A guide to proper use of animation in UX | Hacker News Reader

A guide to proper use of animation in UX

uxdesign.cc

151 pointsjoshdance69 comments

Most animation is visual spam. The proper way to use animation is to NOT use it except when there is absolutely no other way to communicate something.[1] Material Design is fundamentally broken, because it tries to simulate physical motion, which is terrible for accessibility, so I don't think it's a good source to cite. Browsing the Web over the past few years has become an often-nauseating experience.

Edit: here is some reading for the downvoters:

https://usability.yale.edu/web-accessibility/articles/animat...

https://alistapart.com/article/designing-safer-web-animation...

https://simplyaccessible.com/article/animations/

http://accessibility.psu.edu/animations/

https://www.smashingmagazine.com/2018/04/designing-accessibi...

[1] except in gaming, where animation is expected and reasonable

Browsing the Web over the past few years has become an often-nauseating experience.

I'm often amazed by "average" users who seem completely unfazed when they visit sites filled with animated ads and other visual noise; but it likely does negatively affect their understanding of the actual content, since they are being constantly distracted by everything else trying to get their attention. I was a few times forced into this situation and had to physically cover the flashing banners and other crap with my hands in order to actually read the content. Even if it's only in my peripheral vision, I found it to be sufficiently distracting that I couldn't think about the text I was reading --- it was like being half-asleep. I could read each word, but everything felt like a blur.

This brings up my hypothesis that the reason a lot of users seem to not notice things in the UI or read warnings closely is that they're so used to being distracted by useless things and desensitised to it that they've accepted not being able to concentrate and think as being normal, which is quite unsettling. Many times I've had to help someone out with an unfamiliar (to both of us) UI, and the person I was helping would gaze blankly and not be able to find something that I instantly saw.

> "I'm often amazed by "average" users who seem completely unfazed"

They are unfazed because it's the norm, not the exception. Humans adapt. It's what we do. They see X and presume that's the way It's supposed to be.

This explains fashion. It even explains obesity. We look around and unless there's clear and obvious danger we conform, because conforming has historically been s survival instinct.

That "survival skill" is coming back to haunt us.

> "This brings up my hypothesis that the reason a lot of users seem to not notice things in the UI or read warnings closely is that they're so used to being distracted by useless things and desensitised to it that they've accepted not being able to concentrate and think as being normal..."

To be clear, I agree.

> The proper way to use animation is to NOT use it except when there is absolutely no other way to communicate something.

While it is true that some animations are excessive, I find that a certain amount of animation makes it less taxing to follow what is happening on screen. If big changes happen suddenly and without animation then I have to spend a little bit of extra time to figure out what moved where relative to where it was before and what was added, removed or changed.

Mainly I mean when changes happen automatically, like if something was dynamically added, removed or changed by the system. Not, say, if I initiated a page-down with my keyboard in mutt etc.

There's no need for animation to highlight changes. E.g. if a new item appears suddenly in a list, it can use bold text so it stands out. It can then immediately change to non-bold when you interact with it. This has the additional advantage that the change is obvious even if you were looking away when the item appeared. I've never seen a UI animation that was better than no animation.

The problem with your proposal is that the meaning of the "bolded" item is not immediately -or perhaps even with careful thought- obvious to anyone except the person proposing the idea, or the people who have already encountered the same idea beforehand.

We don't have a widely-understood cross-cultural shorthand for annotating an item's 'newness'. By contrast, we are all familiar with objects retaining their identity while they move.

When I minimize my window in macOS, it physically transforms and is sent to its new location. I know where the window I was working on went, and I as a person would have a good reason to be able to somewhat or even mostly comprehend this even if I hadn't ever seen macOS before.

With no animation, the large and prominent window I have been using would immediately disappear with no indication where it went, and the minimized icon in the dock would be highly likely to be overlooked. It would also look almost exactly like "closing" the window, and it would be easy for an inexperienced user to think minimizing a window and closing it are the same thing. I can assure you, this blurred conceptual distinction is not an imaginary problem.

Sure, I don't personally need this animation because I'm an experienced user already. However, in almost all cases it is fundamentally flawed to approach interface design by considering only the needs and preferences of experienced users.

Why does the meaning have to be obvious to inexperienced users, at the cost of making things worse for everybody else? We don't let untrained users drive cars unsupervised, and cars are much simpler than computers. I say that considering only the needs of inexperienced users is a fundamentally flawed approach.

