Designer–engineer here. After 15 years building products [1] ,from
scrappy side projects to multibillion-dollar platforms, I’ve found the most productive way to think about design is as a growth accelerant. The OP reaches a similar insight but keeps it in the realm of product development. That framing is precisely where the friction between engineering and design tends to arise. Design isn’t about making the code look nicer... it’s the bridge between engineering execution and business growth.
Take Dropbox’s much-debated rebrand. Many on HN dismissed it as superficial. What they missed is that Dropbox’s growth had plateaued. The new visual language wasn’t meant to "improve the product" for existing power users. it was engineered to make the product feel approachable to an audience the company had never reached. It worked.
When designers focus on the measurable business impact of their work—and engineers stop treating design as a decorative afterthought—cross-functional frictions fade and growth compounds.
> When designers focus on the measurable business impact of their work—and engineers stop treating design as a decorative afterthought—cross-functional frictions fade and growth compounds.
Measurable business impact -> The dashboard becomes a battlefield and every team wants a modal for their feature release. Dark patterns come hand-in-hand.
I think the world would be better if individual designers focused less on business growth metrics and more on holistic User Experience.
I trust software more if it's ugly. Desktop software interfaces were pretty much solved by the late 90s. The majority of changes since then have been aimed at either dumbing down the software to better appeal to people who are easily manipulated and exploited, or just change for the sake of change so designers can justify their jobs. And I choose to avoid mobile devices because every one of them is slow and frustrating compared to using a keyboard and mouse.
I believe modern UI designers provide me negative value on average. Ugly software is a good sign because tells me no designer was there to ruin it.
"The growth marketing team has identified an untapped demographic".
> change for the sake of change so designers can justify their jobs.
"We've decided to rebrand to focus on XYZ market."
> mobile devices because every one of them is slow
"If we can run the website on the user's device we'll save on developer time."
> Ugly software is a good sign because tells me no designer was there to ruin it.
It just means they don't have the money to hire a good designer. Meaning no marketing department.
For whatever reason designers don't do open source. Looking at you GIMP. Blender finally bit the bullet.
"The growth marketing team has identified an untapped demographic".
Non-experts is always a larger group than experts.
A design prioritizing non-experts does not benefit experts.
Moreover, experts know they are trying to solve hard problems and benefit from tools that recognize the nature of solving hard problems entails doing hard things.
Or to put it another way, though there is a large market of people who don't play the violin, a ukulele is not a violin.
Me too, but there is a hint of survivorship bias here. I trust most of the ugly software I am still using because it was designed well enough in the ugly era to survive til today.
I have largelyforgotten all the bad ugly software.
You seem to be using the most negative expressions possible. Maybe I'm also less upset because most of the software that I use (Linux, FOSS, KDE) is less affected. I'd call it making the software look less intimidating / more approachable and more fashionable. They can both make some sense, though in many cases I also don't like it. Some fashions are objectively worse than others, like hiding scrollbars in non-touch interfaces and making interactive elements look like regular text.
IME, the ugliest software has not received much UX nor design work, and so the UX often sucks, too. Gitk comes to mind, it's very ugly and the weird diff scrolling behavior regularly gets me to where I don't want to go.
Interesting point and I was thinking if I believe the same, but I can't completely agree. Reason is there's a decent chunk of software outside of the "mainstream" app space where the UI is ugly and also broken, likely because the author isn't a great developer. I'm thinking engineering and research type apps. They have the kind of bugs where clicks do nothing, user feedback is faulty, wrong sequence of clicks causes a crash, and it makes me instantly distrust the entire product.
Following 90s design is how you get the atrocious GIMP 2 floating window mess. Furthermore, the standards in the late 90s and early 00s were so low because of the lack of available software packages that programs crashing your kernel was just part of the normal work week. Memory corruption wasn't treated as a problem unless it corrupted data on disk, because it was considered normal that a program couldn't run more than a few hours before committing memory corruption suicide.
