Performance Max Fake Leads: How to Diagnose the Source Before Changing the Campaign
A spike in poor Performance Max leads is a reason to investigate the source, form and conversion goal together. It is not proof that every PMax channel is fraudulent. First establish which submissions are fabricated and which are genuine but unsuitable, then use the reporting and controls available for the affected inventory.
Quick answer
A spike in poor Performance Max leads is a reason to investigate the source, form and conversion goal together. It is not proof that every PMax channel is fraudulent. First establish which submissions are fabricated and which are genuine but unsuitable, then use the reporting and controls available for the affected inventory.
For marketing, analytics, media buying and security teams reviewing paid traffic quality. This guide is technical and operational information, not a legal finding about a specific source.
Why PMax lead quality needs separate measurement
A blended campaign result can hide changes in traffic mix and qualification. Compare accepted submissions with later outcomes for the same cohort. If the campaign's measured success is a form fill but the business needs a confirmed enquiry, a low reported CPL can coexist with expensive sales work.
Fake leads vs cheap low-quality conversions
Cheap does not mean fabricated. Separate invalid contact details and supported automation findings from real people outside the offer's criteria. A lead that cannot be reached after one attempt should remain pending under your contact policy. Use consistent rejection reasons before comparing PMax with Search.
Placement and inventory context
PMax reaches multiple Google channels. Review the channel and placement information actually exposed in the account, together with landing pages and form paths. Missing placement-level attribution is a coverage limit, not permission to assign every rejected lead to a guessed publisher. Track the campaign, asset group and source fields that are reliably available.
Conversion signals and optimisation feedback
Review the conversion actions and goals the campaign uses for bidding. If raw submissions are counted as success, sales rejection does not automatically change that signal. Check duplicates, test events and conversion values before concluding that traffic quality alone explains performance.
CRM evidence you need
Retain the accepted submission ID, available click identifier, receipt time, contact checks and qualification decision. Preserve the reason and time of each later change. Evaluate old and new leads after comparable follow-up periods. If consent or tracking limitations prevent a reliable join, report the unmatched share separately.
How to analyse a sudden spam-lead spike
Locate the first changed cohort. Compare its landing pages, devices, geography, timing, payload repetition and downstream acceptance with a stable earlier period. Check releases, tracking changes, offer edits and sales response delays on the same timeline. A coincident placement increase is a hypothesis to test, rather than a sufficient cause.
Form protection
Validate submissions on the server and examine repeated abuse of the endpoint. Consider email confirmation and appropriate challenges where friction is justified. Validate a control against both rejected submissions and genuine prospects who abandon or fail it. Website controls cannot inspect a platform-hosted lead form in the same way as your own form.
Qualified/offline outcomes
Google documents qualified and converted lead goals and advertiser-configured first-party feedback. Define the qualification rule, account for outcome delay and test matching before using a deeper event for bidding. This requires your measurement and CRM configuration; the article does not imply AdFraud.pl sends offline conversions or changes PMax goals.
Placement exclusions and other campaign controls
Google describes account-level placement exclusions and other PMax controls with inventory-specific coverage. Account-level IP exclusions can also apply to PMax; campaign-level IP exclusions have a different scope. Review the current account settings and documented coverage before changing a rule. A website block and an ad-delivery exclusion act at different stages.
When pausing PMax is not yet supported by the evidence
A broad pause may be premature when the apparent spike comes from duplicated tracking, an unreconciled CRM cohort or one form defect. Immediate containment can still be reasonable if damage is ongoing, but record it as a business decision under uncertainty. Prefer a reversible, scoped change and an agreed quality measure when the evidence supports one.
Review PMax traffic and downstream outcomes before choosing a response
Related reading
Sources and further verification
Platform rules, standards and detection practices change. Check the current primary source and your own measurement scope before applying a control.