[loandigger]

900 hours. That's my rule. It takes 900 hours of concentrated effort in brainstorming, building, tweaking, disillusionment, trashing it all, starting over, brainstorming some more, building, tweaking, demoing, networking and THEN selling before you'll even have the faintest idea as whether this can be a viable company. Put in the time and you'll see.

[mimixco]

Fred Brooks (author of The Mythical Man Month -- highly recommended -- and the director of IBM's System/360 project) said, "Build one you're going to throw away because you're going to throw it away anyway."

[aerovistae]

Have you formed some viable companies after 900 hours?

[loandigger]

Yes, three. One was a total failure. One went public a few years after i left and the last one was a private sale by me to a Fortune 500 company. I'm not saying my rule of thumb is for everyone, but 900 hours is about the length of an average internship. When I'm evaluating a new idea or opportunity, I try to imagine what I would need to accomplish in 900 hours to make it into a viable company==> Sort of a reversal of Parkinson's Law about how work expands to fill the time allotted for its completion. This helps me really clarify and plan out exactly what i have to do, so i don't waste time later or get side tracked. IMHO, 900 hours of grinding it out will give you the insight you need to see if you were really onto something or need to pack it in and try something else.

[g10r]

arbitrary number + that's what, a couple months? That's nothing.

[nyrulez]

Some one else mentioned 9 months above. I am amazed at these "10,000 hours" type of rules for doing a project or startup.