the honeymoon fortnight — and judge your audit findings the same way.

Turn the audit into a hiring decision

Score what you found and it should resolve into one of three situations. Tune-up: tracking is sound, structure is clean, and the findings are one-day fixes — stale negatives, empty assets, a location default, a few zero-conversion keywords. Do them yourself this week; you don't need an agency yet. Rebuild: tracking is broken, or structure is tangled, or both. That's a bounded project worth paying for once — and it should be quoted as a project, not absorbed into a retainer. Management: the fundamentals are sound but there's ongoing work on the table — capped budgets to reallocate, search terms to review weekly, lead quality to close the loop on. That's what a retainer is for.

Whatever you found, the audit document is now your brief. Hand it to every candidate and ask two questions: which of these findings would you tackle first and why, and what would you deliberately not change in the first month? The first tests their judgement; the second tests their restraint. Agencies that neither ask about tracking nor mention your audit findings are selling you a package, not a plan.

For the surrounding decisions — freelancer vs agency, what fees should cover, and what a first 90 days should honestly look like — the Google Ads management cost guide and the agency vs freelancer comparison cover both sides in detail.

When it's worth having the audit done for you

A self-audit catches most of the expensive stuff, but it won't benchmark what you find against what's normal across many accounts, and it can't challenge its own assumptions. That's the case for an independent review — done by someone who isn't selling you management on the back of it, or at least clearly separated from any management decision.

That's how I structure it: a $299 paid account review, standalone, scored and written up as a prioritized action list you could hand to any freelancer or agency — including me, and including nobody. If the honest answer is that you don't need ongoing management, the review says so; it's also credited against the first month if you do go ahead with Google Ads management. Book a 30-minute call and I'll tell you which of the three situations above your account is in — even when the answer means you shouldn't hire anyone yet.