I can’t say I’m surprised by this. The last real large scale UI paradigm shift was the iPhone’s multitouch display. UX is going to follow UI capabilities and form factors.
In the automotive space I’d argue there has been a UX regression as they went pure touch screen for some models.
That trend is starting to change. Mazda and Honda are starting to phase out touchscreens in some cases[0]. Tangentially related, the US Navy is also phasing out touchscreen controls after they were determined to be a factor in the USS McCain collision in 2017[1].
It depends on the manufacturer. Both VW and Mercedes are doubling down on their touch-screen only interfaces. The new Golf, which is (historically) one of the top 10 best selling cars in the world, is entirely touchscreen for everything (and the early reviews are not positive about it). The Mercedes infotainment system which is also entirely touch-based is largely praised, although it is much more expensive and well-done.
It's very easy to do shitty touchscreen UI. Even more, it's possible to make it unusable. For example, mcdonalds screens. They use very bad, slow hardware - is it even capacitive? Paired with slow, unnecessary animations and having to go through 10s of screens to order something, using it is a real pain.
It's also possible to do great touchscreen UIs. I'd hate if my phone had t9 keyboard.
If the safety and consumer satisfaction data backs up the companies that have decided to ditch the screens, I expect to see others follow suit. As others in this thread have pointed out, few automakers have really seemed up to the challenge of developing a touchscreen experience that is truly seamless.
Anecdotally, I bought a used 2019 Corolla recently and the "maintenance needed" alert that beeped and flashed while I was driving down the highway scared the crap out of me. It turned out that it was merely reminding me to get my oil changed, I couldn't believe it. It was so much less polished of an experience than I was used to with the 'Yota.
I was driving a Ford Fiesta for a while and it would hijack the stereo, mute the music and say something like "Please connect your phone to run a vehicle report" while driving once a week or so. The only way to get the radio back was dismiss the modal it popped up or presumably install the app and let it send data home.
I've been thinking for years that automobile manufacturers entered an industry (software) that they had little knowledge of, and it showed, with janky, slow interfaces and last-generation mapping software.
However, they've invested in this in the past decade, and if there's going to be a UX resurgence, it's going to be in the automobile sector, as their teams are maturing more and more.
I've used a fair amount of different cars as I'm part of a city car club.
I can't say I see this at all. Even their tactile controls are generally a mess.
Controls as basic as window wipers or headlights seem to be different for ever car for no discernible advantage.
A different 'start' mechanism for every car is actually quite a nightmare. I once spent 20 minutes in a car park trying to figure out how to get the damn thing to start, even their own instructions didn't say how!
> A different 'start' mechanism for every car is actually quite a nightmare.
I've driven a bunch of cars, over the years. I haven't seen too many different mechanisms.
What have you seen on modern cars that's different than make sure it's in park or neutral (or clutch pedal depressed) and turn the key, or for keyless make sure it's in park or neutral (or clutch in), press the brake and push the button? And if the key battery is low, press the button with the key.
I've driven some older vehicles where you turned the key to enable ignition, but had a separate starter button, and some vehicles with ignition locks that didn't work, so you could turn the ignition without a key. And some cars with keys go off -> accessories -> ignition -> starter vs accessories <- off -> ignition -> starter. If the key retention in the switch/lock is broken, sometimes it's tricky to make sure you pull the key out in off instead of accessories.
You’ve missed the variety of handbrake mechanisms. For example, I was in a Mercedes the other day:
1. It had a fob that worked as a key, but it didn’t look like a key and the key slot didn’t look like where the fob should go.
2. You set the handbrake using the left foot, but release it with a pull handle with your right hand!
Both were unintuitive, and would catch out some people (I am generally well above average in working out how to use execrable user interfaces).
The worst UI is often to release the door to fill fuel. More than once I’ve needed to ask for help on unfamiliar cars, and multiple people also not knowing how.
I had a negative gut reaction. It was mostly autonomous, if there was a problem or high stress situation, this quote is concerning:
Hurley and Behnken worked with SpaceX to "refine the way that you interface with the touchscreen and the way your touch is actually registered on the display in order to be able to fly it cleanly and not make mistakes touching it," Hurley said.
Working in the UX space and seeing all of this across a wide range of projects for over a decade now, there are a few flaws in this observation.
