Multiple flashing locations cancel one another's priority
Aliases: alert fatigue · visual competition · alert prioritization · multiple alerts
What it is
Alert priority dilution occurs when several locations flash simultaneously and each loses the meaning “more urgent than the rest.” Flashing should pull one event from background. When the background flashes everywhere, people face a competing set of transients rather than a clear action order. A critical alert may disappear in the crowd, users may handle whichever is nearest, or they may ignore all flashing to continue work.
Why it happens
Attention is finite and strong transients compete for the same orienting mechanism. Vision can detect that many things changed but cannot infer which consequence is greatest; under load, people choose by location, color, recency, or habit rather than the intended risk order. Repeated equal-status alerts also create habituation and alert fatigue. More flashing does not linearly increase discovery; it reduces the diagnostic value of each signal.
Studying it
Create tasks with simultaneous events of different consequence. Vary flash count, position, urgency expression, and arbitration rules; measure whether the first action is correct, omissions, primary-task interruption, and recovery. Test brief overlap and accumulated alerts separately: the former challenges immediate selection, the latter reveals fatigue and ignoring strategies. Seeing every alert is insufficient; compare handling order to actual risk order.
Where it stops holding
Failures can genuinely create several related messages, and hiding them to pretend normality is not appropriate. Related alerts can become one primary event with expandable detail, while independent high-risk events may need cross-channel escalation. Arbitration must not discard low-priority messages permanently: queue, summarize, or show them once primary risk resolves. The point is not never show alerts together, but never let everything speak through the same forced signal.
Applying it
- Define arbitration for simultaneous events: which interrupts, which merges, queues, downgrades to stable state, or appears in a later summary.
- Give the highest priority exclusive strong signaling in location, timing, and meaning; do not let independent components flash competitively.
- Group causal or cascading anomalies as one expandable event, with counts and a visible queue rather than a separate flash for each.
- Validate through multi-event drills. Users should handle highest risk first and still find deferred items; random order, global ignoring, or primary-task collapse indicates failed arbitration.