I think there's one thing that it is appropriate for: Quickly needed decisions with a "social background", if you are experienced at judging human behavior (I don't know the exact English word for "Menschenkenntnis").
Else: Don't do it. Do not decide based on pure gut instinct. Try to get something more dependable: Ask friends, ask experts, talk to people who know more than you or provide a different point of view and then judge their comments against each other. There's almost always a way to get the information you need if you have the time, and it's very seldom you don't have time.
Then it brakes down to experience and self-confidence, I guess. Luckily, my gut instinct is right, usually :-)
If you're interested, here's a link to a book about this topic:
One experiment I tried was for a very large multiple choice test. (For the Architect exam, 3 days of 10 hour tests). I found that when taking the practice test, 95% of the time that I went back and changed an answer, thinking I had answered it wrong the first time, I actually changed it from right to wrong. So for the actual test, I went through the test once only with one answer, left hours before every one else, and aced it. The moral: The Subconscious is a good thing - you can rely on it.
The moral, in that case, was that a marketer's instinct is a fine thing, but doing online testing gives you real knowledge of the situation. By doing testing not only can you learn if your instinct is right, but you can glean information about your users/visitors as well.
I did notice that my score tended to decrease if I triple- or quadruple-checked my answers, though.