I agree (and often use the same example of cars). I believe the reason is simple and sad - UI is meant to sell your product (or more likely these days, service). UI and UX panders to new users because that's who they have to impress; after purchase/subscribe decision is made, their feedback rarely matters (and there are plenty of ways to make them stick around, including exploiting sunk cost and network effects). Add to that the fact that regular users accept what they're given, because they have little experience to know things could be better, and you have a perfect recipe for UIs promoting sellability over user productivity.

Moreover, users tend to protect what they're given. It is not unusual to hear probably they know better against your criticism of ui, physical (like in cars) or on-screen. Or even worse, lower grade touch screen. People believe that manufacturers are gods of knowledge and engineering, because of that same marketing I think.

> I say that considering only the needs of inexperienced users is a fundamentally flawed approach.

Sure, I don't dispute that for a second. However, we seem to have moved from "UI animation is categorically worse in all cases" to something like "UI animation is categorically worse for moderate-to-experienced users".

A simple method to make animations less frustrating for experienced users could be to speed up the animation each time it's played. The first time, the effect would be made very obvious; but with continued use, it becomes so fast that it's instant.

Someone above said that animations are fine in games. But that is not actually true. I love fallout 1/2, for example, but it is pita to wait for slow-animated enemies to chase you or attack (fire geckos, I'm looking at you!!).

Frustration comes not from animation speed (except unreasonably slow), but from animations that freeze the interface. If you buy item and it jumps into a cart; if new item arrives and inserts into a list; if you change pages with < > buttons -- it is okay if it animates, but what is not okay is that you cannot click another button or simply scroll before the end of entire animation or at least slow part of ease-in curve. UI simply has response delays because of these tail timeouts. It feels slower because you cannot train yourself to catch a right moment on ultraslow ends of ease-in -- and so you tap twice to make an action or wait extra. (edit: remember how you quickly tapped on your phone to emulate g-sync with your finger?)

So, when you click buy, item starts jumping into cart, but you can instantly go back or buy/open another item. When you change pages and tap < twice, it just scrolls/rotates/whatever twice as fast. When new items arrive, space for them expands almost instantly, and then "new" animation goes, not interfering with your actions. If new items arrive continuously, debounce on 800-1500ms to prevent animations spoiling the entire interaction. And so on.

>I love fallout 1/2, for example, but it is pita to wait for slow-animated enemies to chase you or attack (fire geckos, I'm looking at you!!).

You may be interested in this mod: http://falloutmods.wikia.com/wiki/Sfall

I enjoyed playing Fallout 1 and 2 as a child, but I had very low standards back then. I now consider them unplayable without using this mod to speed up the animation.

Thanks, I'm using it (and a restoration project too). But some animations are still slow and annoying as hell, e.g. combat mode change or {un,}drawing a rifle when opening boxes and doors. I usually put one gun back into the inventory when in firendly location.

That's an issue with their implementation/use of animations, not animation in general.

Animations are (if done right) very useful in games. It's what makes you feel, for example, the difference between a weak damage (HP bar drops slowly) and something to watch out for (the HP bar actually shakes as a chunk is wiped out).

It also helps with immersion.

> When I minimize my window in macOS, it physically transforms and is sent to its new location.

That use of animation is okay, but to make your product accessible, just make sure that there is a way to turn off the animation. Easing effects and simulation of real physical motion are not needed. Apple's out-of-the-box UIs are unusable to me. My window manager (i3wm) has no animation.

Page-down should animate (quickly), it makes it much easier to understand how far the page scrolled, to follow the text and pick up reading on the next page.

When I press page-down I want to scroll a page down immediately, not wait 250ms before eventually arriving at a page down. Page-down always scrolls the same distance, so you don't need any hints about how it scrolled; you just move your eyes the same way you always do. If something is messing with the scrolling to make it act in a confusing way (I think Javascript might allow this), that doesn't mean you should use animation to help people understand the confusing behavior, it means you should fix the bug that allows the confusing behavior.

1. 250ms is not quickly, on Firefox Windows a default page down animation is closer to 50-100ms (estimating), it's so fast you hardly even realize it's animating. There are websites that override this with their custom slow animations - those type of page downs suck yes, hence my inclusion of quickly in my first post.

2. Page down doesn't always scroll the same distance when you hit the end of the page, unless you have over-scrolling like some code editors do but websites and browsers don't.

For media naturally represented as pages, like PDF or images, and when one whole page fits on a screen there's a different story though, there it's nice to be able to jump pages instantly without animations, because the page boundary itself can be used as reference. But for reflowing endless wall of text media like web pages and code you kindof need to follow your previous position as a reference.

> When I press page-down I want to scroll a page down immediately

You are among a tiny minority and come off as trying really hard when insisting that animations are useless. Very few people want a 0ms screen flash on page scroll.

