I stumbled upon this site when browsing Stack Overflow and got interested. I started reading the top questions and at first was really excited, but halfway trough page three all I could think was "Seriously, do we really need to nitpick this much?" or in less harsh terms "Almost every choice only has a small majority over the opposite option, why go through the trouble of picking one or the other?".

n.b. I'm a web developer, not developing for a specific OS.

Some examples:

OK/Cancel on left/right?

Windows and Apple use the opposite. So whatever you choose, you will always have a large group that has the opposite of their OS.

Should error messages apologize?

Some say they should be more like a sales/marketing guy would handle the situation, while others state the message should be short and descriptive.

Should a toggle button show its current state or the state to which it will change?

Again, some say Apple has it right with the on/off button where "on" is blue, while others state it's unclear. Either way, a lot of users will like it or dislike it, no matter what you choose.

Why might right aligned field labels be better?

Apparently the right-aligned labels have a lighter cognitive workload on the users. I'm sure its correct, but do we really notice this that much? Isn't this a bit like car sickness? The first computer users might have issues with things like this, but the current generation has been looking at screens longer then they have seen daylight, so aren't they just used to it?

To conclude:

I believe that users can adapt to an interface quite easily these days because of the all the differences in many applications and websites. They are used to it. An Apple users wont be frustrated when the cancel button is on the right and a Windows users can understand an Apple toggle button just fine the second time he uses it.

Now I'm sure that some design choices could lead to people leaving your site quickly (The rotating cube), but can we rely on users to adapt to your site if you offer something of value and a reasonable interface? In other words: Isn't the content of the site still the reason why you have visitors/conversion?

If you say the answer is no: Seeing a lot of differences come from Windows/Apple user, would it be useful to detect/request the OS and show a website that is based on the OS? Or go as far as to allow users to change settings like "left/right align" or "long/short errors".

Update I know that the UX is very important when you develop something. But looking at the top voted answers, I can hardly imagine that these changes would have a positive/negative effect on revenue/vistors. E.G. Could changing a left/right aligned label really make a difference? I'm hoping for some experienced based answers that show minor changes could have big effect.


migrated from ux.stackexchange.com Jul 10 '14 at 14:21

This question came from our site for user experience researchers and experts.

@BennySkogberg > I read the help center and since it stated "slightly more subjective" with some guidelines, I thought this would be OK. I guess the question boils down to "Is it worth spending much time on perfecting the UI?" and hoping for expierenced based anwsers like "We changed our website with some minor adjustments in UI, and noticed an increase in revenue". As I never felt the need to think about these things, I'm wondering if people actually noticed changes based on these choices. –  Hugo Delsing Jul 9 '14 at 12:02
You've chosen examples which have in common that they are smaller aspects with a simple binary choice. And of course the 'correct' answer may indeed be arbitrary and decided by tradition. However to generalise and say that users can adapt to all UI/UX is not supported by your examples. You've cherry picked the evidence and it still doesn't support your conclusion. –  RedSirius Jul 9 '14 at 12:03
@HugoDelsing That is a much better question to ask, so I suggest you rephrase the question and heading accordingly - and you will have the chanse of getting quality answers. I like the fact that you challange the community! :-) –  Benny Skogberg Jul 9 '14 at 12:07
But it would invalidate the answers already given, so changing the content of the question that much, should be rejected as an edit :) But I'll update it a bit. –  Hugo Delsing Jul 9 '14 at 12:10
This looks like a rant, not a question. –  Lego Stormtroopr Jul 10 '14 at 4:44
@LegoStormtroopr I guess it did started as one. Because reading the top questions I did get a feeling that all these small choices that were put to discussion didn't really matter. Luckily a valid answer was provided and changed my perspective, which made me question every site I ever build. And I guess that's the ultimate goal of this community, to make developers/designers think about the details and not just the basic "common sense" guidelines? –  Hugo Delsing Jul 10 '14 at 6:45
Collectively these things do matter. For example, if you are familiar with Network Solutions' web hosting packages, their web management interface is a UX nightmare because of widespread and consistent mistakes of the nature you are describing. Personally, I found it so hard to parse and navigate their site that the resulting frustration (even after 10 months of usage) left my tolerance for service problems very low, and I dropped that account in under one year. I very likely would have been tolerant of service outages had information organization and management been less frustrating. –  Jason C Jul 19 '14 at 22:04
Will users also adapt to your slow loading website? See my answer below for further explanation on this. –  David HAust Jul 21 '14 at 0:01

9 Answers 9

up vote 33 down vote accepted

To respond to your update: these things DO matter, and they can have a VERY large influence on the revenue. There is even a version of user experience testing called A/B testing which shows certain users slightly different versions of the website, or even a preview of an upcoming version on the live site to see what the result is. For example, Google does this with Youtube. If you're lucky enough, you can get a sneak peek of what youtube has planned.

http://unbounce.com/a-b-testing/shocking-results/ is a list of 12 of these A/B tests that show that it can be surprisingly effective, with differences up to 450 % in effectiveness. Google can give you a lot more of similar results.

