Click fraud, conversion fraud, and invalid traffic, explained for people spending real money on ads.
Three independent figures frame ad fraud in 2026: Juniper Research put the cost near $84 billion (22% of ad spend), Imperva measured bots at 51% of web traffic, and the ANA found only $439 of every $1,000 of programmatic spend reaches a real consumer.
Read moreInvalid traffic splits into two classes. GIVT is list-identifiable — declared bots, data-centre traffic, crawlers — and gets filtered by signatures. SIVT imitates humans and needs behavioural and consistency detection.
Read moreGoogle filters obvious invalid clicks and credits some back, but it is opaque and conservative on sophisticated bots. A detection tag scores sessions, and ClickLens turns detected bots into IP and placement exclusions synced to Google Ads.
Read moreAutomated bidding buys more traffic resembling past converters, so a single fake conversion becomes a daily training signal — not a one-time loss. Bots made up 51% of web traffic in 2024 (Imperva), and each automated conversion teaches the algorithm to find more like it.
Read moreConversion fraud is bots completing the actions you optimise for — form fills, sign-ups, purchases. Bots made up 51% of web traffic in 2024 (Imperva), and a fake conversion trains your bidding to buy more of the same.
Read moreA conversion pixel fires the instant a form submits or a purchase completes — it has no model of whether the actor was human. Bots made up 51% of web traffic in 2024 (Imperva), and a pixel counts every automated conversion as real.
Read more