Check measurement before touching bids
Confirm what counts as a primary conversion and whether it genuinely matches the account’s objective. Test form, call and purchase tracking. Review conversion windows, duplicate events and whether campaigns optimise for a meaningful outcome.
Audit the measurement before judging the campaign
Open the account and list the conversion actions used by each campaign. Check which actions contribute to bidding and whether they represent business outcomes. Test phone, form and purchase events where relevant. Review attribution windows, duplicate events and the match between credited conversions and accepted leads. Pull a period long enough to cover the purchase or sales cycle and note recent changes before drawing conclusions.
- Confirm trackingTest which goals bidding uses
- Inspect deliveryReview real search terms, costs and destinations
- Prioritise fixesRecord changes and compare later business outcomes
Inspect where the budget went
Review search terms, targeting, locations, match types and ads by campaign objective. Look for irrelevant terms and separate brand demand from new-customer campaigns. Compare traffic and cost with actual lead or order quality, not just click volume.
Review intent, campaign structure and spend
Inspect search terms to see what triggered search ads. Separate ad group keywords from actual queries. Review query relevance, location settings, budgets and performance by campaign type. Compare how brand and non-brand searches perform and check that negative keywords do not block important queries. Test the ad-to-page journey on a phone and verify that the same offer and eligibility details are visible.
Prioritise and document changes
Fix measurement or major leakage before testing creative. Record the date, campaign and hypothesis for each change. Leave enough observation time to judge a decision, particularly when sales take days or weeks to close.
Prioritise fixes and set a review window
Rank findings by cost, confidence and implementation effort. Broken measurement and wasted spend usually deserve attention before rewriting assets with little data. Make changes in a sequence your team can monitor. Record the date, owner and outcome of each change. Check impressions and spend soon after release, then use a longer review window for lead quality and sales.
A dealership audit example
A showroom sees a large volume of conversions, but the account counts map clicks as leads. The team makes the successful test-drive request primary for campaigns where that is the goal. It reports navigation clicks separately, reviews generic parts queries and tracks confirmed appointments in the CRM.
Review operational gaps that do not show up under campaign performance. Check account access, billing contacts, Merchant Center connections where relevant and whether staff know who approves a budget change. Find sudden changes in the account history and check which were automated, which were manual and what the owner expected to happen. An audit should produce a prioritised work list with owners and success measures, so teammates can verify the changes after the audit ends.
| Finding | Potential impact | First move |
|---|---|---|
| Page views count as leads | Bidding may optimise for the wrong action | Test and correct the conversion definition |
| High spend on unrelated queries | Budget misses the intended buyer | Review matching and add carefully chosen exclusions |
| Ad promise differs from landing page | Qualified visitors may leave | Align copy and the live destination |
Common questions
Should I pause every campaign with poor ROAS?
First check whether the tracking, conversion delay and comparison period make sense. Prioritise clear waste and investigate campaigns with weak evidence before pausing broadly.
Can recommendations replace an audit?
Use recommendations as inputs. Review each against campaign objectives, business data and the constraints of the account.
How far back should the audit go?
Use enough data to cover the conversion lag and normal buying cycle. Compare a recent period with a representative earlier one and document promotions, outages and tracking changes.
What to check next
- ConversionsTest events before reallocating spend
- TrafficInspect actual search terms and location settings
- ChangesRecord hypothesis and observation period




