Hawaii's recent missile crisis - bad UX/UI


#1

So some of you may have been aware of the recent missile crisis in Hawaii.

For those of you who may not be familiar with the story: https://medium.com/ux-immersion-interactions/the-hawaii-missile-alert-culprit-poorly-chosen-file-names-d30d59ddfcf5

The person in charge had to test the drill missile message from the links above…not the most obvious…

Low and behold, the reason for the mistake - bad UX/UI design

Would love to hear all of your suggestions to improve this design and how a crisis like this could be easily averted if companies were willing to invest in good design - particularly with something as crucial as informing citizens of an incoming missile


#2

I think there may be a few options. The first one that comes to mind, assuming they want to have everything on the same screen, is to put the test messages and the “real” messages in different groups.

The operators would generally stay away from the ones they need to use in “real” situations. That would mean the operator would have a visual clue that they are doing something different than testing.

My second idea would be to use colors. The problem I see with that solution is that the operators would get used to the colors and I think that would render them less useful.

Maybe I don’t understand their process though.


#3

Great suggestions @Piper_Wilson! I agree it’s not obvious at all. Perhaps an additional line of check would be an input of a password just to doubly make sure that it’s not clicked on accidentally.


#4

I mean it’s hard to know just how to re-design the page, without knowing anything about the other page elements. Shucks, we don’t even know the whole list on this page. However, I think we can say a few things for certain:

  • The link names are not human readable.
  • Actual warning messages are intermixed with test messages.
  • Messages are not divided by type or severity.

In my view, this is a lesson in the important of having a well-established page hierarchy. At the very least, this means re-arranging the elements in this list into more intelligent groups by adding meaningful headers, dividers, and colors to assist with differentiation (such as watches, warnings, and tests). We’d also definitely want to re-name the items to make them more easily understandable.


#5

I’d lean toward removing the actual alerts from the drills. If drills are truly performed that frequently, it might be better to move them completely out of the way.

They could possibly have the test drills listed out, then display an action button to send out an real alert which you would have to first select the specific one and then also confirm. My take would be that this is probably a scenario where you want to add a couple of extra steps or ‘friction’ to ensure that the user completing this task actually is intending to. Obviously not too many as if a real alert is needed, it’ll want to be completed in a hurry.

I also would second others’ suggestions of updating the link names, and sorting by severity, etc. Basically any of these will help :joy:


#6

I feel like something along these lines with a confirmation popup after selection is a good first step, though there’s definitely more to do.


#7

Interestingly enough… the problem may not have been caused by bad UX, but rather poor communication and an under-performing employee who mistook a drill for an actual threat.

The report finds that the false alert was not the result of a worker choosing the wrong alert by accident from a drop-down menu, but rather because the worker misunderstood a drill as a true emergency. The drill incorrectly included the language “This is not a drill.”


#8

Wow that’s pretty awful. Whilst it still is seen as poor communication and employee underperformance - it still doesn’t deter from the fact that the screenshot from the software to deploy the drill warning wasn’t the most intuitive. Obviously not knowing the entire story, we should still be focusing on the UX/UI side of it.

Thanks @dougcollins we must never forget that there are other contributing factors - one which caused a guy to lose his job and two others to resign.