Google announced the Invalid Traffic Overview Card on November 12, 2026, giving advertisers a clearer view of protection-related savings and credits. The useful question for a marketing manager is what each number means for the next decision. A reassuring summary should not become an unsupported claim about fraud, profit or campaign success.
For a service business, keep three questions separate: what activity was filtered, what billing adjustment was issued, and which enquiries became useful conversations. This guide proposes a review process that connects those questions without pretending they share one denominator. It is a reporting framework, not a hands-on test of the new card.
Name the evidence before interpreting the number
Google’s help guide distinguishes estimated savings from protections and credited cost for earlier activity identified after billing. It also says the card is rolling out gradually. Check whether it is actually available in your account; an absent card does not establish whether a campaign has a traffic-quality problem.
Keep the original metric name beside any exported value. Estimated savings are not an extra sale, a cash payment or proof that a manager recovered money. A credit belongs in a reconciliation record. Neither, on its own, tells you whether the business reached the right people or responded effectively.
Start the review with a specific question: are we explaining a billing adjustment, investigating a campaign change, or checking enquiry quality? Write that question above the report. It prevents a meeting about poor enquiries from drifting into a presentation of unrelated protection totals.
Match the period and campaign scope
Build a small evidence sheet with the account currency, account time zone, date range, selected campaigns, report name and export time. These are proposed review fields, not new Google features. Save the source export privately so another reviewer can reproduce the comparison rather than trust a cropped screenshot.
The Invalid Activity Credit Report documents campaign and network breakdowns for Search and Performance Max. It also warns that an interaction and its credit can fall in different months. Match an adjustment to its service period before comparing it with this week’s enquiries; the month when a credit appears may answer a different question.
Choose one consistent comparison basis and label exceptions. If the campaign selection or date basis differs, leave the rows separate until you can explain the difference. Do not combine estimated avoided cost and an issued credit into a homemade return-on-investment figure: neither measures additional business revenue.
Separate filtered counts from later adjustments
Google describes the Invalid clicks column as filtered clicks that are not charged. Treat that count as a count. Comparing it directly with a currency amount, or dividing it by an unrelated website-session total, creates a misleading interpretation even if both source exports are accurate.
Google’s invalid-traffic guide says past-activity credits do not retroactively change frozen month-end campaign metrics. Its separate credit report provides adjusted metrics; the report documentation defines adjusted cost as invoiced campaign cost minus the relevant invalid-activity credit. Preserve those labels when presenting the two views.
Keep an original reporting view and a clearly labelled reconciliation view. Record which adjustment is included and which period it belongs to. Do not subtract a credit again from a figure that already incorporates it. If a finance colleague cannot trace the calculation, resolve that uncertainty before circulating a performance conclusion.
Review lead quality in the customer journey
Traffic filtering does not guarantee clean enquiries. Google’s lead-generation guidance recommends additional verification, including server-side validation. Choose safeguards appropriate to the form and test their effect on real people, including mobile visitors. A protection mechanism that prevents legitimate customers from completing the form creates another operational problem.
In a private CRM review, distinguish an irrelevant request, an unreachable contact, a duplicate, an enquiry outside the service area and an unanswered legitimate request. Mark unresolved cases as unknown. An unsuccessful call does not establish that the person was fake, and an unqualified enquiry does not by itself prove billable click fraud.
Inspect the message-to-page-to-response path before assigning blame. Did the advertisement describe the actual service? Did the form collect enough context? Was a responsible person assigned? Did that person follow up? Use access-controlled records and publish only non-identifying summaries; a public portfolio does not need customer details to demonstrate a sound method.
Walk through a hypothetical service-business review
Imagine a roofing company serving part of York Region. A manager sees a protection figure while the team reports more enquiries it cannot qualify. This is an illustrative planning example, not a client result or a local demand estimate. The task is to investigate two observations without assuming one explains the other.
First, the manager records the card’s exact labels and reporting scope. Separately, the reviewer checks any billing adjustment against its original service period. The team then samples recent CRM enquiries with the same classification rules and notes where location, service eligibility, contactability or response ownership remains uncertain.
The output should be a decision record, not a dramatic fraud percentage. If the landing page creates the wrong expectation, assign a content correction. If valid requests remain unanswered, fix ownership. If a billing question remains unresolved, preserve the relevant evidence for an authorised account review. These actions can coexist without a proven common cause.
- Record the source, period and meaning of each metric.
- Keep estimates, issued credits and enquiry outcomes in separate fields.
- Use the same lead-quality definitions throughout the sample.
- Give each unresolved issue an owner, next action and review date.
Finish with an action someone can verify
Write the conclusion in plain language: what was observed, what remains uncertain, what will change, who owns it and what evidence will close the issue. A reporting improvement might simply be a clearer reconciliation note. A lead-quality improvement may require a better offer, validation or follow-up process rather than a new traffic dashboard.
Review the outcome after the relevant data has arrived, keeping the chosen date basis visible. Avoid promising a credit, attributing activity to a competitor or declaring a campaign healthy from one protection total. The new card can support a useful conversation; the manager still has to connect platform evidence with business operations.
Key takeaways
- Preserve the difference between estimates, issued credits and useful enquiries.
- Match campaign scope and service periods before comparing reports.
- Do not deduct an adjustment twice or describe avoided cost as revenue.
- Investigate lead quality through the landing page, form and CRM.
- End with an owner, next action and evidence for the follow-up review.
Sources & further reading
- Google announcement: Invalid Traffic Overview Card, November 12, 2026
- Google Ads Help: about invalid traffic and the gradual card rollout
- Google Ads Help: Invalid Activity Credit Report and adjusted metrics
- Google Ads Help: statistics-table columns
- Google Ads Help: high-quality lead-generation practices
Official Google sources checked on November 15, 2026. The announcement is dated November 12; supporting help pages show no visible publication date. These are same-vendor descriptions, not independent performance evidence. Account-level rollout, specific credits and the new card were not tested in a live account. The review framework and roofing example are original analysis; no client figures, fraud attribution, savings result or credit entitlement is claimed.
Farshad Ghobadvand