I think you can apply the same towards design of anything. Tastes change. While a certain design philosophy might be a godsend in 2010, it might be considered trash in 2015.
To be fair, I also think a lot of design decisions are poorly made, and a lot of businesses make poor user experiences.
Just like female jeans that have tiny pockets.
We'd quickly branch into 'busines formal' UX patterns that change infrequently, only splashing a company color in places but largely uniform, all the way through to 'haute couture' UX patterns like the recent 'brutalist' trend, trusting they only appear on trendier sites and dropshipping fronts.
Objectivity is an ideal we can strive towards. And several crazy and expensive designs are worn in the real world. I have seen it with my own eyes. The people who are able to afford those clothes either:
- Have bad taste (A subjective judgement which is logically wrong, since it is absolute)
- Really like that look
- Don't care what anyone thinks of them
- Want to display their high status
- Want to stand out from the crowd
It's not an ideal, it's a legend.
They do, but do they change as quickly as every season?...or sometimes people do those who make money off changing tastes tell you that tastes change that often?
Female jean pockets being small is entirely a driver for handbags and the like. It’s a poor decision for practicality but entirely a business decision — it’s sorta a dark design pattern I guess similar to how amazon buries privacy settings
Like, “How many interactions/seconds does it take our users to find X?”
It’s difficult because you have to know what “X” a user was looking for, which requires some varying degree of asking lots of users and reading those responses to be accurate.
Measuring “engagement” is far easier to orchestrate practically and organizationally, even though it produces awful software.
Engagement is such a terrible metric. I'm not saying it should be thrown out, but it is classic example of Goodhart's Law[0]. Rage is engagement. Too much reliance on engagement is what got social media into a race to the bottom. That's not the mark of a quality product.
Measuring is hard, you're right. But if you're going to use bad measurements (which often is our only choice) they shouldn't be targets. We should understand where they fail and use the information accounting for that, not ignoring it.
Lacking objective numerical comparison, perhaps designers and product teams attempt to do metrics-based comparison from telemetry (we raised metric X by changing the design). That leads to an ever changing design that tries to optimise for business objectives, rather than what the user wants.
Backend code can be optimised and honed and improved with objective measurements (CPU cycles used, latency, perceived performance, download time for user), but design doesn't give these.
In saying that though, rewrites in new languages and frameworks do seem to be becoming the problem for backend - optimisation isn't seem as exciting or career advancing, I guess... Few are optimising code for performance these days though, I'd wager - everyone just assumes your user won't mind your software using 15% of their CPU when sitting idle.
Maybe it is just as simple as that metrics individuals are measured on for job performance?