There's no "screen flash" with a non-animated page scroll if done correctly. I'm not complaining about double buffering.

Nobody is talking about double buffering. We’re talking about non-animated page scroll, which instantly replaces content with other content, a preference that you share with very few people.

Citation of a UI study showing that preference? And was it with inexperienced or experienced users?

There are many other people in this HN discussion expressing a preference for fast interfaces over animated interfaces.

Literally nobody thought your next response would be anything but “CITE???!?”. You and your kind are entirely predictable: Disguise a minority opinion as an objective truth then shift the burden to anyone who calls you out.

Nobody in this thread besides you has expressed a desire for a 0ms scroll. Nobody.

False.

JoshMnem: "The proper way to use animation is to NOT use it except when there is absolutely no other way to communicate something." (scrolling is communicated by the fact that you pressed the page-down button)

yorwba: "The first time, the effect would be made very obvious; but with continued use, it becomes so fast that it's instant."

iforgotpassword: "I usually try to disable animation entirely"

th0m4s: "On my last phone I turned the animations off complete"

kqr: "I turn off animations entirely on my phones, and when watching others use their phones I'm always baffled by how slow they feel."

devgutt: "You can also don't use animation at all. I really don't like animations even when used correctly."

Ah yes, that well-known and beloved feature of the iPhone: the page scroll button. This is pretty sad and weird of you dude.

Only half those comments were about phones, so even if those posters would make an exception to their stated animation preference in the case of page-down, "literally nobody" is false. If wanting low-latency scrolling is "sad and weird", why does Mozilla expose this option in Firefox's main Preferences page? They have no compunctions moving genuinely rare preferences to about:config or removing them altogether.

I'm referring to accessibility, not just aesthetics.

+1 for mutt. Text-first is the way to go.

I agree, I built a small UI for work and used animations (with 200-300ms transition times)

Then out of curiosity I disabled them all on a release and the feedback was pretty much the same except I got a few "it's much faster" responses.

Fwiw internal business application not consumer but lesson learnt.

Personally I very very rarely think animation is an improvement.

Except where you are using animation to cover loading times and at the end of the animation the UI is ready perhaps.

Yes, agreed.

What I've exexperienced is that the team / client falls in love with senseless doodads and whatnots. They lack they empathy to ask:

- "If I saw this for the first time, what would I think?" as well as

- "If I had too keep seeing this - knowing it's not helping me - what would I think?"

The user needs is a distant third to the team's wants and lack of empathy. You'd think after Adobe Flash being abused we would have learned? That doesn't seem to be the case.

You provide five links to back up your claim that

> Most animation is visual spam. The proper way to use animation is to NOT use it except when there is absolutely no other way to communicate something.

But they don't back up that claim. They don't come close to backing up such a sweeping statement.

Indeed each of the linked articles ends with a paragraph more or less stating that animation is useful, but it should not be excessive and that one should give users options to opt out or tame them down if it is necessary.

Not all animation is bad. Most animation is bad. My comment is a guideline to help people figure out when something actually needs animation. If you're adding animation because you think it's "delightful", it probably is making the UI unbearable for some users.

Examples: animations that simulate real physical motion or cause flashes to appear on the page (like Material Design) on sites where users are supposed to read text are nearly always a bad idea from an accessibility perspective. Subtle animation that makes a certain part of a user interface clearer when the user is not reading content is okay if a lot of moderation is used. Don't think "how can I add animation?" -- think "how can I remove animation so that this still works?" Designers should also test the interface and make sure that it still works with animation disabled.

"Numerous researches have discovered that optimal speed for interface animation is between 200 and 500 ms"

They are still too slow.

If you have an android phone, go into developer options and turn animation scale to 0.5x (animation duration is halved, so they are twice as fast). You'll be amazed how much faster the OS feels.

Did this on a jailbroken iPhone once and I found the effect over time to give me the feeling of being frantic. I think the speed of these animations are chosen for various reasons, and serenity is an important part of how you feel in general while using the device. Just like you would need to teach a self-driving car not to accelerate too fast, your mind has similar expectations about what’s a comfortable pace.

Interesting. I usually try to disable animation entirely, but that sounds plausible. Did you try to disable animations and did it make a difference?

On iPhone the animation speed was/is there to help give apps time to launch, for example, so while the app is launching you see the animation and it hides the fact of the app launch time. If you disable the animation / speed it up, sometimes you run into apps that you notice are unresponsive until they have finished launching. You can’t disable the animation completely, but you can reduce the animation time to as low as it can go. This was a while ago as I haven’t been jailbroken for a few years now. What I describe above is my experience with “disabling” the animations.

