The Batesian Theory of Warning Dialogues

Appprops of the work I’m doing on the MailOnline CMS, I think the following is a thing. And I’m going to call it a theory:

The probability of reading a warning dialogue decreases by the square of the number of other warning dialogues you have seen before in that session.

This might be useful in explaining why adding messages to help people understand how to use a system is usually futile. And if people end up simply hitting “OK” without reading the warning, it may quite possibly also be counter-productive.

A dialogue, yesterday.

See also the Jef Raskin adage, “Whenever you sit down to write a message, consider first whether the system can be designed such that the message isn’t necessary”

And see also Bruce Tognazzini’s advocation “Always have an undo“, partly for this reason.

There are of course other considerations in all this, because it’s UX after all. But as a principle I think it’s worth noting. The fewer the warning dialogue you permit to exist in your app, the better your app will be.

—

Some possible examples to ponder

A $444 billion dollar mistake? (2024)

“The FCA said the trader was able to manually override a pop-up alert, without being required to scroll down and read all the alerts within it”

Hawaii false missile alert (2018)

“He clicked the button to send out an actual notification on Hawaii’s emergency alert interface during what was intended to be a test of the state’s ballistic missile preparations computer program and then clicked through a second screen, which had been intended as a safeguard, to confirm.”

Always have “undo”.