Settings are not a design failure

There are a lot of people who are willing to engage with your software and use their critical thinking skills in that process. I think we're losing something valuable by dumbing down all software and all interfaces to the lowest common denominator in the pursuit of "no friction" for users. Some friction is totally fine.

I wouldn’t say I like to muck with the settings, but I’d consider it a critical design flaw if my desk chair didn’t let me.

Same with my car’s seat, steering wheel, and so on.

It does make one wonder though to what extent automating that customization optimally might be possible. It would be cool if every time a different driver sat in the car the mirrors, seat, wheel, and whatever else automatically adjusted for maximum comfort and safety.

Settings can be well designed or poorly designed though. In a way, settings are really part of the design and should be thought of as such. If the settings are obvious to find and obvious in utility, it's not a matter of willingness to fiddle, it's just part of using the software.

The modern UX ethos seems to be that settings indicate a failing of the product and if people don't like the path laid out for them, they're either wrong or they can go use something else. I think the book also makes the argument that constantly breaking workflows isn't very user-friendly either, but SaaS products routinely change their UI and workflows around. I suppose I can see the business justification for that, but that doesn't necessarily make it good UX.

Of course, Norman could just be wrong or my interpretation of his text could be wrong. Either way, something feels off to me with modern UX. I suppose if you've run actual user studies (not just A/B tests), then you have more contextual data than I do. But, I came away from reading the book really wishing more software products (both SaaS and the recent spat of "opinionated" frameworks/tools) followed the principles he championed.

Exactly.

Don't think about settings as a failure.

Use reasonable defaults and let users override them in a reasonable way.

For example, my kids endlessly change the icons, colours, sound and many other visual and functional behaviour of their phones, the vast majority of which I would consider ‘settings’. But for them it’s highly important and fun

The defaults are fine for me, but changing those settings isn’t a cost to them, it’s just plain fun.

It’s easy to dismiss that as not real work, but I see the same in workplaces. Being able to customise your tools to be that little bit more bearable can sometimes be the difference between having a great day and crappy day at work. A little bit of control can feel very empowering.

Imagine there were no mouse/controller sensitivity settings for example. These and many other settings can have huge impact on how usable something is.

Or a setting that once set doesn't stay set. If I arrange the settings for something, especially part of a UI, I do not expect those setting to go back to defaults every time an update happens. Somewhere between functionality and not having functionally is only maintaining functionality through constant user effort to rebuild settings.

Like the rental car I had that constantly reset the heated steering wheel every time I got in. I don't want you to be on. I told you I do not want you on. So stop turning yourself on every time. THGTTG predicted this frustration decades ago (the Nutrimatic drink machine scene).

Passive consumers don't.

Power users, domain experts, creators, engineers, and owners love settings.

People have no problem with settings as long as the settings are reasonably easy to understand and actually help solve the problems they have.

In fact, Apple setup a store that did nothing but sold settings…the iPhone ringtone setting.

Other countries have had mobile companies selling more settings. Wallpapers, hello tunes (messages that play when you call someone instead of a ringtone), etc.

Settings are not inherently bad. Unnecessary settings and bad defaults are bad.

If it's something I use every day for years on end to make a living, then I probably want to be able to tweak and customize it (and I won't mind doing it). Just don't make me keep doing it because your evergreen software keeps updating and changing how things work.

If it's something I "barely" use, then I have absolutely no interest in customizing it. It should just work.

For example, if all I want to do is securely share a file with someone, then don't make me jump through hoops to use it. If however I was looking for software that let me share organized groups of files, should sort in a specific way, would expire after a certain amount of time, and to track engagement with these files (downloads / views), then are probably a few settings I will want to tweak.

What I think often happens is software will evolve / pivot to do more (maybe sometimes less?) then it was originally designed to do. Unfortunately that evolution some times lacks a different design perspective. You might have been religious about clean and simple in the original product, but now it's a bit more nuanced, and you'll need to be more flexible. I don't think there's an easy solution. We just need to recognize that as features are added, design still needs to be part of that process.

The cars are delivered with handbrakes on, wrapped in with wrapping paper/red ribbon. Some people do not realize they need to release the handbrakes and remove the wrapping paper. They are driving for years with default settings, wondering why the mileage is bad and it smells like burned brake pads.