@HugoDelsing, for a local SE example: If you come to an SE site in "incognito mode", you'll see a box at the top with a short summary of our model, and a button that takes you to the "Tour" (sometimes called "About") page. The current text there "Take the 2-minute Tour" was 85% more effective in A/B tests than the exact same box with a button labeled "Learn More". –  Jaydles Jul 9 '14 at 16:00
And I did take the tour, because it was only 2 mins :) –  Hugo Delsing Jul 10 '14 at 6:37
@HugoDelsing For another local SE example, they've been A/B testing the job postings ad on SO for a while now. Assuming you have AdBlock off, you will see either a blue or an orange color scheme and title text for this pane. Not sure what test results so far have been. –  Jason C Jul 19 '14 at 22:07

To understand and apply good UX is to understand and work with human nature and human biology (and all it's evolved quirks).

Our users are not lab rats, running around a maze and learning to push complex combinations of buttons so they can get a "treat". They are people that want to accomplish something that is important to them, and it is our job to help make this as easy as possible.

If we don't pull out all the stops understanding how to make a website or application easy and friendly to use, then someone else will. If you want to keep your customers you should treat them well.

When you have a massive customer base, even tiny "nitpicking" changes can translate into bigger profits for your business.

It's always good to keep "the principle of least astonishment" in mind when designing an interface. People like using things that work in a way that they expect to (from past experience), and don't like making mistakes or having to learn by trial and error. For "OK" and "Cancel", for Windows you do it the normal way for Windows, and for Mac you do it the normal way for Mac. Users are happy. Documentation might require two pictures instead of one, if you want to be thorough. –  Phil Perry Jul 9 '14 at 17:28
+1 for tying it all back to commercial realities. –  edeverett Jul 11 '14 at 16:02

First, thanks for contributing with your very pragmatic perspective on UX.

You are right, it is a messy world with many contradicting practices and sometimes it may seem tempting to just give up and go with whatever you feel like. Which, in truth, most UX designers do all of the time.

Whenever you face a paradox like those you mention, there is no great answer. But if you have done your user-research and have an idea of what your users have come to expect, combined with a knowledge of where the market is heading in terms of trends, you should be able to make a somewhat justified decision.

You are right, content is king. That's where you start and there are numerous examples of horrible UI-design that still manages to attract hordes of users because the content is truly unique (Snapchat, Facebook (especially in the past), Hackernews are examples that pops up in my mind). Same goes for most enterprise software. UX might be horrible, but users have no option but to use it anyway, so they figure it out and adapt.

About assuming that "users will figure it out" whatever you do. Usually, if you ask a developer to build a frontend, he will build it the way he would prefer to use it. Problem is that he is heavily biased and probably much more tech-savvy than the average user. That's why we test with real users.

Yet, companies are increasingly willing to pay big bucks for people like us, nit-picking on tiny little UI details. Why is that? UX has become a major differentiator all of a sudden. People have come to expect all software they use to be aesthetic and pleasing to use.

In theory I suppose you could make your websites responsive to what platform users might be using, in the same way as interfaces are adapted for different design patterns on Android and iOS, but in most cases it is not feasible to maintain many simultaneous frontend interfaces, but for smaller OS-specific traits such as having confirm on left/right I think it would make good sense to have an OS specific override. Unfortunately I don't think it as easy as just differentiating between mac and pc users on the web. There are many contradictory practices within each platform as well.

Excellent point on UX being a differentiator. –  Tim FitzGerald Jul 9 '14 at 12:17

Franchesca answers your global question very eloquently. But I feel in reading your question there's something else bugging you. Am I right in suggesting that you find the examples you give are too nitpicky? Their answers too measured and ambiguous?

If so, Something tells me you're not going to appreciate my response, but here it is: it all depends.

