[hbn]

> They had implemented that already a while ago, then reverted the behaviour, and now implemented it once again.

This is, among many other reasons, why I finally dropped Android after the better part of a decade. The constant A/B/C/D/E testing makes every single thing they put out feel like it's a constant state of beta testing. It's to the point where you don't even know what to expect when you do something as fundamental as opening the app store. You'll seemingly have some kind of server-side flag activated one day that gives you a totally new UI in an app you use every day, hiding things or removing features you rely on. Then maybe in another few days it'll be back to how it was.

Not only do they not seem to value their users, they actively punish you for being one of their users, jostling you around between new UIs or even entire services that are always worse than the last.

[roughly]

I heard a phrase a while back: “the subtle gaslighting of A/B testing” - that feeling that you’re pretty sure that button used to be over there, or the app used to have that function, but not entirely sure, because one day it’s just Different, no release or upgrade or reinstall, just - it’s not the same anymore - or, is it?

[ephimetheus]

This is super frustrating. I’ve described it like feeling like you have Alzheimer’s as everything changes all the time.

[nescioquid]

My wife once handed me her newer-model android phone to fix something for her. I thought I was having a stroke because I couldn't find the settings icon anymore.

[mdp2021]

> the subtle gaslighting of A/B testing

It surely would be so... What could trigger that in an application? Some of us have never seen any such behaviour (fortunately for anybody responsible and huntable). Maybe said applications are web-based, mostly front-end?

[ephimetheus]

The native Revolut app (at least on iOS) had been doing this excessively for a while. I think they toned it down now.

[roughly]

Ain’t just web. Plenty of well-defined feature flag frameworks for mobile, too.

[gifnamething]

The lack of valuing their customers is what made me finally give up on Android. Android's biggest problem is the same as it was 5 years ago - the support doesn't last for long enough - and all they've done about it in that time is some half-hearted upstreaming of <1% of their kernel patches (project icebreaker) as yet-another skunkworks alternative to an existing project. The attitude seems to be that they assume Android will always have its market share and the users are captive. So just chug along in mediocrity and let the e-waste pile up.

[2OEH8eoCRo0]

It's funny how smart yet dumb A/B testing is. On one hand you can intelligently gauge the effect of changes, on the other hand you can push stupid shit since you have that power.

How about intelligently designing applications that you yourself want to use? Too hard.

[eurasiantiger]

That would require designers to deeply understand system and software engineering, or systems and software engineers to understand design. Funny how that gap keeps on manifesting itself, even though both parties work on the same domain.

[notriddle]

> How about intelligently designing applications that you yourself want to use?

The people who use the apps I make aren't always me. Many of them are very good at things that I know nothing about, but I usually know more about computers than they do. Sometimes the differences are even harder to account for, having different preferences, different ways of thinking, and participating in different cultures.

[collaborative]

All the A/B tests I have done point to one conclusion: assume users are illiterate and have an attention span of 5 seconds

[eurasiantiger]

Keep it simple, stupid.

[dtgriscom]

There are facts about human behavior which can only be learned by actually testing them. People are complex in ways that you, whilst sitting at your keyboard, cannot figure out.

[tornato7]

All which can be learned through focus groups and internal feedback, no need to further confuse and frustrate your users.

[szundi]

After you have the engine running, the marginal cost of an A/B is probably smaller than a focus group test. So if you really believe in focus groups for the results and you are also ambitious enough - probably this means go for A/B.

Disclaimer: just thinking, not knowing the costs

[tornato7]

Definitely agree, though I think the point of this thread is we don’t know the “true cost” of A/B testing in terms of long-term lost user engagement.

Have we done an A/B test for A/B testing where we test A/B testing…

[fshbbdssbbgdd]

It’s very common to think something is a good idea based on focus groups and internal feedback and then learn it’s a bad idea from launching it (or running an AB test).

[ffhhj]

> one day that gives you a totally new UI in an app you use every day, hiding things or removing features you rely on. Then maybe in another few days it'll be back to how it was.

We are approaching the age of Schrodinger's Apps.

[cmroanirgo]

The alternative is iOS, a more consistent UI/UX, but you lose out on projects like F-Droid, where you can bypass Google HQ nonsense.

