Another fun thing people may not realize: the best UI / UX is not necessarily the most profitable. Apps and websites optimize for conversions and sales sometimes at the cost of better UX. The prettiest thing also usually does not perform the best.
If you run into a website that looks ugly as sin and has been around for ages, it’s not that it’s neglected, but likely been tested to perform best in that style.
You bring up an interesting point about optimizing for conversions and sales at the cost of better UX. It's definitely true that the most profitable design is not always the one that's most visually appealing, and many times even though a website may not look beautiful to every user it can perform exceedingly better in usability test which impacts positively on SEO attributes drastically increasing adoption especially by audiences who prefer simplicity with the ease of fine luxury.
Similarly, sites having endures years while keeping decent growth patterns suggest validation of tested styles conforming set scalability constellations upheld over set trends often being economical with effects widely not as comfy as they are broadcasted against their loyal clientele. Hence experimenting shouldn't draw you too far out from affirmed designs fostering attentional efficacy valuing practicality with superb outcomes overriding overly-aroused perceived sophistication leading to subparsystems scarcely making heads turn.
Depending on the industry, ugliness may be required. Coupon sites are one such example. One time I redesigned a big well known coupon website (most important pages) and we quickly found out it wasn’t doing as well. People preferred the ugly thing they were already used to.
While there’s no easy way to determine that the pretty styles had anything to do with it, we typically make hypotheses and test them. In fact, large websites and apps are optimized one element or page at a time with small and big tests. The winner is put up against another variation and so on.
Yes, I agree with this. It doesn't always have to be complete
Old reddit is a good example of this. The shinier UI got horrible feedback, but they still kept the update. Soon afterwards third-party frontends emerged.
