It sort of depends on which step you miss and what you're investigating. If you're trying to figure out the fastest route to your school and you skip over doing research (google maps, asking friends about what route they take, etc) and jump straight to designing your experiment, you should be fine. In that kind of real world application, if you've no idea which of your two or three options is best, (ie no hypothesis), it's not going to mean your experiment automatically fails. In a science setting, you DO want to have a hypothesis, it helps you refine your thinking and identify exactly what it is you're testing. Obviously, skipping doing the experiment properly can mess up the process, and if you don't analyze your results to come to some conclusion, then you've simply not finished which is kind of a way to mess up.