| ▲ | dccoolgai an hour ago | |||||||||||||||||||||||||||||||
This form of A/B testing or "optimization" if you like (unaccountable 3rd party script spawners) should probably not exist in the modern agentic era. It was a flimsy premise _before_ but borders on malpractice now. Even as a business case, using these kinds of frontend-based frameworks slows down performance (which is known to impact conversion/ ux) for the sake of letting some twit change button colors from red to blue (which will almost certainly impact conversion less than the performance loss). It's like saying "let's drop conversion by 4 percent by ruining performance and then see if we can raise it 1 percent by changing button colors". There's no reason for it on the modern Web. | ||||||||||||||||||||||||||||||||
| ▲ | teeray an hour ago | parent | next [-] | |||||||||||||||||||||||||||||||
I also cannot describe how much it fills me with rage when I am part of a UI experiment with no way to opt out. I think this is most noticeable with Facebook’s apps. I usually just do the opposite of whatever action they clearly want me to take to do some small part in ruining the results. | ||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||
| ▲ | robrenaud 16 minutes ago | parent | prev | next [-] | |||||||||||||||||||||||||||||||
The testing cost is temporary and the gains are permanent. | ||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||
| ▲ | deepfriedrice 8 minutes ago | parent | prev [-] | |||||||||||||||||||||||||||||||
My marketing team had a contractor run AB tests on my static pages via Optimizely. I had to insist that they run each test by me before going live. They would essentially re-implement my design system in Javascript, inevitably with tons of mistakes, in order to build their experiments. (In hindsight it could've maybe been a use case for web components). Fortunately it was shut down after enough people questioned the statistical significance (there was none). | ||||||||||||||||||||||||||||||||