Because it is so good for most, but apparently not good enough for a high enough percentage of people. We don't have much data on that and most of what I do have from treg testers is old and thus may not still apply. I'm just going with what I have.
I can't begin to say what level must be achieved, but just as a theoretical, suppose you have 10% with problems. In some ways that would be considered a low number. But I don't think 10% would be considered good enough for a shipping product. I know, when I worked in a manufacturing plant, it sure wouldn't be good enough! I know a bit about that since, besides working there, I had relatives who were foremen and engineers.
So, let's work our way down to lower numbers. I can't say what that needs to be, but let's say 2% with problems. And you have orders for hundreds of thousands already and more to come when shipping starts and more people hear about it. That 2% is going to mean 4,000 people (assuming just 200,000 to start) with real issues plus the following orders. There will be many others who, because of the new tech, think they have issues, but are just user error. But they will be just as upset as those with real problems. I suspect this will at least equal those with the real problems. So now we have 8,000 who are calling tech support with a small company. The result is a lot of criticisms of the product and service all over the internet. Upset people tend to post more than satisfied people so the influence would be much greater than 2% who have those real problems.
Don't know how big their tech support will be, but I doubt it will be anything like they'd prefer if the company was well established.
They also have to consider the effect when you are bringing out new tech, especially from a new company. Samsung get survive explosive batteries because they are big enough to absorb even that huge loss in sales. And have a good enough reputation where most people won't write them off permanently. And smart phones aren't a whole different way of doing things anymore.
WT has none of those advantages. Are they being too conservative? I have no idea. No one outside of WT does. None of us has nearly enough information to determine that. But I can say that I've seen enough from testers to say that I can see not shipping yet MAY be the right thing to do.
I know, you don't have access to the info from testers that I see and you will be much more frustrated as a result. I'd love to see more testers. A lot more. But that may not make business sense. Just what I'd like to see.