Every Google Ads charge in July was reconciled against Google's own credit records. Of the clicks Google reported, including those it filtered as invalid, the ones we could not observe on your site are counted on page two as unmeasured, not as checked.
This sentence appears only over a period whose reconciliation closed. It speaks for the charges, not for every click: what was observed is counted on page two. When a reconciliation has not closed, this document says so in its place.
Billed spend our detection flagged as invalid activity that Google had not credited. Halvern & Co documents and files this as an invalid-activity claim; accepted claims settle as advertising credit to this account.
Working · Detected US$21,512.96 − credited US$3,108.64 = US$18,404.32 recoverable · formula-a.3 · computed 2 August 2026, 03:14 SAST
Of the clicks above, those Google had not credited make up the recoverable figure on page one, documented click by click and kept for 90 days from when each click reached us.
Google reported 24 980 clicks, including those it filtered as invalid. 2 240 of them were observed on your site. The money was priced on a different count: Google's 22 874 billable clicks, which leave out the 2 106 it filtered.
Counts, not a rate. A rate is published only where the observation count can carry one. Every reported click we could not observe, filtered ones included, is counted as unmeasured: a count, never a value; no amount attaches to an unmeasured click.
Out-of-region could not be assessed on one of five click clusters this period; that cluster reached us through a proxy network. The finding on page one does not rest on it.
Two records, kept apart. This document is prepared from the platform's own billing and credit records, and from independent measurement of arrivals on your site. Neither is edited to fit the other; where they disagree, the difference is what gets reported.
Events, never people. Invalid activity describes events: clicks and sessions matched on pattern. It is never a claim about a person or a device.
Floors before rates. A rate is published only where the observation count can carry one. The floor in force this period: 500 observations per cluster. Below it, the counts are published and the rate is withheld.
Retention. Every click record is kept for 90 days. The 90 days run from the moment our system receives the click, not from the date of this report.
An issued report carries a code here, and entering it at kliksight.com/verify answers one question: whether Kliksight issued that exact report, and whether its figures are unchanged since it was issued. It is issued only where the agency elects it, report by report.
This code belongs to this sample and to nothing else. Entered at the verification page it returns exactly that: a specimen attached to the sample report, establishing nothing about any real report. The check on a real report confirms the document, never the platform's own figures.