A premature decision is the worst decision of all.
In my experience, that's wrong: No decision is definetly the worst decision. If you take a wrong decision, or a premature one, you can always fix it ... but many people/managers tend to wait and wait - till the time of decision-making has passed (and then they are surprised that they get nothing - not even an 80%-solution which can be improved).
Just my (professional) experience - yours may differ!
To be fair, marketing and product strategy by necessity involves lots of guessing. Ideally those guesses are turned into testable hypotheses, tested, and the results used to make better next guesses.
Product research and development, even in IT, is often the same but using different data and metrics. There's no one right way that's readily apparent to build a finished project. We build mock-ups, testable units, unit tests, integration tests, and end-user quality tests. We reuse what works and throw other things away.
Anyone afraid to guess and test is just going to make clones of other products. Anyone who doesn't test is just guessing, which is less successful but may work if you're really good at guessing and really lucky.
Uhm, i think the answer to the poll really depends on what part of the IT business you are in. If your duty is to keep some undocumented, crappy, closed source software running, then guessing in your best and often only option.
On the other hand, nothing good has ever come out of "i guess this encryption scheme i just made up is secure".
"For me, programming in Perl is like my cooking. The result may not always taste nice, but it's quick, painless and it get's food on the table."
Perhaps if you start with a premise(guess) and you try it, and test it and it succeeds. And then you test it some more, then give it to someone else to test and it still succeeds. Then you give it to a small group of people to test and it continues to succeed, you may be onto something. Then you can start developing you premise(guess) in earnest.I guess.