For many projects we've been involved in with our UX/UI Design service for SaaS companies, founders worked with us before releasing their product. So before the product was released we could eliminate common pitfalls based on our previous experiences. When a product is released you can then use data to optimise things.
An other issue I have with the article is that in some cases I've seen upclose, a tiny tiny percentage improvement in UX equaled an extra 6-figure increase in revenue. When a product is in a stage of optimising UX for more revenue, it's in the company's best interest to focus on the improvements that have direct impact on company revenue. So improving the UX on pricing pages, cancellation user flows etc.
This means a smaller UX improvement, sometimes really really small, but a huge upside for the business. So smaller UX improvements does not mean it has less effect on businesses. If anything, it's more likely the impact is growing because UX optimisations don't happen in random places anymore.
I don't see any flaws that you mention when I read the article.
"This doesn’t mean that we can expect 50% of all design projects will have an improvement score within that range. It’s possible your own impact could be a 500% or 5,000% improvement, but our data suggests that an improvement score that high is unusual."
The UX improvement I most want to see in modern software is to optimize for reducing the amount of time I spend in your app.
There are business models where companies are optimizing for the opposite ("engagement"), and that's unfortunate. But there are also apps and sites that have no need for that and are just cargo culting the same UX features, and that's avoidable.
On iOS when I was turning off notifications from this one app, the system asked if I only wanted to turn off marketing notifications, and I did. That was a nice middle ground.
Ugg. The App Store Guidelines disallowed using notifications for ads...even though Apple itself violated this. Apparently, they changed this policy last March.
I'm convinced that Windows developers have perverse incentives tied to optimizing engagement or telemetry metrics that happen to be highly correlated with irritating power-users.
I doubt that's because Windows wants you to use it more. It's the freaking OS! I think it has more to do with UWP and the Settings app itself. On the (slight) plus side, they have improved their search so much over Windows 7.
I'm not sure if it's intentional or incompetence but here's an example from Firefox: Count the number of actions it takes to delete specific cookies (say all matching "google", which I do every time I log into YouTube so it won't log me into Search etc.)
Firefox's UI is just appalling in a number of ways.
Ever since the new preferences dialog was introduced in Firefox, the UX has been quite slow. What was previously achievable through a few keystrokes now requires several.
For your use case, you could use the Cookie Auto Delete extension and configure it appropriately. Combined with Multi Account Containers and Temporary Containers, your life will be a lot easier.
I wonder how much of this is because of establishment of more well-known best practices and diminishing returns for incremental improvements.
I'll use PC gaming as an example: go back a decade of more, and it wasn't uncommon to need to use a significant portion of the keyboard (or on a console, more awkward controls). These days things are much better in a lot of games. The use of contextual controls allows the use of the same button for multiple tasks. Tabbed menus mean you don't have to have a separate key for each thing, even if many games still provide that option.
That's not to say all applications have better UX than in the past. The ERP system where I work is truly cumbersome. But there's still a certain commonality of UX "language" in lots of newer projects that mean software engineers aren't having to figure this stuff out from first principles.
Interesting examples, as I see many of those as regressions in UI. The use of contextual controls means that you're never quite sure what action is going to result. For example, Mass Effect 2 had quite a lot bound to 'E' as "interact", which could be anything from calling an elevator, talking to someone, crouching behind a wall, or picking up an item. Try to talk to someone near a wall, and you might hide behind it instead. Similarly for the tabbed menus, they hide information several clicks deep, rather than showing you what you want right away.
I'm not sure if that's a good explanation. It's becoming a standard practice for many organizations to involve UX designers and researchers from the very initial phase of designing a product or feature. They use existing data, user-research, user-testing, and popular best practices before pushing changes to production.
So more UX focus is happening before release phases instead of after release, which means there will be much less opportunities for UX improvements after the release of each feature.
Also analyzing and tracking is a default everywhere now. So many feature requests are already backed with data related to user-behavior. Which means many of them kind of include UX improvements.
I think UX impact is bigger than before. But, it's becoming part of the process and it's more mature now.
It's because UX is increasingly becoming standardized, starting to resemble a "solved problem" in some parts.
It's like website design. Far fewer jobs available in that field because so many now just use lightly configured, pre-built themes. Website designs, in other words, are standardized far more than ever.
