There's a large social component, too: beyond simple resume-driven development engineers get social status by doing interesting technical challenges. You rarely see a project make the front page of HN, for example, because it has a really polished UI and great documentation.
Since that's what the community generally respects and focuses on, people seek out areas to add complexity – who wants to work on a boring “solved” problem? – and, especially younger engineers, often feel they have to do things the same way Google or Facebook does (it's web-scale!) even if they have 0.001% of the engineers and 0.000001% of the traffic. It's easy to miss that most people with similar problems have so much less volume that they aren't forced to do the really cool (i.e. expensive) engineering and, since they're using something stable, it's not considered interesting enough to talk about it.
That's not the same as under-engineering. They simply engineered for a different target. The engineering of the product is only part of the issue. The bigger issue is engineering the entire production system. They may well have optimized for maximum return on investment in the factory, not maximum quality for your headphones.
* no one ever got hired for under-engineering
* a lot of people have died from under-engineering
* you never notice how much was engineered juuuuust right (tm). Because it was. Just. Right.