I set about analyzing the problem over a period of a week or so and documented my findings with some recommendations. It ultimately boiled down to far too many settings, most of which were footguns, and no sane defaults that served most users most of the time.

In the end, I instructed the development team to make a number of changes which were mostly around picking defaults, or having the import process use some heuristics to make a reasonable guess, but letting users override them if the software did something wrong -- we modeled very heavily off of various MS-Office import processes, particularly MS-Access and MS-Excel. We ended up getting what was something like a 40+ step process down to 3 clicks that more or less eliminated the support tail and turned training into a 3-5 minute show and tell.

One of the problems with lots of software IMHO is the lack of sane defaults, the blizzard of hard to grok options and settings, and the multitude of ways of doing nearly-exactly the same thing.

Sometimes, there's multiple viable ways to use a piece of software. That's settings. That's facts.

Looking at you Microsoft

Have as much settings as you want, sensible defaults are important though.

Really, the list of hidden features on products is such that nobody has solved discovery. Affordances that work are ones that mimic already learned behaviors. But shared learning is not as universal as folks think it is.

Interactive systems that let you ask "what will this do?" Or "why did that happen?" Are very good. But even that is hard to work with sometimes. Consider how few folks use the apropos utility.

how many users is this in proportion of the silent majority which has no trouble finding what they want in the menus ? you can't base a judgment on complaints alone

Some people want to take the effort to become proficient with their tools, others are doomed to wish they had features that are right in front of them.

But being less powerful, efficient, and adaptable to keep the design clean and simple and make everything discoverable should be an anti-pattern for serious apps and tools.

I only wish it applied to more places. E.g. it searches help files, but they're basically always worthless, and it doesn't search non-menu things (settings, in-window toolbars, etc). Intellij has a nice cmd-shift-A which does most of this ("search everywhere" iirc), and I use it absolutely endlessly.

Any modern application that has more than 20 settings should have a good search function in its settings menu. That way you can have as many configurable options as you like without them being impossible to find. The defaults should ofc be the most common (or best) ones, but some users really do value customizability and respecting the users is (almost) always a good idea.

If the software doesn't do something they want, most users will either use it as-is but be somewhat annoyed, or develop a (possibly manual) workaround. Most will not explore settings or read instructions, so spending a lot of time on "discoverability" is probably time better spent elsewhere.

> If you can do the right thing automatically

I think a lot of software will shoot for doing the right thing automatically when they don't actually know what the right thing is. Or they'll remove the options because they think that the users don't want settings at all. But users do often want all of the behaviors (even the contradictory behaviors) that settings enable -- they just also want those settings to be discoverable and intuitive, and ideally they don't want to think about what they're set to most of the time.

As an analogy, as a user I like automatic high beams in my car if they work well. And having an automatic mode that's turned on by default might mean that I don't need to spend as much time messing with my high-beam brightness, and that's great. In a world where they worked perfectly, I might never even need to learn how to adjust the beam brightness myself. But I still want the behavior of different beam brightness in different contexts.

There's a trap designers fall into sometimes where they say, "settings are too hard for users, therefore the headlights should only have one brightness." If a bunch of users are asking you about something, that means they're engaged with and care about the functionality they're asking about. It might of course indicate that different defaults should be set, or that UI should be reorganized. And if you can do the right thing automatically, then you might be able to get rid of a lot of those support calls by doing that instead. But make sure you actually can first, because users are signaling with those support calls that they do care about that feature/behavior.

I wonder to what extent command palettes are the solution. At least for me, they are the pinnacle of discoverability. Throw the list of all settings and actions into an easily discoverable command palette and let the user search for them by filtering the list. Make it fuzzy enough to account for different formulations of the problem. VSCode does it (mostly) right.

A hard to find setting is better than not having the feature to begin with. The main prblen with "modern" UIs is that they are strait up removing feature and acting like its a good thing.

I know it's partially off-topic... but Instagram removing the ability to view content without and account and pretending they did it for the benefit of the users is pure BS and it gets even worse when you consider how agressive they are at requesting your phone number.

Arguably until users are willing to fund good R&D - and there aren't other unreasonable pressures on the production, and that design isn't perverted by designing to optimize for ad revenues and to capture engagement of a person to unhealthy levels, then it's going to continue to be the status quo.

