Starship Flight 7

Still, many of these are more successful than Starship:

The first GSLV was still able to deploy a satellite, just in a lower orbit than intended.

The first Delta IV had the same problem, satellite deployed, but in a lower orbit than planned.

The first Long March 5 is classed as a full success on Wikipedia, I couldn't find info there about a failure (the second one did blow up).

The Rocket Labs' Electron did get destroyed. However it was later found that nothing at all was wrong with the vehicle, it was a failure in the ground software, and an identical vehicle successfully carried out its mission 7 months later.

In contrast, the first two Starships blew up completely due to engine issues, and no Starship has deployed even a test payload of some kind to orbit. In fact, until today, none even carried a payload of any kind, they have all been flying empty.

The Wikipedia entry describes it as "suboptimal but workable initial orbit", which I interpret as a partial failure (coming from a military entity that's universally opaque about its failings). They're not inclined for language like "partial failure" that we get out of transparent countries—contrast that first Delta IV-H, which also reached a "workable" orbit—just not the intended one.

- "However it was later found that nothing at all was wrong with the vehicle, it was a failure in the ground software, and an identical vehicle successfully carried out its mission 7 months later."

Also true of the Ariane 5 explosion: that was a software bug (unhandled integer overflow) in the flight control unit. The important part isn't whether it's hardware or software, but whether they got it right or not, before launch.

They have explicitly and publicly chosen to rapidly iterate without spending billions to make sure the first try goes well - it's simply different culture. The first Starship wasn't even something you could actually call a rocket, it was a water tower with a bunch of rocket engines.

They wanted data about the engines and got them - mission 100% accomplished, that's not a failure in any way except for media shock value because "wow such boom". Come on, you call yourself an engineer? Do you not try your software or hardware before 100% completion? You don't have CI with integration and e2e tests? There's no other way to do this cheaply and quickly, you have to try.

Call me when any other company achieves what Falcon9 did, then we can discuss issues of SpaceX engineering culture and how others are better. But they are not, few test flights are not interesting, what's interesting is that they are 10 years ahead of everybody else and offer by far the cheapest and by orders of magnitude most reliable orbital lift service.

Others should stop waiting 10 years before the first flight and accept some risk, the world would be much better by now.

Your definition of success doesn't leave room for anomalies. Your mindset seems to be "if you try and it's doesn't turn out perfectly, it's a failure" -- which results in spending tons of time and money iterating behind closed doors (or even worse, trying to model/calculate the whole thing without many test runs), and only unveiling the result when it's "perfect". This approach costs more time and money, and more embarrassment if/when the product fails in public. It also doesn't build a culture of learning a lot from anomalies.

Meanwhile, SpaceX doesn't care about iterating, testing, and failing in public. So they skip all the costly effort of trying too hard not to fail, setting expectations that they get it right the first time, and not learning as much from anomalies.

Anomalies, properly understood, are opportunities to learn and improve -- and never something to be ashamed of. The only true "failures" are to give up because it's too hard, to stop learning from the data that anomalies provide, or to never try in the first place because you're too afraid of anomalies.