Invalid traffic & fraud · also called Sophisticated invalid traffic
SIVT (sophisticated invalid traffic)
SIVT (sophisticated invalid traffic) is invalid traffic designed to look human or legitimate, which can only be found through advanced analytics, multi-point corroboration and often human review.
SIVT is the part of invalid traffic that tries not to get caught. Instead of announcing itself like a crawler, it copies the look and feel of real people and real publishers, so simple lists cannot stop it.
The MRC IVT guidelines list SIVT types including: bots and browsers that mimic human behaviour, hijacked devices, sessions, ad tags or creatives, adware and malware that load ads users never asked for, invalid proxy traffic such as abused residential proxies, incentivized manipulation of measurements, domain spoofing and app spoofing (misrepresenting where an ad ran), cookie stuffing and cookie harvesting, falsified viewability, manipulated location data and, in streaming TV, SSAI spoofing.
Detecting SIVT means corroborating many signals together: IP reputation, device and browser consistency (device fingerprinting), behavioral analysis of mouse, touch and timing patterns, supply-chain checks via ads.txt and sellers.json, and statistical anomaly detection. MRC requires SIVT detection to use these advanced methods, and accredited vendors combine automation with analyst review. A low SIVT number is only meaningful if the vendor is actually accredited for SIVT.
Think of it like this
If GIVT is a known troublemaker on the banned list, SIVT is a pickpocket in a good suit: you only catch him by noticing that his behaviour, his story and his timing do not add up.
An example
A video campaign looks clean on GIVT checks, but analysis shows 40,000 "viewers" whose devices claim to be iPhones while their screen and rendering traits match a Linux server, all watching from residential IPs in the same few city blocks. That is SIVT.
Related terms
GIVT (general invalid traffic)
GIVT (general invalid traffic) is invalid traffic that can be identified with routine, list-based or parameter-based filtration, such as known crawlers, data-center traffic and non-browser user agents.
Invalid traffic (IVT)
Invalid traffic (IVT) is any impression, click or visit not coming from a real person with a genuine chance to see the ad, whether caused by bots, fraud or error.
Bot traffic
Bot traffic is visits, ad impressions or clicks generated by automated software rather than people, ranging from honest search crawlers to fraudulent bots built to look human.
Domain spoofing
Domain spoofing is misrepresenting the website where an ad will run, so that inventory from a low-quality or fake site is sold to advertisers as if it came from a premium publisher.
Behavioral analysis
Behavioral analysis in fraud detection is examining how a visitor interacts, such as mouse movements, taps, scrolling, typing rhythm and session timing, to tell real humans from bots and scripted activity.
Fraud detection
Fraud detection in advertising is the process and technology of identifying invalid or fraudulent impressions, clicks, installs and conversions, so they can be blocked in advance or excluded from billing.
Sources: MRC Invalid Traffic Detection and Filtration Guidelines Addendum (2020 update), MRC standards and guidelines, TAG: fighting fraud