Key takeaways
- A Google Ads audit checklist should run in dependency order: ownership and access, then tracking, then bidding, then what the account buys, then what it returns.
- Tracking problems hide everything after them. I've audited one account that counted only clicks and another that reported zero conversions in a month with 12 real leads.
- Waste usually sits in three places: loose match types, negative lists that are missing or in conflict, and brand spend nobody has tested.
- Platform numbers aren't the last check. Join spend to closed deals, and for multi-location accounts check whether your own locations are bidding against each other.
01The Google Ads audit checklist
This is the checklist I run, in this order. Each row is a check I've done on a real account, and the order matters: if the account can't count a conversion, every later number is wrong.
| No. | Check | What I look for | What it found |
|---|---|---|---|
| 1 | Account ownership and access | The company owns the account and has full access, not a monthly report | Purolator: accounts under the agency's login. I asked for access and didn't get it |
| 2 | Conversion actions | Leads or sales are tracked as conversion actions and tied to the CRM | Purolator: the account counted clicks. No conversion actions, nothing tracked through Salesforce |
| 3 | Primary goals | The lead actions are primary, so they fill the Conversions column | Pipe17: 12 click-ID contacts in HubSpot in a month, zero conversions in Google Ads. Both form goals were secondary |
| 4 | Bidding strategy | Campaigns bid toward conversions, not clicks or impression share | Pipe17: six of eight campaigns were on clicks or impression share |
| 5 | Match types | Search terms that match what the keyword meant | Pipe17: 75% of generic clicks were unrelated. Phrase to exact cut irrelevant clicks about 85% |
| 6 | Negatives and shared-list conflicts | Negative lists exist, and no shared list blocks keywords you want | Pipe17: a shared list was blocking 51 of 146 new keywords |
| 7 | Brand versus non-brand split | How much budget buys your own name, and whether that's been tested | Pipe17: about half the budget was on brand. A 22-day pause barely moved total leads |
| 8 | Observation audiences | Audiences attached as observation, not targeting, unless you mean it | Pipe17: one campaign at zero impressions for about 11 days |
| 9 | Deal-level ROAS | Spend joined to closed deals over a window as long as the sales cycle | Pipe17: 4.0x over six months, one keyword at 38.2x, five keywords cut |
| 10 | Locations bidding against each other | Two of your own locations in the same auction | Dental Team FL: ten practices and five agencies bidding on the same searches |
| 11 | Per-location targets | A cost target for each location, set from that location's own costs | Dental Team FL: $75 at one office and $120 at another |
02Checks 1 to 4: ownership, tracking and bidding
Ownership and conversion actions
At Purolator I was brought in to audit an outside agency. The audit was short because the problems were basic. The account counted clicks, with no conversion actions, so nobody could say which keyword produced a lead. Results were reported as reach and impressions, with nothing tracked through Salesforce. And the ad accounts sat under the agency's own login. The rule I've kept since: the company owns its ad accounts and has full access, and that goes in the contract before work starts.
Primary goals and bidding
An account can have conversion actions and still report nothing. At Pipe17 the tags and the offline upload were fine, but the demo and contact form goals were set as secondary, so form fills never reached the Conversions column. Six of eight campaigns were also bidding on clicks or impression share. I made both goals primary and moved all eight campaigns to conversion-based bidding the same day. The full diagnosis is in Google Ads showing zero conversions.
03Checks 5 to 8: where the money leaks
- Match types.Read the search terms report against each keyword. At Pipe17, phrase match on a broad category term matched all kinds of unrelated software. I paused 19 phrase keywords and created 19 exact versions. Don't cut by rule, though. At Dental Team FL, broad match stayed where it worked: in one 28-day window, 136 of the account's 439 conversions came from searches the phrase and exact versions would have missed.
- Negatives.Check that negative lists exist and that someone reviews search terms. Pipe17's list grew past 1,500 terms. I had also installed a script that added negatives to a shared list every morning, and removed it once it became a risk to new keywords.