UI controls were standardized and programs stuck to them because they didn't have the CPU cycles to waste on styles, but that didn't prevent programs from using text areas as listboxes, abusing picture elements to stylize checkboxes, and picking whatever cursor the dev decided fit most (despite its name and intended use) when hovering over controls. The control schemes for anything 2D or 3D graphics were also designed by dice roll. Programs also regularly didn't fit on your screen unless you adjusted your resolution, because people often ran on lower resolutions to be able to use more colours. "This website works best in 1024x768" didn't just apply to websites.
The 90s software that stuck around until today is good because it survived decades of competition and reinvention or because people who don't want to learn new software just refused to try something else. The thousands upon thousands of expensive software packages that looked just like it found their deserved deaths years ago.
This is the type of twaddle that I have come to expect from “designer” types. Constant boosterism and exhortations to respect their craft. And when left to practice said craft, the majority of the actual work product is visionary PPTs for execs and a pile of figma screens that barely cover the happiest of happy cases.
Sad thing is that I very much agree with the importance of design. But practitioners seem often insecure, inward looking and unconcerned about whether their dreams are actually workable. But they will talk your ear off about the “texture” of a font.
My pet peeve with the Figma drivers is that they tend to think of design as a 'phase' that happens before engineering. So as an engineer you are asked to implement a bunch of painstakingly laid out user interfaces that have never actually been used to do any work. IMO it's essentially impossible to design a good interface without feedback obtained from using it. Designers would be way more useful if they'd partner with engineers throughout the development process. But there seems to be huge cultural resistance to this mode of working. In fairness, a lot of business processes make more sense if you assume that having to talk to an engineer is one of the worst possible outcomes.
Nice, sequential flow charts are easy to understand and pitch and they also make it nice for the folks at the front to say they did their part and any issues are now someone else's fault.
When people try to diagram the actual process (closer to what you're talking about) it's just a giant circle (or "double diamond" for folks charging money for powerpoints) which isn't a tool that makes management feel easier or "on track".
A problem I've seen in programming too: understanding and articulating the value of quality is necessary but nowhere near sufficient for actually doing good work! So you end up with folks who've gotten past the initial hurdle—developing taste and confidence—but, either individually or in an organizational context, are only able to do okay quality work. Not awful but not especially elegant or insightful either.
And so we get people making totally reasonable arguments about the value of design and then producing polished but tasteless and shallow corporate products. Or we get people who start understanding the value of maintainable, understandable code, but get stuck on design patterns, "clean code" rules and "best practices" rather than conceptually clear, coherent and effective code design.
I have some sympathy. Getting past that initial hurdle rests on skills and tacit knowledge that take time, practice and mentorship to build after you've developed an appreciation for good taste. And, especially in a corporate setting, you have to get most of that practice and mentorship in public.
Oh, inevitably for a blog, this is about web site design.
Confidence in manufactured physical objects is more interesting. Discuss.
Raymond Loewy on a good day - The Honeywell Round.[1] The standard little round thermostat. It's still manufactured.[1]
Raymond Loewy on a bad day - the first attempt to make a steam locomotive look streamlined by adding a sheet metal body.[3] This bad idea caught on in the UK, for some reason, resulting in a whole series of difficult to maintain locomotives.
Eventually he designed the look of an electric locomotive, the GG-1, which was very successful, looked very good, and had good access to the important working parts.
For a good sense of how to think about industrial design, here's Lowey's autobiography, "Never Leave Well Enough Alone".[1] Look for the before and after pictures of common objects he redesigned. In most cases, sales went up.
As a bad example, see "Design for Dreaming", a 1956 General Motors promotion around the time the auto industry was reaching peak chrome and peak tailfin. This also introduced the first self-driving car, the Firebird 2.[2] (The self-driving didn't actually work in the Firebird 2. It did work in the Firebird 3, on a test track equipped with a guide wire.)
