News for hackers. Stuff that matters.
Re:Ask HN: Dark UI Patterns in Gmail? by legitster on Friday July 21, 2023 @04:18PM (#36819100)
"Never attribute to malice what can adequately be explained by incompetence". I hate "Dark Patterns" because UX is hard and 90% of accusations are complete assumptions of intent.

Advertisements are almost universally served from a secondary/external source. So when you make a request to load your inbox, the ad may come in a half second or so slower than the native content. In a world where content is loaded in Async that means it may come in slower than your reaction time to read the page and click.

If you load up your console and look at the network requests, I would bet you a crisp $49 bill that you will not see all of the content loaded at the same time and the ad being delayed intentionally.

Re:Ask HN: Dark UI Patterns in Gmail? by BugsJustFindMe on Friday July 21, 2023 @04:27PM (#36819222)
Never attribute to incompetence instead of malice after a reasonable amount of time has elapsed for someone to have noticed and decided whether the behavior was undesirable and should be urgently fixed (not malice) or not (malice). Never attribute to incompetence instead of malice when a trillion dollar company that should have trillion-dollar-company-appropriate internal QA processes misses or ignores something blatant and terrible, especially when that problem is as old and as well understood as content shifting.
Re:Ask HN: Dark UI Patterns in Gmail? by legitster on Friday July 21, 2023 @05:09PM (#36819804)
Uh, I have worked cross team enough at huge companies to know that there is nothing magical about the size of a company when it comes to QA and bad UX choices.
Re:Ask HN: Dark UI Patterns in Gmail? by mostlysimilar on Friday July 21, 2023 @04:27PM (#36819228)
Choosing "No" instead of "No thanks, I don't like free stuff" is not hard. Choosing to give the "No" button equal weight and placement as the "Yes" button is not hard.

It's malice to hide the opt-out option in tiny gray text at the bottom of the modal but make the opt-in button huge, orange, and unmissable. Dark pattern, not incompetence.

Re:Ask HN: Dark UI Patterns in Gmail? by radley on Friday July 21, 2023 @04:30PM (#36819260)
> UX is hard

Odd. At one point, it was possible to specify the size of an image before it loads, thus reserving the space and avoiding content jumping around. I guess it could work if the UX designers were aware that there would be ads, so they could design space for them and then find a way to communicate that to the developers.

It might help if Google could communicate with the company producing the ads (DoubleClick?), so they might know in advance the exact sizes they will get. Even better, they could provide some sort of specifications that the ads be a specific size.

If only there was a way...

Re:Ask HN: Dark UI Patterns in Gmail? by awinter-py on Friday July 21, 2023 @04:28PM (#36819232)
also never attribute to malice what can be explained by hyper-optimization -- if all ad clicks were misclicks of this type, an AB test on a speedup might say to keep the jank version, without saying why

normally I'd say in-house users would notice any gmail issue + it must be intentional, but most in-house users won't get shown ads

Re:Ask HN: Dark UI Patterns in Gmail? by BugsJustFindMe on Friday July 21, 2023 @04:30PM (#36819272)
Hyper-optimizing for ad clicks instead of user experience is basically a synonym for malicious treatment of users.
Re:Ask HN: Dark UI Patterns in Gmail? by legitster on Friday July 21, 2023 @05:07PM (#36819772)
> if all ad clicks were misclicks of this type, an AB test on a speedup might say to keep the jank version

The problem is that misclicks are not actually super valuable for the ad purchaser. So if Google changed their ad behavior and your bounce rate shot up, you (or enough reasonable customers) would reduce your ad spend to compensate.

Hyper-optimizing for clicks would eventually be noticed.

Re:Ask HN: Dark UI Patterns in Gmail? by cameldrv on Friday July 21, 2023 @06:38PM (#36820912)
But everything is A/B tested. What starts out as a bug can end up organizationally unfixable if it has a positive impact on conversion.