I'm surprised to see so many people reacting so negatively this claim. This research doesn't claim that flat UI is inherently unusable:

If your designs are already strong, any potential weakness introduced by flat design will be mitigated.

It does claim, however, that by merit of flat design being flat, there are less signifiers for users to grab onto, and that a lack of strong signifiers can cause users to hesitate. This is certainly true in my experience.

Think about the old distinction between inputs and buttons—the former had a shadow on the inside, the latter on the outside. This made it evident which was which, without hovering. In flat design, both of these inputs can appear identically, and often do. Popular design systems like Material do away with borders and background altogether, which leads me pretty regularly to question whether or not something is clickable.

It isn't that your app will be totally borked without drop shadows or gradients or whatever else. It's that every time you remove a signifier, you subtly increase the amount of mental effort that's required to use your interface. And since cognitive processing and willpower draw from the same pool of resources, you are increasing the likelihood that your users will make poor decisions both inside and outside of your app.

This isn't something that you can learn yourself out of. As designers, we are probably more skilled at picking out what can be clicked on and what can't, but here are two important truths:

1. We aren't designing for us, the strongest 5% of computer users. We're designing for the 95%.
2. Even if we were just designing for us, flat UI still induces subconscious hesitations. Why subject ourselves to that for an aesthetic?

Design is intent. When you're making a visual change, it pays to know the research.

lol, no.

It's all about HOW you use flat design. The examples they gave are all horrendous and they clearly have no clue what flat design is and how to use it. They're all outdated corporate designs.

Also: Comparing gradient buttons with border-only buttons is not accurate. The flat equivalent of a gradient button would be one with a solid colour. A text link that is colourful versus a text link that is the same colour as the text is not flat design either. What? Text links are flat by definition. All they're comparing there is the efficiency of colouring your links versus not colouring them, it has nothing to do with their argument.

Maybe if this article provided proper data with designs by actual design industry professionals, it would be more relevant but, as far as I can tell, it's just pointless, whiny, "I'm afraid of change and I want websites to stay how they looked in 2001" non-sense.

So, flat design is somewhat more difficult to use for the first time. I would love to see a see a study that investigates if people can learn what a flat interactive element looks like and if that'll speed up the remaining part of the task.

I don't really get the point of this kind of research. That's not about flat vs. 3D and that's not really an open-minded scientific approach, I think.

I've always felt this was obvious. Nice to see research to back it up. Flat design can also cause issues with those that are visually impaired. Where not a lot of contrast exists to properly differentiate elements.

A nice example of a flat design failure is on this very website. The "New Story" is very much not obvious. To use flat design the interactive elements need to stand out, and not fade into the background. I'd personally make the button yellow to draw attention to it. The blue on blue is very bad.

this entire thing makes very little sense to me because:

1. the result are quite predictable, as this proposition:

"every time you remove a signifier, you subtly increase the amount of mental effort that's required to use your interface"

clearly doesn't need backup from a study.

1. The data is just wrong as there aren't real definition of what "flat design" is. (ex: in their test the "flat" version had the link underline removed...how un-flat is an underline ?)

In the end, this makes up for yet another catchy headline article full of BS.

Aren't we past Flat design now?

IMO the balance swayed back a little in 2016 with the introduction of subtle shadows, gradients and less use of outline icons.

Flat with a pinch of skeomorphic AKA Flat-o-morphic design (I'm joking with that term)

The research is deliberately misleading. They used a full colour '3D' button, with their assumption being that making the button 'Flat' will reduce engagement. They didn't just make the button flat though, they removed all colour entirely. It's not a fair test.

It's not flat design, it's BAD flat design. The title is misleading.

Can nngroup.com explain or research on how Snapchat UX works? :D would be awesome to see.

I'm curious on how UX Researchers from Google Material, iOS Apple Design, Microsoft Design, Facebook Design and Airbnb Design and even Dropbox would respond to this? These gods are all flat.