- Shared-list conflicts.Compare every shared negative list against the keywords you're bidding on. After a rebuild at Pipe17, old competitor-name negatives in a shared list were silently blocking 51 of 146 new keywords and had left two ad groups dead. I removed 35 entries, then ran a 619-keyword serving audit that cleared 16 more.
- Brand versus non-brand.Split spend and conversions by brand and non-brand. If brand takes a large share, test it. At Pipe17 I paused brand and compared 22 days after with 22 days before. Leads with a Google click ID went from 9 to 7, and organic and direct contacts both ticked up. At Dental Team FL, most of the gap between agency reports and the booking system was branded search. At Purolator I spent heavily on brand and never tested it, which I'd do differently now. The method is in how I run incrementality tests.
- Observation audiences.Open the audience settings on every campaign. A target account list attached without an explicit observation setting restricted one Pipe17 campaign to list members only. It served within 24 hours of the fix. More on that in Customer Match in Google Ads for B2B.
04Check 9: return on closed deals
Platform ROAS only knows what the platform can see. The last general check joins spend to revenue in the CRM. At Pipe17, with a sales cycle of six months or more, I joined six months of Google Ads spend to HubSpot deals through the Google click ID, keyword by keyword. The account returned 4.0x on closed deals. One keyword returned 38.2x, and five keywords had spent real money and produced nothing, so I cut them.
The same idea works outside B2B. At Dental Team FL I compared what each agency claimed against call tracking data and the practice management system, and the reports didn't match the bookings.
05Checks 10 and 11: multi-location accounts
Dental Team FL is a group of ten practices. Five agencies each ran a channel, and the offices bid against each other for the same searches, including the group's own brand name.
- Overlap. Look for two of your own locations eligible for the same search. The fix was one account, with one Search campaign per office fenced to its own area and one Performance Max campaign across all ten.
- Targets. A click in one city doesn't cost what it costs in the next, so each office got its own cost-per-conversion target. At one point that was $75 at one office and $120 at another.
With tighter match types and proper negative lists on top, average cost per click fell from $9.40 to $5.80 across all locations.
06How to use the checklist
Run it top to bottom and stop at the first failure in checks 1 to 4. There's no point reading search terms in an account that can't count a lead. Checks 5 to 8 need the search terms report and the settings pages. Check 9 needs CRM access and is the one that changes budgets.
The list is short on purpose. It only holds checks I've run and can show a result for. If you'd like me to run it on your account, that's my Google Ads audit.
FAQQuestions people ask
- What should a Google Ads audit include?
- Farhan Munshi's audit runs 11 checks in order: account ownership and access, conversion actions, primary goals, bidding strategy, match types, negative keywords and shared-list conflicts, the brand versus non-brand split, audience settings, return on closed deals, and for multi-location accounts, overlap between locations and targets per location. Each check comes from an account he has run.
- What is the first thing to check in a Google Ads audit?
- Who owns the account and who can log in. At Purolator, Farhan Munshi found the ad accounts sat under the agency's own login and the company received only a monthly report. He asked for access and didn't get it. Without access, no other check on the list can be run, so ownership comes first.
- How do you audit Google Ads conversion tracking?
- Check that conversion actions exist, then check which ones are primary. At Purolator the account counted clicks and had no conversion actions. At Pipe17, Farhan Munshi found 12 CRM contacts with a Google click ID in a month and zero reported conversions, because the form goals were set as secondary.
- How do you find wasted spend in a Google Ads account?
- Read the search terms against match types and negatives. At Pipe17, Farhan Munshi found 75% of clicks on generic campaigns came from unrelated searches. Moving 19 phrase keywords to exact match cut irrelevant clicks about 85%. He also checks the brand split, since about half that account's budget was going to its own name.
- How do you audit Google Ads for a business with multiple locations?
- Check whether your own locations meet in the same auction, and whether each has its own target. At Dental Team FL, ten practices bid against each other under five agencies. Farhan Munshi moved them to one account with one campaign per office, fenced to its area, with a cost target per office. Average cost per click fell from $9.40 to $5.80.