Seems to me we're regressing to the view they had before this story: "If I just had more data, I could find out the real seat dimensions that will fit everyone! Then I can finally do away with all those pesky knobs and levers!"

Not saying its a bad idea, I just don't think it'd fly with some of the anti-settings camp

From a user point of view, the problem is that satisfying beginners will get more click/impression/downloads/etc.. If the company is managed by accountants, you end up with products that get shittier every version.

> I remember, back in pre-windows XP days, me and some of my friends would spend ages going through the appearance options and customisimg pretty much everything. Even later, I knew of people who had a meticulous desktop ordering or spent substantial amount of time ordering favourite icons in the browser.

> All that was great fun until the first time, the settings were lost. Maybe you had to reinstall the OS (or got a new PC) or maybe the program updated and simply erased the settings. In any case, trying to redo everything you had arranged before looked to be an enormous amount of work and not fun at all. So after the second time, this happened, we gave up and accepted the standard.

I was like this, and felt exactly the same. When I discovered the magic of unix windows managers, it felt like a miracle. Suddenly, I could adjust every frivolous detail to my liking, and _have it reproducable_ in a version control somewhere. When my laptop broke, I just reintroduced my package list, inserted the config and it was just like the old one! I think it's a shame so relatively few people will ever reap the benefits of a customized system over the fear of losing it randomly to a hardware failure.

I used to love to fiddle with settings. I gave up because the constant churn of modern software made it too tedious to maintain non-default settings in most applications.

What use is spending time to make a more comfortable environment when the software will randomly auto-update in two weeks and might undo all my settings, or worse, change its behavior in such a way that my custom settings just make things worse.

Non-technical users may not care about settings to begin with, partly because defaults work well enough. But there will be a point where they'll have to change them, so giving them the option is important.

Just because most users don't use a thing doesn't mean it's not a useful thing to have. Even if almost never called the emergency services doesn't mean I don't want or need the ability to do so. Further, being able to adjust the cursor speed is great for improving accessibility.

> If you have sensible defaults, you can make users do actions that you want them to do

What about what the users want to do? What entitles you to decide what is good for other people? That's just creepy as fuck.

I wish I could align my springboard icons to the bottom. And choose a shiny metal shelf again instead of that acrylic bullshit, but that is just my bad taste.

I'd call myself a power user and to an extent I am (give me a deb or a PPA over snap and the app store any day) but tinkering is just a pain in the ass now.

Sometimes, just sometimes, users want customizability.

>> First of all, remind yourself that users love settings.

Maybe this is a better way to explain it: Some users like settings, some users have to change settings to use your product. But most users will not change them if you pick good defaults... for most users.

Some settings are even more important than others - like changing default browser in the operating system - that even if no one ever changes them they should be there basically by law.

Funny how often this gets repeated. Usually (as here) with a conspicuous lack of supporting data.

> They do not change their mouse speed, their background image, their default browser. That's why a meme exists of people thinking the internet is the blue E icon on their desktop.

Could just as well be -- and, I think, is -- the other way around: It's not that they're ignorant because they don't change their settings, but that they don't change their settings because they're ignorant. Maybe most (or even all) of them would want to change their mouse speed, background image, and default browser, if only they knew there are settings to do it. Removing the settings wouldn't exactly help with that -- on the contrary, the solution would be to make them more obvious.

(there are other reasons too of course, sometimes not really good ones)

as long as the app does what they want out of the box, they love unnecessary stuff like setting a custom background image or theme color.

Yes, it's an exercise for the developer to know their audience and their affinity for settings.

As a user of software, I agree with the author: I am generally happier with software that is highly configurable to my preferences versus software that is rigid and limited to the tastes of the developer.

Of course, this is a continuum; I am not so obsessed with configurability that I build my own customized OS from source. But a healthy settings/options/preferences panel is a way to earn my interest as a user.

Native means that interactive elements like toolbar items, nav or icons are just show the regular cursor and links in text use the pointer (Mail app)

Web style is what is common with many web apps and sites where every interactive element uses the pointer (Slack).

The reason for the setting that people seem to be one way or the other on which they are used to.