  • Questions arise big and small, and depending on the maturity of your product and the feedback you receive (comments from the client, user research conclusions, etc.) you may get more granular questions.
  • It may also be a sign of the maturity of this StackExchange if larger, more black and white questions have already been discussed and put to bed. I was so proud I had my first question to post yesterday ("Why can't we undo the call to a floor in an elevator?") only I to find the question had been asked and answered satisfactorily two years ago.
  • There is often no one right answer, and the answer for you depends on context. Who is your audience? What are their goals? Your goals? What medium are you communicating through? If answers can appear neutral or undecided, it can often be to present the options to let you decide which fits your situation.

At any rate, I hope you stick around. I'm relatively new to the site and I come from a developer background, as well, and I find it great that we have a place to ask questions that challenge us and at times our preconceived notions.


If you were making a website, and you contemplated every single detail to the level often seen on UX.SE, it would take several years and probably damage you as a human being. That doesn't mean this level of discussion isn't useful to working practitioners, though.

If you do something twenty times a week, every week, then it's worth taking an hour or two some day to really think about that thing. Especially if it's something you always had nagging doubts about. You don't have to sweat all the details all the time, but someone needs to sweat every detail at least once, or else design could only progress at the glacial pace of natural selection.

Quantitative measures like revenue and A/B testing are important, but they don't drive design. They measure which design is better, which is of little use if you don't have a better design for these tests to detect. And developing better design requires meticulously taking stuff apart to see how it ticks.


So it looks like we have several extremely interesting perspectives or points of view that, far from opposing, draw the "UX design golden polygon" where each of us wants to be :

  • smart enough users : users have shown they are smart enough to adapt to a variety of efficient and self speaking designs
  • several coexisting bases : Windows and Mac, iOS and Android, Facebook/Twitter all work and have contributed to building large user bases' habits that are givens of the equation
  • a sum of tiny bits can make a difference : removing many largely unnoticed small sources of friction sometimes results in a very different global experience
  • let's bother for what matters : significant changes in business and revenue cannot derive from insignificant changes in the users' experience

Most of us probably feel at home near each one of the above points. We might want to add another one that reminds us that UX design, although building itself as a science, remains some kind of art :

  • "new can be good" : all the things that users love and are comfortable with were once new.

Maybe some "shoulds" could be translated into "This thing proved effective, one explaining theory being this or that. Do not necessarily mimic this thing but you might be successful (and quicker) by applying this theory in your design.".


I've seen (and done) tests that show the position of a padlock icon on a payment page have a significant affect on the money the money that a company makes. That's why UX designers get paid - because the choices we make make our clients money.

Imagine saying to a client: "We could make you more money, but lets not worry about it - I'm sure your customers will be able to adapt."


making these decisions would depend on your target audience. i suggest you take a data driven approach and identify who is visiting your website/app. if your analytics are telling that majority of your users are windows users go for the right side.


Would you expect your users to adapt to your slow loading website?

Of course not. As a dev, you would be busting you ass to make it as fast as possible. See Jeff Atwoods' Performance is a Feature for examples of how/why this is implemented on the Stack Exchange network. It also has info that answers your request for: 'experienced based answers that show minor changes could have big effect?'

[Google found that] the page with 10 results took 0.4 seconds to generate. The page with 30 results took 0.9 seconds. Half a second delay caused a 20% drop in traffic. Half a second delay killed user satisfaction.

In A/B tests, [Amazon] tried delaying the page in increments of 100 milliseconds and found that even very small delays would result in substantial and costly drops in revenue.

Performance is an example of how making small changes improves the user experience (or 'competitive advantage' as Jeff puts it) of your website. The exact same theory applies to visual and psychological nuances like label alignment, wording, colour choice, etc.
There are plenty of studys that have proven this.

If you need an gentle intro to web usability and user experience, check out the very well respected Don't Make Me Think by Steve Krug.

Nice of you to put it in bold, but just saying "because performance is important, UI is too" doesn't make sense if you don't back it up. –  Hugo Delsing Jul 21 '14 at 6:47
@hugo Fair call. Added link to article covering the results of a whole bunch of usability studies. –  David HAust Jul 21 '14 at 13:30
Thanks for the link, it was an interesting read. But it's also exactly what I'm questioning. Some researcher in a lab found that labels on the left are a little harder. So? They could still do the task. The goal was to find experiences where they changed something as little as alignment of labels and found that it actually matter in a real environment. –  Hugo Delsing Jul 21 '14 at 13:47

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .