Bot detection logic — summary
AdFraud Shield is not an "AI black-box". We analyze five layers in a defined order — Network, Protocol, Browser, Identity, Behavior — and we explain the verdict to the customer. This page is the summary; the Technical Brief has the depth.
FP rate < 0.1%
Estimated false-positive rate (expected from the ratio-guard design; not a guarantee)
Inbound clicks pass through five filters from top to bottom. Each layer evaluates independent signals; multiple layers contradicting each other lift the verdict score.
INBOUND TRAFFIC
01 · Network layer
Network
UA/OS · TTL hops ※
02 · Protocol layer
Protocol
TLS JA4 ※ · HTTP/2 ※
03 · Browser layer
Browser
Canvas · WebGL · webdriver
04 · Identity layer
Identity
Datacenter IP · Geo/TZ
05 · Behavior layer
Behavior
Frequency · Honeypot · Behavior · Ad-click fraud
HUMAN / SUSPICIOUS / BOT
※ Edge-dependent signal — requires a network-edge layer, not enabled in the current configuration. Disclosed in full: 11 implemented — 8 active / 3 edge-dependent.
We can tell you, per detection, why the verdict was bot — signal name, category, what was matched. That's structurally different from "AI says so."
A 24-hour cooldown plus a 7-day human/bot ratio guard actively detect and reverse mistakenly-excluded IPs.
Canvas/WebGL fingerprinting, behavioral honeypots, datacenter-IP and frequency analysis — all techniques validated in academic literature and OSS. Nothing arcane.
Zero false positives is impossible for any bot detector. AdFraud Shield's guard mechanism reverses mistakenly-excluded IPs within 7 days — minimizing the time real users are blocked from your ads.
Registration no. 17002460
Valid until 1 November 2027
Assessed by JUAS (Japan Users Association of Information Systems)
Search registration number 17002460 in JIPDEC's Privacy Mark holder search to verify.
View the official registrationRegistration no. IR0268
Valid until 5 August 2028
Certified by IMJ Certification Centre
Verifiable in the ISMS-AC register of certified organisations.
View the official registrationISO/IEC 27001 scope (the part relevant to this service)
"Planning, development, operation and maintenance of cloud-based web marketing tools and mail-order systems"
See the ISMS-AC register for the full scope.
AdFraud Shield is a cloud-based web marketing tool planned, developed, operated and maintained by FID Ltd.
The PrivacyMark grant and the ISO/IEC 27001 certification apply to FID Ltd.'s management systems. AdFraud Shield is not separately certified, and neither guarantees that the system is free of vulnerabilities.