Ask HN: Why is tech UX so bad in general?

36 points by friendly_chap ↗ HN
I have just tried to get my access back to my bank account.

Since I switch numbers frequently due to moving countries, 2FA is hell for me. To get my account back I need a selfie with both my face and ID visible. Very hard. Guess it's due to security.

After some hours wasted I got it done. Time to transfer the file. Of course gmail does not send my email. Fear not, I am smarter than that. Upload to Drive. Nope, does not work either. Weird.

Upload with facebook messenger maybe? Hmm they compress images like hell. The other day fiverr was not loading orders for like a day. Wise.com turned chinese for no reason. Google cloud resets my functions to private randomly.

Am I losing my mind? Getting old? Too busy? Frustrated? Am I just a shitty developer? How do non techies cope?

39 comments

[ 0.78 ms ] story [ 98.9 ms ] thread
The H-1B program allows companies and other employers in the United States to temporarily employ cheap foreign workers in occupations that require highly specialized knowledge. Many businesses outsource work to other countries for additional cost savings. Companies rely heavily on automated testing to mitigate the downside of hiring less skilled developers. You have worse UX but they spend less money. What can you do.
Are you implying that immigrants or devs working remote from other countries are less competent? If so, consider how much additional hurdles those guys had to overcome in order to find a job in the US.
What about being from a country that's not the United States makes a developer less skilled?
It's not that they arent skilled, they just don't have the incentive to demonstrate them.
Nothing.

From browsing software engineering literature from the early 2000s, the common issue seems to be communication or processes, which is the fault of the hiring company, not that of the developers/consultancy themselves.

Top of my head most of it is compliance to varying laws in different jurisdictions, and of course optimization for growth and profits.

It's hella surprising how much users will keep using a product with shitty UX as long as it still solves a problem for them.

The tech world functions by selling products that are unfinished, and promising to keep working on them after you buy them. These days, almost everything is de facto early access.
”finished” is just a label, all software can be worked on infinitely, at some point production usage HAS to start
>I have just tried to get my access back to my bank account.

This is related to tons of regulation regarding KYC. Nothing to do with "tech".

UX in tech is bad for a great number of reasons.

1. Good UX is hard and even harder to get right from the start. When teams* are faced with tight deadlines to deliver functionality, they invest too little time in figuring out how to best solve a problem. Even less time and effort is expended to follow up on something that was delivered in order to see if it works (it doesn’t) and how it can be improved.

2. Developers* have their mind set on optimization. This applies to ticking off requirements on a backlog item (I refrain from using the term "User Story" on purpose; the context of features isn’t properly communicated either). With UI development, there is a myriad of details to cover that require attention from everyone and developers* won’t complain if there is no detailed spec for input constraints, form validation, error handling, responsive behavior, keyboard interaction, a11y. Works as designed, why bother.

3. UI Development (for tech, the interface to the user makes/breaks the UX) is seen as "lesser" or "a complicated mess to avoid" by senior developers*. My subjective feeling is, that the seniority in UI dev is/stays behind the mean. There are more juniors around.

4. To a degree, the product quality in tech isn't great, it's not only bad UX. Bad UX is merely more visible, despite it not making the news as much as the backend engineering blunders of not sanitizing input or exposing databases to the interwebs.

5. https://userinyerface.com

* Not all of them. There are exceptions - appreciate them and treat them nicely, they are hard to replace :)*

“ My subjective feeling is, that the seniority in UI dev is/stays behind the mean”

With regard to quality/competency or organizational influence?

Good question.

Long answer:

What I meant to say is that there seems to be more of a tendency for full stack or front end engineers to bleed into backend / system than the other way around. If we were to assume that seniority (here, strictly referring to days spent with bum on chair) is correlated with organizational influence, it would follow that there are less devs with focus on UI who end up in leadership roles. But as I said, I don’t claim to have an objective truth here.

"UI Development" is not gate keeper of great UX. If you want great UX invest in professional UX designers.

If you want great house to live in, hire an architect, not civil engineers.

[Edit]

After hiring UX designers, listen to them, fund them (enough dev bandwidth to realize the design proposals), do not triage away "nice to fix" UI bugs.

That is, change the engg' mindset.

If I had to choose between a great architect and a mediocre builder or a master builder and a mediocre architect, I would choose the former to build my house. I say that as an (UX) architect :) Plans are worth little if you don't have the right people to translate them into reality with the right quality - and in my experience that involves more than painting by numbers. Imo great UI developers are essential for building great UI experiences.
Yes. But they are already there and established, so why then this 'Ask HN'?
What are the financial incentives to produce good UX? In your case, it’s regulatory compliance (KYC), which is the lowliest incentive of all. Also, customer losing access to an account happens infrequently and is not a large source of loss for your bank, I would guess. And you are strongly incentivized to plow through whatever UX they throw your way, because you literally need to do it to get your money. The incentives are aligned entirely against you in this case.