As long as it performs the core function people keep trusting it. That's why people bother to deal with all the BS about passwords, captchas, user agreements, crashes, updates, reboots that systems designed by the "best of the best" put people through. Even Apple products no longer "just work".
Don't for get that everything out there is a slop in some way. Apple's core promise since ever was to make computers easy to use and approachable and they improved a lot, enough to charge significant premium over the alternatives.
Computers, software in particular is very low quality across the board. Not just in UX but overall technical implementation is also comically low quality, to the point that contains huge security issues that wouldn't ever pass as acceptable in any pre-computer utilities like microwave ovens or blenders.
Software products most of the time offer no guarantees and it gets very expensive when quality and guarantees are involved through SLA.
> “If they couldn’t be bothered to align the text, what else did they overlook?” But a clean, coherent layout creates harmony. It shows someone cared. And that care is contagious. It makes the user care too.
It may only show that they cared enough to fool you into thinking the product is good. Good design can be hijacked to be a fake proxy for a good product. It's basically fruits and vegetables that look good to sell and taste bland.
That's a similar thing to clothing though. You may argue how it can be manipulated all you want, but people do think in this way and it's a socially agreed upon contract to show you care enough about the occasion and have enough respect for the other people present. No head of state shows up to a meeting in shorts and slippers for a reason.
It's mindblowing how the majority of the comments here seem to be rather cynical and distrustful. I'd be really discouraged if I were the author. I (as an engineer) found the article rather well-written and convincing! Well, at least it did get 100+ upvotes.
If you can ignore the polarized politics and eye the Cybertruck objectively: sales are awful due to it being ugly. Everything about it is solid in terms of design, yet it looks awful and is destined to be a partial failure for Tesla.
Readit News
Take Dropbox’s much-debated rebrand. Many on HN dismissed it as superficial. What they missed is that Dropbox’s growth had plateaued. The new visual language wasn’t meant to "improve the product" for existing power users. it was engineered to make the product feel approachable to an audience the company had never reached. It worked.
When designers focus on the measurable business impact of their work—and engineers stop treating design as a decorative afterthought—cross-functional frictions fade and growth compounds.
[1] https://fairpixels.pro
Measurable business impact -> The dashboard becomes a battlefield and every team wants a modal for their feature release. Dark patterns come hand-in-hand.
I think the world would be better if individual designers focused less on business growth metrics and more on holistic User Experience.
To me, this is exactly what the OP is going on against.
If it's not cohesive, it's bad design.
I believe modern UI designers provide me negative value on average. Ugly software is a good sign because tells me no designer was there to ruin it.
> dumbing down the software
"The growth marketing team has identified an untapped demographic".
> change for the sake of change so designers can justify their jobs.
"We've decided to rebrand to focus on XYZ market."
> mobile devices because every one of them is slow
"If we can run the website on the user's device we'll save on developer time."
> Ugly software is a good sign because tells me no designer was there to ruin it.
It just means they don't have the money to hire a good designer. Meaning no marketing department. For whatever reason designers don't do open source. Looking at you GIMP. Blender finally bit the bullet.
Non-experts is always a larger group than experts.
A design prioritizing non-experts does not benefit experts.
Moreover, experts know they are trying to solve hard problems and benefit from tools that recognize the nature of solving hard problems entails doing hard things.
Or to put it another way, though there is a large market of people who don't play the violin, a ukulele is not a violin.
I have largelyforgotten all the bad ugly software.
IME, the ugliest software has not received much UX nor design work, and so the UX often sucks, too. Gitk comes to mind, it's very ugly and the weird diff scrolling behavior regularly gets me to where I don't want to go.
Interesting point and I was thinking if I believe the same, but I can't completely agree. Reason is there's a decent chunk of software outside of the "mainstream" app space where the UI is ugly and also broken, likely because the author isn't a great developer. I'm thinking engineering and research type apps. They have the kind of bugs where clicks do nothing, user feedback is faulty, wrong sequence of clicks causes a crash, and it makes me instantly distrust the entire product.
