It definitely doesn't apply to Netflix. You can get good raises and promotions by optimizing an existing process. If you can show improvement in internal metrics, such as improving latency or lowering cost, that's a huge win. Even if you increase costs you can get a raise and promotion if you show that it increased code quality, such as creating the Chaos Monkey/Gorilla/Kong.

I guess it helps that Netflix essentially has one product.

Yes that probably helps a lot. Also they don’t hire fresh college grads, so you generally have less of the “wow this was clearly written by a junior engineer” type of maintenance. The low hanging fruit is mostly picked when it comes to architecture and optimization.

I work at Amazon and I don't believe that applies here. There certainly is an expectation that you can coordinate increasingly broad efforts as you rise through the ranks, but the focus isn't on launching new products (in my experience). Impact can be anything that is positive and measurable. At the end of the day if you make the company $1million then it doesn't really matter where you do it.

Oh just go ahead an name Facebook.

The irony, is that this pattern at FB and and Big G isn’t indicative of a good engineering culture. It’s pathological engineering.