How good does you software have to be?

I've got this PC video game...
Besides, thinking that it might work yourself and a few friends agreeing, how do know if your software is good enough to sell?
Also, saying that it is; how would you go about doing it?

How do you decide how good a program is?

If it's full of bugs, it's not good. Other than that, it's highly subjective. A game is good if it's fun enough that enough people buy it that it was worth producing. Other software is good if it's useful enough that enough people buy it that it was worth producing. Often, only actually making it available can tell whether this is the case.

That said, I have no idea how you would go about publishing software.

A wider beta test for say 100 to 1000 people.

http://en.wikipedia.org/wiki/Development_stage
I would say what you've done so far qualifies as alpha stage.

Companies usually have testers on staff to do initial testing of their software (test for bugs, see how good the game is, etc.)

Then, once the product gets far enough along, many companies will open up a beta testing stage in which more people (often outside the company) will be able to test the product and give feedback.

To expand on what others have said bugs are normally rated by seriousness. If a bug rarely shows up then it is normally not a priority since most will never see it. However if the bug is very serious and yet very rare this could warrant attempting to find its source. Frequent bugs and easily produceable bugs are high priority and should not be in a shipped product.

Given the size of programs today I would venture to say it is nearly impossible to release a completely bug free product.

As far as the marketability of the product I have no idea how that is all figured. I would guess it's done via various surveys, sales figures, etc. Nearly every game I've bought has some type of survey when you register. This information combined with sales of different genres of games gives developers a fair idea of what games will sell and what ones will flop. For applications I would say much the same is used to determine marketability.

Every now and then a company comes across a sleeper app or sleeper game that just sells like wildfire when it really was not expected to do well.