I think there are two factors causing developers to choose too-long animations:

1. Pride. We worked hard to build this animation, and the user is damn well going to watch it! Because they'll definitely like it as much as we do. (This is an example of not putting the user first.)

2. Inaccurate conclusions from user testing. I think users go through two phases. Early on, they don't know their way around the UI and can't navigate quickly. Plus the animations are novel and they're still enjoying them as eye candy. So for two reasons, they don't mind slower animations. As they use the UI for days or weeks, both these factors flip. You start navigating the UI by muscle memory, and novelty is gone. Thus you can only find the optimal animation speed by testing how a user feels about it after weeks of usage, which isn't how user testing is usually done.

Fun fact: The default Start Menu pop-up delay in Windows XP was 400ms. Once of the first things I always did was reduce it to <100ms on every new machine I used.

While for you (and most of us here) that delay was too slow, it also served a purpose -- users who can barely hold the mouse would diagonally go from parent to e.g. 4th child element too slowly and lose the submenu on the way, because it would disappear after the same delay if cursor was not on parent-or-child. Otoh, windows devs could arguably make it independent and the problem could disappear.

In short, with 100ms you have to go ¬-shaped and with 400ms you can go \-shaped.

Same here, I miss the days of Microsoft Windows XP PowerToys

Now almost nothing UX-based is configurable

On my last phone I turned the animations off complete, and then with my latest I'm waiting until I want a new phone to turn them off to give me a breath of fresh air.

In contrast, on early Android 2 phones, I did the opposite -- increasing animation time to 2x. It felt smoother. I'd argue, that animation can cover unresponsiveness of application, and improve perception of performance.

You need very high frame rate to make fast animations feel smooth and natural. I suppose older Android phones had quite low frame rate and required slow animations to feel smooth.

Imagine moving a window 500px over 100ms in 30FPS, that is just 3 frames! It means your window will jump 166px per frame. On average resolution display this translates to a few centimers jump per frame. Not very smooth.

Why not just tell the user something is taking long to calculate or load (maxbe due slow network)?

I turn off animations entirely on my phones, and when watching others use their phones I'm always baffled by how slow they feel.

However, I've notice Android occasionally uses animation to convey things to the user, so I'll try your idea of halving it.

Yeah, I initially turned off animations on my phone maybe a year ago or so. Works great unless you occasionally play games (QuizUp for example!). Then animations needs to be able to show something, otherwise the game UI is laggy. Setting the animation scale to 0.5x fixed the issue, and still my UI feels very snappy.

If you turn one of the three animation settings to 0 on a Keyone phone, the popup to choose charge/fast charge/data connection when connecting to USB doesn't appear and there's no alternative way to select it.

I forget which but stuck with 0.5x for that reason.

Best tip ever! My phone feels snappy again!

Someone please explain to me, why we have all these animations nowaways other than 'because we can' - what purpose do they actually have?

It can make it clear what happened after user interaction.

Was an article added to the list during a refresh? Did the correct item actually go away after deleting it? It's easy to miss updates without animations. Even if you don't miss the update an animation will provide an additional confirmation that the correct thing happened.

On touch devices there's also the fact that animations mimic input gestures. Imagine swipe-to-delete without the item actually moving, that would be super awkward to use...

Of course, like anything, it is possible to mis-/overuse.

I think one of the original ideas was to "show where something had gone" - when you minimised a window, both on Win XP (if I remember right) and OS X, it wouldn't just "disappear" but would animate down to the taskbar / launch bar. That way you get fewer people confused about where their window went.

I suppose you could just about make a case that when you have a modal element like a form to edit an item, if it takes up most or all of a small screen then having an animation to expand it makes a visual difference between "you're now on a new page" and "this form has popped up, but when you close it you'll be back on the page you were before - it hasn't gone away".

They may be useful for beginners.

Otherwise mundane app will look much more sophisticated. It can also make the experience more fun, which makes it easier to addict the user. Just think why slot machines are so colorful and flashy. Would they have same appeal if they would look and behave like a parking meter?

I would not mind animations if:

- they would not slow me down - I'm faster than them and I have to wait. It would be interesting if animations would be faster or disabled when user interacts faster than some threshold.

- they would be interruptible

- they would not cause performance degradation

- they would not add more bugs. As with most asynchronous actions they are not properly tested when things happen simultaneously or are interrupted.

An automatic option to disable animations if the user is faster certainly seems like an interesting one.

Though it may be interesting just to let users choose whether they want to see animations in general. That'd solve a lot of problems here, and additionally give you some metrics in regards to how many people like the things.

> Though it may be interesting just to let users choose whether they want to see animations in general.