This oversimplification is dumb. My friends and family feel dumb for not knowing how to naivgate the dumb UX decisions

Sure, I can disable pointer acceleration, and I can change the overall scroll wheel speed, but I can't disable acceleration. I feel most comfortable with 1 scroll click being 3 lines of scrolling. It should ALWAYS be that way, no matter how fast I move the wheel. If I quickly flick the wheel and do 5 clicks, that should be 15 lines of scrolling. If I'm very slowly going click....click....click....each click should be 3 lines.

Instead, if I scroll quickly, each click might give 5 lines, and if I scroll slowly, each click is a pixel or two, and there's no option to disable this awful scroll acceleration behavior, even on the command line.

I also love settings with the exception of web browsers for some reason.

For me this goes further. Anytime I rent a car I browse settings. "Oh cool, this car has lane-assist, and radar-powered cruise control". People look at me weird.

I think this is the same mindset that drew me to linux 20 years ago, when compiling and tweaking your own gentoo was not-crazy.

"Let's just say, if your VCR is blinking 12:00 linux is not for you"

Your examples are valid except when zip code lookup fails or wifi lookup is not correct and the systems tries to think for me. It takes away that option. This thinking is a disease and the reason I am left with half assed systems that i ahve to work around for my friends and family everyday.

A friend felt dumb for not knowing how to plugin a line out cable on her iphone because apple decided 3.5 mm is outdated. I told her she isnt dumb and it is not her job to know that.

This is thinking is causing more dumbing down of people than anything else.

My parents feel dumb for not knowing how to use a smartphone. The dumb android dialer calls randomly and goes to background with no way to hangup. My dad cant figure out what to touch to close it. With a physical button he can close it easy. Windows phone was miles ahead than latest ios and android in call screen which is what a phone should be for. So are "feature" phones. As a cutting tech engineer, I prefer those over the ad mining machines that are iOS and android

My mom feels dumb for not knowing how to order on a website because everything now requires a login which requires a unique password and 2fa and a unique workflow with godawful hidden icons that every UX team wants to lay their marks on and eveything useful gets moved to the overflow menu, the gear or hamburger on some random side of the page.

Just have clear wordings with accelerator keys and tab bindings

Also because Joel spolsky is right on some things, doesnt mean we should listen to Everything he says. That leads to a monoculture and people are different and diverse. So should be designs. They should be more accomodating of differences, isntead of the current my way or the highway of restricted UXs and walled gardens

https://www.joelonsoftware.com/2000/04/12/choices/

https://www.joelonsoftware.com/2006/11/21/choices-headaches/

A setting might have been added because the creator legitimately couldn't make up their mind about a feature. Or because end-users have divergent use-cases, preferences, abilities, situations, or hardware. Or because the creator felt that the only way to support two different use-cases was to make two different interfaces with a settings toggle.

Each setting has an incremental cost in terms of development, testing, maintenance, documentation, initial onboarding, finding other settings, etc. The best response is to consider whether each setting is sufficiently valuable, and whether there are any settings that can be eliminated with better defaults or more flexible interfaces.

Because many people don't have time to do that, products typically end up at one of two extremes. Either they have no settings beyond the ones the creator themselves uses, or they are the union of every setting anyone has ever asked for.

The chance is high that the "almost no one" group is actually the ones that use your app the most. Casual users won't care much about some setting five layers deep in the menu, but the ones who are dedicated enough to actually find it will care.

The constant "optimization" by taking features away that "no one" uses is what drives so many users away from installing updates. Please stop that bullshit. I'd like to not fear updates again.

And I fxxking hate it.

Because I am the user that will open many tab and make the tab small.

So does this mean I should not use firefox because I use it wrong? It's really a shitty attitude if you are exactly the one that they claims 'not-important'. Pray for you not being lucky enough to be the one 'not-important' if you don't know the pain yet.

So the newer settings dialogue changes more stuff about the desktop than just the background?

How would the menu label "Change Desktop Background..." make all that other stuff more obvious or discoverable than "Desktop Settings"?

The author of this article is saying they disagree with that orthodoxy.

I don't think it's helpful to just repeat the orthodoxy without saying more.

In particular I'd be very happy never to see the claim "people only add settings because they're shirking their job of making a decision" again. I don't think it's true, and I think it's impolite to make such a claim without justification.

Is it really fair to your users to presume the are clueless? Is it a good idea to take all decisions from them, and rob them of the ability to become a designer?

I have a suspicion that this is a self-fulfilling prophecy: take away agency, and you'll end up with a demographic which never wanted any.