UI controls were standardized and programs stuck to them because they didn't have the CPU cycles to waste on styles, but that didn't prevent programs from using text areas as listboxes, abusing picture elements to stylize checkboxes, and picking whatever cursor the dev decided fit most (despite its name and intended use) when hovering over controls. The control schemes for anything 2D or 3D graphics were also designed by dice roll. Programs also regularly didn't fit on your screen unless you adjusted your resolution, because people often ran on lower resolutions to be able to use more colours. "This website works best in 1024x768" didn't just apply to websites.
The 90s software that stuck around until today is good because it survived decades of competition and reinvention or because people who don't want to learn new software just refused to try something else. The thousands upon thousands of expensive software packages that looked just like it found their deserved deaths years ago.
Sad thing is that I very much agree with the importance of design. But practitioners seem often insecure, inward looking and unconcerned about whether their dreams are actually workable. But they will talk your ear off about the “texture” of a font.
When people try to diagram the actual process (closer to what you're talking about) it's just a giant circle (or "double diamond" for folks charging money for powerpoints) which isn't a tool that makes management feel easier or "on track".
And so we get people making totally reasonable arguments about the value of design and then producing polished but tasteless and shallow corporate products. Or we get people who start understanding the value of maintainable, understandable code, but get stuck on design patterns, "clean code" rules and "best practices" rather than conceptually clear, coherent and effective code design.
I have some sympathy. Getting past that initial hurdle rests on skills and tacit knowledge that take time, practice and mentorship to build after you've developed an appreciation for good taste. And, especially in a corporate setting, you have to get most of that practice and mentorship in public.
Deleted Comment
Confidence in manufactured physical objects is more interesting. Discuss.
Raymond Loewy on a good day - The Honeywell Round.[1] The standard little round thermostat. It's still manufactured.[1]
Raymond Loewy on a bad day - the first attempt to make a steam locomotive look streamlined by adding a sheet metal body.[3] This bad idea caught on in the UK, for some reason, resulting in a whole series of difficult to maintain locomotives. Eventually he designed the look of an electric locomotive, the GG-1, which was very successful, looked very good, and had good access to the important working parts.
[1] https://www.smithsonianmag.com/innovation/meet-product-desig...
[2] https://www.honeywellstore.com/store/products/the-round-non-...
[3] https://allthatsinteresting.com/wordpress/wp-content/uploads...
As a bad example, see "Design for Dreaming", a 1956 General Motors promotion around the time the auto industry was reaching peak chrome and peak tailfin. This also introduced the first self-driving car, the Firebird 2.[2] (The self-driving didn't actually work in the Firebird 2. It did work in the Firebird 3, on a test track equipped with a guide wire.)
[1] https://archive.org/details/neverleavewellen0000raym/page/38...
[2] https://www.youtube.com/watch?v=tG_KxSM6bTI
[3] https://youtu.be/tG_KxSM6bTI?t=427
A small number of designers who created the iconic objects of midcentury modern. Who do we have today? Johnathan Ive?
There’s a lot of nice looking crap out there.
Don't for get that everything out there is a slop in some way. Apple's core promise since ever was to make computers easy to use and approachable and they improved a lot, enough to charge significant premium over the alternatives.
Computers, software in particular is very low quality across the board. Not just in UX but overall technical implementation is also comically low quality, to the point that contains huge security issues that wouldn't ever pass as acceptable in any pre-computer utilities like microwave ovens or blenders.
Software products most of the time offer no guarantees and it gets very expensive when quality and guarantees are involved through SLA.
It may only show that they cared enough to fool you into thinking the product is good. Good design can be hijacked to be a fake proxy for a good product. It's basically fruits and vegetables that look good to sell and taste bland.