It's all about how you do it. I think the consensus is to fail small and fail often. Actually, scratch that, it should be: prototype often, so to keep failure

It's all about how you do it. I think the consensus is to fail small and fail often. Actually, scratch that, it should be: prototype often, so to keep failures small and easy to rebound from. Regardless, you do not want to spend countless hours and dollars bringing a product to market only to realize it was a failure 3 weeks later. That's a rookie "big company with lots of money" mistake that Microsoft should have outgrown decades ago. To quote Jason Fried on failure: "You might know what won’t work, but you still don’t know what will work. That’s not much of a lesson." After all their success, Microsoft should know better.