Yes, there are exceptions, as others here describe, but they don't negate a clear trend.
In consumer software? No. UX has regressed and most software is now written by employees who don't have much of a say or have given up.
remember that function is only important in closed software insofar as it keeps people from moving to something else. If the cost of the move is higher than the cost of bad UX there's no buisness incentive to fix it.
Not at all, the field is exploding, and people's expectations are constantly evolving. Because UX is fundamentally about people, there will never be a final 'end state'.
It's a bit like saying software is a "solved problem" - in reality, the needs of businesses and people will always change.
Here’s a (perhaps weak) alternative explanation for the headline and chart of the graph: suppose there is some UX quality number representing how good an average app will be if it is updated. Now suppose this number increases at 10% a year. If you update your ux every year, you might therefore observe a 10% improvement each time you update. If you update your app only every 5 years, you would instead see a 60% improvement.
So my question is: how does this study compensate for changes to the frequency of updates?
Yeah in the hypothetical world where things improve by 10% each year. But I don’t really see how it’s relevant. The way to compensate would be to say something like “average growth rate := growth ^ (1/#years since last update)”
Average UX improvements are negative and have been for quite a while. (There's enough noise that it's hard to say exactly how long a while, admittedly.)
Agreed. I suspect designers deliberately harm UX in an attempt to drive away unprofitable power users. E.g. Firefox recently made the about:config UX much worse.
Previous version only made you click past the warning once. New version makes you click every time, and there's a noticeable delay before the config settings appear.
Previous version had comfortable spacing, with everything clearly readable without excessive wasted space. New version adds an absurd amount of padding to waste your screen space.
Previous version let you sort by column, so you could easily see all the setting you'd changed. New version removes this feature.
about:config is mostly used by the kind of person who complains about user-hostile policies and declines to accept advertising and spying, so if you want to do these kind of things then making it worse can help drive those users away. Simply deleting it is too obvious, but sabotaging the UX has plausible deniability. If you dumb everything down enough you'll only have dumb users, and dumb users are more easily exploited. I can think of no other reason why somebody put serious time and effort into actively making it worse.
(The old UI is actually still available at chrome://global/content/config.xhtml, but now that it's no longer supported it's easier for them to remove it later.)
> I suspect designers deliberately harm UX in an attempt to drive away unprofitable power users.
You might be partly right. I mean, I've worked in UX for more than a decade and I've never met anyone who did this deliberately. But over-optimizing for the novice experience essentially produces the same result.
A decade ago it was trivially easy to find improvements that benefited all users, but these days platforms are a lot more mature and there are fewer universally beneficial usability improvements like that left. At this point making improvements for one persona often result in regressions for others. It comes down to a business decision of which users matter and how the changes impact the bottom line. And you sort of have to make a call because if your company decides not to optimize for a profitable set of users then your competitors will.
I've never worked on UX for a web browser, but my guess is that they're mostly optimizing for novice users and safety at the expense of the power user experience. I don't see that changing as long as their business model depends on appealing to the largest possible number of users. Of course I wasn't in the room when the team made that call and you may be right that they're engaging in some dark pattern shenanigans to actively sabotage your experience and make you leave but I seriously doubt it's that explicit -- no one's actively trying to drive you away, they just don't particularly care about you. You might be able to find a niche web browser just for expert users that meets your need, but I'm not not sure how they'd sustain the business.
Having said that, there are lots business models in which power users are profitable. I've worked on expert systems (military, finance, ETL, data science) that often have crazy high learning curves for novices because whenever there's a tradeoff they optimize for the efficiency and stability for expert users. Having said that, these systems can be expensive (for example ~24k/yr/seat) so the businesses can afford to focus on optimizing for a small number of expert users.
Yeah. There was a time that you could use chrome://global/content/config.xul to get to the old about:config, but they removed that too. Don't know why they did that, only people who wanted the old behaviour would use it. Being able to sort by column is really useful, that's what I miss.
Edit: Didn't see your last line about chrome://global/content/config.xhtml , now I can get the old behaviour back. Thank you, thank you, thank you.
> I suspect designers deliberately harm UX in an attempt to drive away unprofitable power users.
