The PPC audit we run before taking on a Google Ads account: five checks in order, what good looks like, what each costs when wrong, plus a 20-minute pass.
You're paying for Google Ads every month and something doesn't add up. Leads are flat while spend creeps up, or you're about to change supplier and want to know what you're handing over. A PPC audit is how you find out where the money is going before you decide what to change. This is the one we run on every search advertising account we take on, in the order we run it.
The order matters more than people expect. On a handover call earlier this year, a prospect asked us to start with bidding. We asked to see the conversion setup first. Two of the five conversion actions were double counting, so every bidding decision had been made on numbers roughly a third too high. Nothing about the bidding needed fixing until that did.
The short version
- Five areas explain almost every wasted pound we find: conversion data that isn't real, search terms that could never convert, structure that stops budget reaching winners, automation nobody is supervising, and budget concentrated on brand rather than growth.
- Check them in that order, because changing bids on top of broken conversion data teaches the algorithm the wrong lesson faster, and on a £10,000 a month account the first two areas alone typically account for £1,500 to £3,000 a month.
- Clear Click runs this audit before taking on any Google Ads account, and we charge for it, because the discovery is what protects the work that follows.
1. Is the conversion data real?
Smart bidding is only as good as the conversion it optimises towards. If the number Google is chasing is inflated or the wrong event altogether, every other finding is built on sand, so this comes first.
What to pull. The conversion actions list, with each one's source, counting setting (one or every), primary status and value. Then the last 90 days of conversions by action, next to the real number from your CRM or inbox.
What good looks like. One primary conversion per commercial outcome, counted once per click, with a value that reflects margin. A newsletter sign-up and a five-figure enquiry should never carry the same weight.
What it costs when it's wrong. Take an account where 30% of counted conversions are duplicates or low-value events. A reported £80 cost per lead is really £115, and a year of budget decisions has been made on the £80. The algorithm has also been steered towards whatever is cheapest to trigger, which is usually the thing worth least.
The 20-minute version
Open Goals, then Conversions. Count the primary actions. More than three, or two describing the same thing, is your first fix. Then compare last month's platform conversions to last month's real enquiries. A gap over 20% in either direction means measurement gets sorted before anything else. If that gap is a familiar frustration, we've written up why GA4 and ad platform numbers never match.
2. What are you actually paying for?
Not the keywords you bid on. The queries you matched to. Broad match plus smart bidding has widened this gap, and the waste hides inside averages that look fine at campaign level.
What to pull. The search terms report for the last 90 days, sorted by cost, with conversions alongside. Then the same report filtered to terms with zero conversions and more than £50 of spend.
What good looks like. The top 50 terms by spend all describe something you sell, to a buyer you want, in a place you serve. In most accounts we inherit, a meaningful share of spend sits on queries that could never convert: careers searches, competitors' returns policies, informational queries with no buying intent.
What it costs when it's wrong. Take a firm spending £10,000 a month with 15% of it on terms that could never convert. That's £1,500 a month, £18,000 a year, and it is the fastest saving in most audits. Negatives take an afternoon.
The 20-minute version
Set the search terms report to 90 days, sort by cost, and read the top 100 terms out loud. Anything you'd be embarrassed to have paid for goes into a negative list today. Multiply the spend on those terms by four for the annual figure.
3. Can budget flow to what works?
Over-segmented accounts starve the algorithm of data. Over-consolidated accounts hide losers inside winners. Once the measurement is honest and the obvious waste is gone, structure decides whether the account can improve on its own.
What to pull. The campaign list with budget, spend, cost per conversion and the "limited by budget" status. Then a note of why each campaign exists: product, geography, match type, or a decision someone made in 2022.
What good looks like. You can tell, in one view, which parts of the account are profitable. Campaigns are split by how the business makes money and by what needs a different target, not by legacy logic. Each smart-bidding campaign gets 30 or more conversions a month, the rough floor below which the algorithm is guessing.
What it costs when it's wrong. The pattern we see most often is the best campaign capped. Take a £10,000 account where campaign A converts at £40 and is limited by budget at £2,000, while campaign B converts at £130 and spends £5,000 freely. Moving £2,000 from B to A trades 15 leads for 50. That is 35 more leads a month for the same money, and the structure was the only thing stopping it.
The 20-minute version
Sort campaigns by cost per conversion. Check whether any of the cheapest three are limited by budget, and whether any of the most expensive three outspend the cheapest three combined. If both are true, you've found the biggest lever in the account, and it needs no new creative, keywords or tools.
4. Who is in control of the account?
Auto-applied recommendations, broad match defaults and asset automation can each quietly change an account over months. Automation is not the problem. Unsupervised automation is.
What to pull. The auto-apply settings and the change history for the last 180 days, filtered to changes made by Google rather than a person.
