The experiment was not 'flat' vs 'not-flat'.

It is weak-signifiers vs strong-signifiers.

They have a hypothesis that states that users will be able to accomplish their task with less effort given strong signifiers.

They tested the hypothesis with two versions of the same page: one with weak signifiers and the other with strong signifiers. Both versions were created to _test_ the hypothesis. To either confirm or disprove the hypothesis.

They measured user eye movement activity and produced heat maps.

The heat maps confirm their hypothesis. With weak signifiers the users spent more time searching for the elements they wanted.

How is this not scientific?

Like all good science the conclusion is not 'flat bad, skewmorphism good.'

No, it says whatever you are doing ensure you provide users with signifiers.

So the final result is not a false dichotomy of flat vs non-flat but a useful set of guidelines: "Early pseudo-3D GUIs and Steve-Jobs-esque skeuomorphism often produced heavy, clunky interfaces. Scaling back from those excesses is good for usability. But removing visual distinctions to produce fully flat designs with no signifiers can be an equally bad extreme."

Thanks for your response. I agree that the basic process used here was scientific, but the way it was framed was misleading.

> The experiment was not 'flat' vs 'not-flat'.

> It is weak-signifiers vs strong-signifiers.

I agree completely.

The title of the article is "Flat UI Elements Attract Less Attention and Cause Uncertainty".

Read a few of the comments here and you'll quickly see how the article is being misinterpreted.

Somebody needs to publish: "Clickbait articles attract attention but reduce credibility"

Agreed. The investigation was good but there are issues with the reporting.

Maybe they did choose things to confirm their hypothesis, but these aren't some mom and pop shops nobody recognizes, these are international well known brands - Hertz, Barnes and Noble, Nest etc...

If you read the entire article, you can see that the point was not Flat UI is bad, but rather Flat UI done bad can strongly hinder UX.

It makes no sense to just pick sites at random, you want to check the extremes so the evidence will be significant. To put it another way - you choose the most obvious candidates so that if you end up without a significant difference you can be positive that the hypothesis isn't correct. Following research can be more nuanced.

> you can see that the point was not Flat UI is bad, but rather Flat UI done bad can strongly hinder UX.

Yes! You're absolutely correct, and that was ultimately the point I was trying to make, if poorly.

My comment was more of a response to this post's title and a handful of particularly torch-and-pitchfork comments.

Right, they aren't "nobodies", they are a shop with a long history of creating biased experiments to validate their initial assumptions. There's a LOT of material available on the net regarding their "methods", a simple google search will do.

First, I'm pretty sure that's ad hominem.

Second, I said that it's probably ok that they chose the sites to validate their claims, saying they do that often isn't a counter argument.

Third, twisting my word so blatantly makes you look childish.

No, it's not ad hominem. It's me expressing my dislike of the Nielsen Norman Group. They have contributed to the state of the art but they also tend to overblow their claims - famously they said "people don't read on the internet" which is absurd, a slew of more serious research on the subject shows how reading and scanning patterns are not substantially different on screen or in the press, and that users on the internet tend to read MORE than people reading newspapers.

Not to mention the fact that we're interacting in a text only site...

Third, what I've done is not "twisting your words" but connecting to your reply using it as a jump off point. I don't remember the name of that rhetorical figure but it's pretty common.

All in all I don't think there's anything you should be taking personally unless you have some vested interest in the author of the article (haven't checked, are you the author?)

This. Even Jonny Ive can't ship a "view all colours" link like the one on their sunglasses page. It would have been better if they were comparing material design or metro rather than their interpretation of flat design.

> "Explain to me how adding an underline to a link is somehow not-flat? "

"Hey, I'm definitely a link"

vs

"I may or may not be a link".

As the study explains, link colour difference was enough to balance the scales. It's when link text looks the same as normal text you have a problem.

Flat design can go too far and everything looks the same, requiring more effort from user to figure out what's what, and also remember what's what for next time. I find this with iOS, my memory of the interface is quite poor ever since they introduced flat design. I'm only an occasional user of my iPad, but still... the flatness is irritating.

> Flat design can go too far and everything looks the same

Flat design, done well, makes signaling easier because there's literally less visual noise. Bad design is just bad design.

Sure, if there's sufficient contrast within your flat design, I'm fine with that. Lots of good examples of this in 'material design' where buttons are easily recognised, with nice colourful backgrounds. When buttons get wishy-washy grey in some attempt to "blend in", then we have bad flat design.

>Most of the "flat" designs strike me as being created or chosen to confirm the hypothesis.

And yet, it's totally representative to things I frequently see on the wild (if not better than most, including Google's crap).

Exactly. Here are some bad designs, which happen to be "flat". QED