Considering how much things can be achieved via about:config registry in Firefox, then you're probably right. Not sure if entries can be still loaded with userChrome.css and userContent.css - back then I was using these to preconfigure browser for friends. These could be dropped into freshly created profile (via -P switch, without launching browser) and act like sorta "answer files".
1. Attempt to register
2. Find out I already have an account / I wait for a registration link that never comes
3. Find the “Forgot password” link, often buried.
Multiply this friction by 3X to 5x on mobile devices.
It’s a challenge with many issues, but the go-to design patterns for the most impactful interaction a user will have with a platform is totally broken.
Just this week my son had to drop a class and re-add it due to issues with his username. The issue? The one he had wasn’t recognized on login and when he went to create it he got an error message (unfriendly of course) that it already existed.
Agree with the premise, but there are definitely sink holes in today’s UX/UI.
That's positively the worst UX for logging in I've seen. Having to connect switch between apps and waiting for/fishing out the login email over 3-5 folders is so painful. And oh yeah - doesn't work in Incognito/non-default browser because the email app always opens the default browser in full cognito mode!
Intuitively, this makes sense that the lowest-hanging fruit from when UX was a younger discipline has been steadily eliminated from the market as design has become less of an afterthought and a certain baseline of best practices has taken root in the industry.
Thinking back to my own experience over a decade ago, the role of design in the development process has changed much for the better - it’s that much harder to make obvious UX mistakes now that design and development teams are used to working collaboratively rather than throwing things over the wall.
> "We weren’t surprised to see such a wide range because the outcome of any benchmarked design project depends heavily on the following factors:"
I would have expected to see "users are becoming more savvy and sophisticated about UI's", as well as something along the lines of "Many experiences and sub/micro experiences are becoming standardized. It's difficult to optimize that which is already perfected."
The conclusion (generally improved UX) sounds a bit dubious: bloat (including bloated graphics and splash screens, as mentioned there) seems to increase, not to decrease with time. Unfortunately the linked report doesn't explain how it is estimated (beyond hinting that ROI for UX work is taken into account), and the full report (which possibly explains it) is paywalled. It's an interesting topic, but the linked document is not very informative.
It looks like the article focuses on websites. But things can get very different if you compare desktop to mobile, web apps to "normal" websites, and web apps to native apps.
The web seems to get all the attention, and one can argue that the UX is improving, especially on mobile. However, everything that is not the web, like desktop apps, or worse, special purpose hardware, is being shoved into the web model and with it comes a significant degradation.
Before the iPhone, most companies didn’t bother to hire designers, or if they did, didn’t give them much power. After the iPhone, everything changed, and not just in the mobile space. The iPhone was such a vast, visceral improvement over what came before it, and the experience was so personal and eventually ubiquitous, that it shocked decision makers into action.
>> does this actually imply UX advancement any more than the possibility that people have become more familiar with/adept at using UIs since 2006?
Not really. The kinds of changes they're taking about get tested. A site will collect data, make changes and observe the impact on the data. A good way to go is present different users a randomly selected version of a site and track which version does better. With tests like that we can rule out people simply adapting over time.
> Not really. The kinds of changes they're taking about get tested. A site will collect data, make changes and observe the impact on the data. A good way to go is present different users a randomly selected version of a site and track which version does better. With tests like that we can rule out people simply adapting over time.
I've always wondered if the reverse study is done; keep the old versions of the UX around for months/years as an experiment to make sure the improvements are still there, or if there's a regression to the mean. This sounds much harder to do because most development is roll-forward with new features and such.
Also; practically how does attribution play into conversion measurements? In theory it should be easier now to tell whether it was landing-page+UX or view-through/click-through that made the bigger difference.
Context will be hard for AR. Fascinating and ‘the future’, for sure, - but my gut says it will end up being severely limited (in the way that there is only minimal personalization to the interactions with my Apple devices) or a total mess.
I actually think this too. Sadly I think that the real good stuff will only start happening in a good few years (decades maybe) when hardware catches up with our ideas.
I dont think that could apply to Windows at all. The same problem with Microsoft as they were in the 80s. That is 40 years now. They dont understand UX. They dont understand design.
I am rather sad only Apple is really doing much in this field. Especially when I continue to loss faith in Tim Cook for some strange reason.
In the automotive space I’d argue there has been a UX regression as they went pure touch screen for some models.