What good looks like. Auto-apply is off, or limited to a short list someone chose on purpose. Every bidding target traces back to a commercial number: a cost per lead the business can afford, a return the finance team signed off.
What it costs when it's wrong. The common one is match types widened automatically. Take a £5,000 campaign whose cost per lead drifts from £60 to £90 over a quarter after exact match was switched to broad. That is 83 leads a month falling to 55 on the same spend, and nobody chose it.
The 20-minute version
Open Recommendations, then Auto-apply, and turn off anything you didn't choose. Then filter Change history to the last six months and count the changes attributed to Google. If it's in the dozens, the account has been running itself. Our position on where a person belongs in the loop is set out in how we run AI and automation governance.
5. Where does the next pound go?
Finally we map spend against margin and growth, not revenue. This is where an audit stops being hygiene and becomes strategy, and it comes last because you need the first four areas clean to trust it.
What to pull. Spend split between brand and non-brand terms. Brand impression share, and whether your brand ranks first organically. Non-brand impression share lost to budget, by campaign.
What good looks like. Brand spend is a deliberate decision with a reason attached (a competitor bidding on your name, a launch), not a default. The way we see it, the account should be buying growth the business couldn't get for free.
What it costs when it's wrong. Take a £10,000 account with £4,000 on brand terms where the site already ranks first organically and nobody else is bidding. If 60% of those brand conversions would have arrived anyway, that's £2,400 a month, nearly £29,000 a year, defending clicks you already own while an unbranded category with headroom sits capped.
The 20-minute version
Filter the search terms report to your company name and its misspellings and note the spend. Search your own brand in an incognito window and see who else appears. If nobody does, and you're first organically, pause brand in one small region for two weeks and watch total enquiries.
2,700 searches a month: The combined UK demand for "ppc audit", "google ads audit", "ppc account audit" and "google ads account audit", a rough count of how many businesses are asking this question every month.
Source: SEMrush, UK database, October 2026.
What should you do with the findings?
Sequence them the way you found them: measurement, waste, structure, automation, growth. The temptation is to jump to the biggest number, usually area three or five. Resist it for two weeks, because clean conversion data changes what the algorithm learns and the structure decision looks different afterwards.
Put a number next to every finding before you present it. "Brand spend is high" gets a nod. "£2,400 a month on brand clicks we'd get anyway, and here's the test to prove it" gets a decision.
To be honest, most of the audits we run end at the same place: the settings were a symptom, and the missing piece was a target the business had agreed on, whether a cost per lead the finance team can live with or a pipeline number the sales team believes. Everything is fixable once that exists.
Turning audit findings into a testing programme, rather than a one-off clean-up, is usually where the account starts compounding. See our PPC testing systems work
When should you get an independent audit?
Run the 20-minute versions first; they'll tell you whether there's a problem. An outside audit earns its fee in three situations: you're changing agency and want a fair baseline for both sides, spend is above about £5,000 a month so the findings move real money, or the in-house and agency views disagree. Our guide to in-house, freelancer or agency PPC management covers the first in more depth.
We charge for audits. Our reasoning is that discovery is what protects the work that follows, and a free audit tends to become a sales document rather than a finding. Ours take two to four weeks and end in a prioritised list with a pound figure against each item, which you can act on with us or with anyone else.
Questions we get asked
What is a PPC audit?
A structured review of a paid search account that answers one question: where is the money going, and should it be? A good one covers measurement, search terms, structure, automation and budget allocation, in that order, with a cost against each finding.
How often should you audit a Google Ads account?
A full audit once a year, or whenever the account changes hands. The 20-minute checks above are worth running monthly, and the search terms review fortnightly on accounts spending over £5,000 a month.
How much does a PPC audit cost?
Ours start from £2,000 plus VAT, with the price set by account size and how many markets it covers. A free audit is usually a proposal in a different outfit, which is fine as long as you read it that way.
What should a Google Ads audit include?
At minimum: conversion tracking checked against a real source, a 90-day search terms review, a structure review with budget flow, an automation and change-history review, and a brand versus non-brand budget map. Each finding should carry a pound figure and a recommended order of fixes.
Can you audit your own Google Ads account?
Yes, and the 20-minute versions in this piece are built for exactly that. You'll find the first two areas on your own. Where an outside view helps is areas three to five, because the person who built the account is least able to see its structure fresh.
Where we fit
Clear Click runs search advertising for B2B companies with longer sales cycles, where the audit has to connect ad spend to pipeline rather than to form fills. The same discipline sits behind Convoy ATX reaching a 32% share of voice in its category. If you're weighing up what management should cost once the audit is done, our guide to PPC agency costs in the UK sets out the bands.
If you'd like us to run the full version on your account, or just want a second opinion on what the 20-minute checks turned up, get in touch and we'll tell you where we'd focus first, and why.
