There's a common anti-pattern that happens with what I'll call "short attention span" development processes (excessive reliance on A/B testing, agile sprints, etc, no long term vision). A poorly implemented, buggy, poorly documented, and/or not well integrated feature doesn't see much use so the developers avoid putting work into it, which just makes all of its problems worse until eventually someone has the bright idea to cut it because "nobody uses it". Sometimes this is warranted, but more often than not it's just laziness and short-sightedness. Nobody ever puts in the effort to fix broken core functionality, all the effort ends up dedicated to superficial turd polishing or gimmicks.