Prioritise Search Fixes by Impact: Focus on fixes that affect reliable conversion signals, not just high spend.; Use business records to estimate impact—avoid over-precise forecasts from uncertain data.; Track each fix with expected effect, confidence, owner and review date.
Image: Paid Search Desk

Account Audits

Part of Search campaign account audits

Prioritising paid search fixes by expected business impact

Turn Search audit findings into a fix queue using observed exposure, plausible business value, evidence, downside and a clear review plan.

Prioritise an audit finding by the business result it could change, the evidence behind that expectation and the downside of acting.

An untidy account item may matter less than an unreliable conversion signal used by a large campaign. High spend alone does not make a proposed fix valuable if it could also remove useful customers.

State the decision

Turn each finding into a specific proposal. Instead of “improve keywords”, identify the visible requests that appear outside the offer and assess a scoped response.

Instead of “increase bids”, confirm that the selected conversion action represents a useful outcome before changing its target. Name the affected campaigns and the customer outcome the proposal aims to protect or increase.

Possible effectEvidence to seekLimit
Reduce unsuitable trafficVisible queries, cost and service eligibilityThe report omits some queries; an exclusion can block useful demand.
Restore operating controlMissing administrator or unresolved dependencyThe consequence depends on the account and connected systems.
Increase useful enquiries or salesAccepted requests, orders and capacityPlatform conversions may measure an earlier stage.
Improve the basis for decisionsConversion settings, status and business recordsA measurement fix can change reported totals without creating customers.

These are possible effects, not a fixed ranking. An access or measurement fault can warrant early action because it affects later decisions even when an immediate revenue gain cannot be estimated.

Estimate impact without false precision

Use the business's own records. For a traffic mismatch, identify spend on visible unsuitable terms; do not treat it as a guaranteed future saving.

For a sales opportunity, consider the contribution from work the business can accept and fulfil, using its agreed treatment of variable costs. Keep unknown acceptance rates, margins and delayed sales visible.

Separate observed exposure, plausible change and business value. A query's past cost establishes exposure. It does not establish how much a future exclusion would save or whether redirected spend would produce a sale. Avoid counting the same traffic shift once as saved cost and again as increased sales value.

A qualitative estimate such as high, moderate or uncertain, with a reason, can be more useful than a precise score built on unknown lead quality. If dependable records support a range, state its assumptions.

Key Metrics for Impact Estimation

Observed exposure
Past spend on visible terms (e.g., Google Ads reports)
Plausible change
Estimated shift in traffic or conversion potential
Business value
Contribution from fulfilled work (considering variable costs)
Avoid double counting
Do not count saved cost and increased sales from same traffic shift

Weigh confidence and downside

Ask how many campaigns the issue affects and whether the evidence is a current setting, a repeated observed pattern or one ambiguous case. Check whether a proposed change is reversible and whether it risks cutting eligible demand.

Negative matching differs from positive matching, so assess the queries a broad exclusion could affect. A conversion-goal change alters the signal available to conversion bidding and deserves a dated record.

Keep an action queue with the proposed fix, expected business effect, confidence and downside, owner and review date. An access gap may need prompt attention without a spend estimate. A copy idea supported by one click may belong in an investigation queue.

Decide how to review the result

For each fix, specify later evidence that would support or challenge it. Query controls need later visible terms and accepted-enquiry checks. A measurement correction needs event and business-record checks before a performance comparison. A bid-related change needs time for conversions and sales assessment to mature.

Record other material changes during the review period. If the offer, budget and conversion definition change together, a rise in reported results cannot confidently be credited to one fix. Close the item as supported, not supported or still uncertain, with the reason.

More from Account Audits