I recommend reading Alberto Savoia's book, "The Right It: Why So Many Ideas Fail and How to Make Sure Yours Succeed"<p>One of the examples that Savoia provides in his book is especially helpful to recall here:<p>Back in the 1960s, IBM was trying to figure out if they should make a massive investment in building a speech recognition product. Rather than betting the company on an expensive idea without being certain of the market for it, they simulated the speech recognizer by putting a secretary in another room with a mic, leaving the test subject with the impression that their commands were going directly into the computer in the room.<p>By doing this, IBM figured out that even if they made a perfect speech recognizer, their customers wouldn't use it! They were concerned about privacy, found typing to be faster and more precise, etc. Rather than investing huge sums of cash on a dud, they were able to put it towards the System 360 line, which turned out rather well :)