Applying for a loan, though, well I’ll bet you’d breeze through that interface on any device.

UX follows short-term profit incentives.

When UX is your competitive advantage, and especially when switching costs to another product are low, that’s where you’ll see the best UX. Google Maps and Google Search come to mind.

> when switching costs to another product are low, that’s where you’ll see the best UX

In my career as a designer I've found that good design (functional, fast, works predictably) is primarily limited by business model incentives.

You hear the most about the latest trends and frameworks in situations where the best anyone can do is putting lipstick on a pig.

The dirty secret is that no self-serve tech UX is good. For valuable clients you hide the pain behind an account manager.

As an aside, get Authy if you can and replace the SMS 2FA. As someone who worked in account security, changing phone numbers is account compromise like 99% of the time.

Generally, all examples you mentioned don’t have competition or compete in different things than UX, so there you go.

One example, banks, especially old big banks. They compete first and foremost on public perception of them being secure. As we know, security and usability are almost always diametrically opposed, so it’s a concept of a bank that makes banking hard to use. Some companies are trying to change it, but it’s hard.

Banks are a something else entirely when it comes to art of horrible customer service. The other day, I noticed a "security alert" when logging into my bank account. "Someone logged into this account from an unknown device. If it wasn't you, panic" basically. I wanted to write a suggestion that they add an IP address and timestamp to make that message useful and actionable, but they don't even have a contact email address, just phone numbers (even for the IT department!). The "chatbot" doesn't connect to actual humans.
I'm guessing it's because focus has shifted from the "basics" of being able to do the actual thing to something else.

That something else is probably "having it JustWork(tm) in 98% of the cases" is better economy than "having it work in 100% of the cases". Your product will win if it works and "feels nice any easy" for most people, compared to a product that is flexible, and works and utilitarian, but maybe "looks worse" because it actually was made to work, also when things don't go smooth.

It's definitely a trend, whitespace everywhere, UI's void of buttons or obvious hints of what to do or what can be done, a sacrifice to the gods of UX.

I agree with the 2 paragraph of this comment. In the case of a frequent traveler there is got to be a bank which cares about such case. With some time some banks will be better at this when more people will travel which is unlikely after COVID.

On a tech side: it’s very hard to create, support and sustain a super flexible system which wouldn’t be perfect for everybody.

People don’t spend their time configuring 100 settings in their Zoom profile and struggle finding invite link.

And at the same time in some place of this world a person lost a bank card and smartphone being in a foreign country… rare case but turns life into hell quickly.

While I do empathize with bad UX this story does feel like a bit of an unnecessary dive across services without trying to fix the root issues.

If your contact info is not stable then you are going to have a hell of a time with 2FA through your phone. There are alternatives, both hardware and software that will fix this problem for you. Have you looked at what your bank supports?

If the answer is "yes and they only use SMS based 2FA" then fine - the next issue is figuring out why your email provider isn't sending an image. Is it too large? Gmail supports multiple quality settings and has a will tell you the attachment is too large to send - for a single image you should be able to select a lower quality to send it.

This dive into other services sounds like a waste of time. "Good developers" would at least try and understand what is broken before throwing random OTS storage services at it.

You asked "in general."

It's because "UX" is presently way too full of itself; it passes off a lot of mere fashion and preference as "objectivity" or "science" and then other UX people get suckered into believing it, not just continuing arbitrary trends, but also reducing usability and optionality along the way.

Poofy vs. flat buttons is just a preference, it's not science.

So fashion is focused on instead of real usability. Fixing this will take time, and probably regulation type measures. The analogue here is architecture. It's great to make something beautiful, but if you don't have wheelchair ramps or the bars on the railing are wide enough for a kid to get a head stuck in you don't get to build the building.

UX needs more of this.

The high point of UX design was Windows 95. It was consistent. Windows had menus in the same place, and everyone knew what a menu was. Buttons looked like buttons. Apps generally used the same Win32 API and had the same appearance. Key bindings were underlined. You could teach it to someone and they could learn it. And once they knew it they could be fast.

Now I have to explain hamburgers and ellipses to my 80-year-old relatives. They have no idea what's going on. It's bewildering for them. It makes me realize how large a part of my brain is forced to run a simulation of the UX designer in my head, to play a game of long-distance charades with them.

Google is to blame for a lot of this. They kicked off a minimalist trend with Chrome, which everybody aped. Now everything is a nearly-featureless shape with a few ambiguous hieroglyphs.

