Frustration, detected automatically.

visz.ai evaluates every session server-side in a nightly pass and derives six signals: rage clicks, error clicks, dead clicks, quick backs, U-turns and excessive scroll. No black box, no ML guess.

Rage clicks are three or more clicks on the same element in under two seconds. Dead clicks trigger no reaction at all. Quick backs, U-turns and excessive scroll are caught at the session level.

An error click is a click that coincided (within 500 ms) with a JavaScript error. The click most likely had no effect, and you see straight away which error was behind it.

Rage, error and dead clicks are tied to a concrete element. The Worst Offenders list points you straight at the buttons and links that snag.

The Frustration Score condenses the six signals per session into one number. Each signal has a fixed weight, its contribution grows log-dampened with frequency, and the sum is capped at 100.

Four bands place the value: Healthy, Elevated, High and Critical. Detected bots are excluded, and because the score comes from events, it exists only for consented sessions.

Each tile links into the matching filtered sessions, each conspicuous element straight into the recording. You jump from “something snags here” to “this is what the visitor saw”.

So a signal is never an anonymous metric. It is a concrete spot you can fix.