[hbn]

If you're running Google Play Services, you're not really bypassing Google HQ nonsense.

[devit]

Just don't use proprietary applications (or don't expect them to serve you).

[hbn]

I switched to an iPhone and I'm using proprietary applications, but ones that don't randomly change their UI on a regular basis whenever some nameless product manager decides they want to use me to gather some new engagement metrics by rearranging UIs on my phone

[heleninboodler]

They do make arbitrary changes to the UI, though, and when that happens, you can be damn sure it's because some turtleneck somewhere decreed that the new way is the One True Way and there's no amount of user complaining that will fix it.

[shrimp_emoji]

No, they just entomb you into a comfy walled garden where it's only easy to do what Apple lets you do and where you hope some change made by some nameless product manager/CEO autocrat doesn't force you to buy more expensive hardware.

[shadowgovt]

Yeah, it's pretty great. Best walled garden on the market by a country mile.

[serf]

> Yeah, it's pretty great. Best walled garden on the market by a country mile.

a lot of us avoid supporting such behavior from corporations because we view it as unethical or immoral and damaging to the sector in general -- regardless of how good the ux/ui may be.

[shadowgovt]

The nice thing about living in a world of free people is that is a choice a person can make.

[jchw]

That’s funny. I mean I use iOS and sure the UI is nice… but if you are thinking there’s not needless silly UI changes… how many major versions have you been through?

Some of the bad phone UX ideas started on iPhone. Like removing the physical button at the bottom in favor of annoying gestures and no touch ID. And iOS 7 removed all of the borders everywhere, it’s arguably more radical than Google Material, a UI design I also am not really that fond of.

I guess if you mean there’s no A/B testing or it moves slower then probably. But, it definitely moves. That becomes apparent any time you load an app from the App Store that hasn’t been updated in a while and suddenly your phone looks and feels like it did 2 years ago across the whole UI.

[hbn]

They've definitely changed the overall look, but a good amount of their apps are pretty much identical in how you use them from the original iPhone in 2007. Notes, Messages, Contacts, etc are all relatively unchanged, except for additional features. The biggest overhaul was probably the recent change to Safari where they brought the address bar to the bottom, which was a consideration based on how big phones have gotten, and allows you to reach things easier.

And at the very least, these changes come from normal app updates either from the App Store or OS updates. And it's usually a pretty big deal when they change something, and gets a formal announcement months in advance where someone high up gets up in front of the world and pitches why the change is an improvement (not to say it always is). Whereas Google just randomly shunts out new UI updates on a regular basis and enables them for random people. Usually someone posting about their new UI on reddit is the first place you'll hear about the redesigns.

[bjelkeman-again]

At least with Safari address bar, I found a setting to move it back to the top.

[WorldMaker]

I appreciated when iOS Firefox added a setting to move it down to the bottom. As an old WinPhone user, I missed having that key navigation tool at the bottom. It really does make one-handed phone usage easier.

[ajross]

> ones that don't randomly change their UI on a regular basis

I'm genuinely curious what apps you're talking about here. Everyone does this nonsense. Everything changes all the time. Everything. I don't like it either, but to state that it somehow doesn't happen in the Apple ecosystem seems like a pretty big whopper.

[hbn]

Here's a random example: I think it was like 2017, 2018 when Google launched messages for web. For starters, when it launched, it was located at messages.android.com

I don't think it was much time later before they moved it to messages.google.com, which i think was in line with their SMS's apps like 5th rebranding, this time from Android Messages to Google Messages.

Originally the app had an overall blue theme, and for individual contacts you could change the color of your conversation with them so each chat thread was themed. This even had the neat effect that it would sync with the web version. However, it only lasted like a week maybe before Google completely redesigned Messages to be all white themed, and killed the chat themes entirely so all message threads were now blue and white to look like the iPhone messages app.

That wasn't a rare experience, and I haven't encountered anything like it since switching to an iPhone.

[richardfey]

Wow..imagine the churn amongst the teams working on this

[pavel_lishin]

Isn't that equivalent to telling us not to install apps at all?

[devit]

Use free and open-source ones like the built-in AOSP apps or the ones on F-Droid.

They have no incentive or resources to do A/B testing or make unnecessary UI changes.

[mdp2021]

I guess the poster meant, "either use Open Source or code them yourself".