Internationalization and cost-saving may also be part of it. The old Windows was very verbal: It used words, English words. But it's expensive to translate those to various languages. And expensive to make sure it still works in Arabic and Chinese. So instead, Tech companies have ended up with an iconographic pidgin that they don't have to translate. The burden is pushed to users.

It also has to do with the move to WebApps and the absence of standardized GUI APIs. There's no "Win32" for WebApps. So people invent their own.

(Actually, I take it back; there is a standard: It's the default form elements, that you had at least as far back as HTML4. When was the last time you saw a native checkbox or radio button?)

But I don't think that's the main thing. I think it's mostly fashion. People would be faintly embarrassed to release a program with Win32 clicky buttons -- even though other people would find it easy to use.

Sure, it's fine.

But I think there's a bigger point here.

I could change Windows 95. Dark theme, light theme, lots more. See also Linux for even more possibilities.

That's the great big ball of stupid that all of UX presently misses, and fundamentally why I don't much respect it. ALWAYS at least allow the damn user to decide if they want to change something.

To your point, I'll add that semantic HTML with CSS was like this. It was maybe a little technical, but a motivated person could fairly easily override their browser's stylesheet. And I imagine that this was not just some esoterica to, say, people with limited vision. I don't think you can do this anymore, practically?
Yes yes yes. Exactly. For all of its slowness and bloatedness, this is one reason why I'm slowly but strongly coming around to GUI apps made with html/css, like Electron et al.

Obsidian strikes me as a great example of this; it's pretty great UI out of the box, and also you can edit it to your heart's content.

As much as I like to point out the emptiness of such trends, it would be unfair to place it fully on the UX folk. It comes down to incentives and economics. We the users won't look at a piece of useful software if it doesn't 'look' like the kind of thing that is up-to-date and modern in its behaviors. Same goes for everyone else in the software production industry, except for the core function developers who complain, myself included.

In short, fashion exists because it sells. It's also good to know who you're selling to, sometimes it's not the person using your product, e.g. Apple is selling to those who don't already have their latest product, so it's important to round-off (or square) those edges, etc.

I'm not putting it on the people, but I am putting it on the "field" or the "discipline."

Perhaps in the same vein as "Economics," which purports to be a science but in practice is something else. UX may even be worse, because at least Econ has had the time and space to be established and is at least arguably a "discipline."

I'm quite literally unsure as to whether "UX" should even exist as its own thing at all. I'm not sure you can mix "engineering" and "advertising" and "fashion" and "psychology" to this extent and have something coherent at the end.

In an older age, UX would have been industrial design which I can certainly appreciate. Historically there's more of a separation of industrial design and how a thing operates, the variation in the exterior of cars doesn't dictate the drivetrain or controls. With software there tends to be more overlap. How things are grouped, displayed/hidden is explicitly affecting the displays and controls.

We do need good UX, usability, ergonomics, etc. It can't be left to engineers who will merely attach a display/control to every element of the data model with 1:1 mappings. I agree that UX doesn't seem to be its own thing but it's hard to separate, with Conway's law and all.

The best we can do is be frank with who it serves and/or why. I don't think people really want to know this level of truth on a daily basis though.

"I don't think people really want to know.."

But why not go ahead and actually find out? That's my point, and that's what makes this field potentially SO MUCH DIFFERENT. It's true, in the past you couldn't do this for toasters. You had to do a ton of homework and pick a design and hope people liked it.

But this is software! Just bake in a little "hey, you might break this but here's how to fiddle around with it" and let people work-slash-play.

This sounds good in theory, but then we could end up with something like the various Linux desktops. It's hard to find a complete and coherent one that doesn't also have annoying bits too. Usually if such a thing exists it's from implementation of a single vision rather than a collection of hobby experiments.
One, there's no good reason that you can't have both. None at all. Lots of times in Steve Job's history where he arbitrarily (or at least, not for reasons of his expertise in UI) shut down greater user configurability -- not for the sake of design, but for the sake of market control.

But also, two, I think "various Linux desktops" would be far preferable to what we have now. Again, I appreciate Steve Jobs' accelerating things in many ways, but also, he made interfaces perhaps too simple, to the point of frequently being just dumb.

First anecdote that comes to mind is my local pizza place’s buggy mobile web UI, with generic HTML inputs for all their fields that make typing in my phone and email a pain due to autocorrect kicking in. They also shove their app down your throat on the web ui, and the app somehow manages to be even more unusable!
UX is generally OK but you unfortunately are dealing with a corner case.

So you’re experiencing scenarios which have not been taken in consideration.

99% of Programmers are not product designers. The more you look into it the more you understand why Steve Jobs is so special.