That would be the ideal solution. Make a setting in browsers and devices where users could turn off all animations. Designers could test their UIs with the setting on and off. Everyone would be happy.

Tricking the managers into thinking that the designers are actually contributing value.

I agree with your sentiment - there really isn't a use for animations in the context of most UI.

Except when there is - for example, first-user introduction/onboarding segue's, which guide the users attention around the app by highlights/flows, perhaps overlaid in a fashion that makes it clear that the overlay itself is impermanent (by fading it) .. but only if its got a real purpose, as you state.

I appreciate this degree of assistance in learning a new app personally, especially on mobile.

Animating the UI is what we want the user to do anyway - and many users are very animated about how they use their apps (i.e. selfie culture), so using animation to train the user to use the app properly, such that the user animates back, seems obvious.

Probably because playing with CSS3 is fun and most people have little aesthetic restraint or awareness of accessibility for visual motion sensitivity.

- huge sticky header

- some sticky like and share feature on the side

- weird contrasts which are not comfy for the eyes

- too many animations (this might not count at all)

This sums up most things I hate about modern web pages actually. Why use all these things? Especially huge sticky headers are just super annoying. Please don't use them. Thanks.

Honestly, I never got the hatred for sticky headers and navbars. Yes they're ridiculous if they're too big, but having the menu in an easily accessible place is usually a good thing. I mean, there's a reason mobile apps and desktop programs have menu bars rather than some other design solution.

Does that mean they're always done well? No, things like Medium's 'please join our service/subscribe to see more from [name]' are ugly and pointless, and nothing more than a visual distraction. Sticky ads (or for that matters, ads that try to take over large parts of the page in general) are also terrible for obvious reasons.

And yes, a lot of 'fancy' design techniques just hurt the usability of the page, as seen with anything that hijacks scroll/overuses parallax effects.

But I don't think sticky elements are bad in of themselves, and they can make the site easier to use if done correctly.

The issue is that I can already scroll to the top to see the navbar, and interacting with it is a small fraction of what I do on a website. In fact, when clicking into media (like a news article or blog post), I'm not even going to interact with it once.

Yet I must endure all of its downsides the whole time I'm on the page, from taking up precious screen real estate to the even more annoying approach of popping in and out when I scroll up/down one pixel.

Just think of how seldom you interact with the navbar of most webpages and why it then needs to follow you around the page in the off chance that it's more convenient than just scrolling to the top.

So far I remain unconvinced that it can be done well since the entire premise is flawed: my device already gives me a one-click scroll-to-top shortcut. Like inertia scrolling, it's a bad smell when every website thinks it has to bring its own implementation of feature devices should already have.

Sticky headers are an example of something breaking the normal scrolling behavior as mentioned here: https://news.ycombinator.com/item?id=18045207

A sticky header changes how page-down works, so I have to adjust my eye movement to compensate. Because of this I consider allowing web pages to set sticky headers a browser bug. Browser devs disagree, but at least I have the tools to disable this kind of user-hostile behavior.

> Web animation is treated in a different way. Since we are accustomed to an almost instant opening of web-pages in a browser, we expect to transit between different states quickly as well. So, the duration of web transitions should last about 2 times shorter than on mobile devices — between 150–200 ms.

So, let me get this right. Animations on th web should be fast, because we're used to the web being fast. Animations on mobile should be slow, because we're used to mobile apps being slow?

> In other cases, the user will inevitably think that the computer freezes or has troubles with the internet connection.

But on mobile, that is normal?

If I ever need to write a mobile app, I think I've got an idea to make my app feel a lot faster than its competitors.

They talk so much about Material Guidelines. If only Google themselves followed them, https://grumpy.website/post/0QcZsPBJl

You can also don't use animation at all. I really don't like animations even when used correctly.

Overall a good article. I learned that speed should depend on the screen size, something I had never thought about.

What I don't understand is that we are not supposed to use motion blur. Is it only because it's hard to do right?

Also I always thought that the "when the moving objects transform their size disproportionally, they should move along the arc rather than in a straight line" thing felt wrong. Anybody else feels this way?

Gratuitous animation like the ones showed in this article are not "proper".

Animation can serve a communication (and hence UX) purpose, but when everything is animated it defeats its purpose.

It also has to be noted that the more animations there are, the better GPU you need, which can create a false need of a new device.

The author of this article should read some guide to proper posting animated images on the web. Having examples implemented in CSS would probably be easier on resources than those awful fullHD GIFs.

> Please don't do things to make titles stand out, like using uppercase or exclamation points, or adding a parenthetical remark saying how great an article is.

--https://news.ycombinator.com/newsguidelines.html

Reply on news.ycombinator.com