Testability
A scientific hypothesis, no matter how simple or complicated, must be testable. This is one of the important features or characteristics that a scientific hypothesis must have. If you produce a hypothesis that cannot be proven or falsified in an experiment, then that means that it is not an intelligent guess that can be proven and instead qualifies as a "hunch."
Falsifiability
The second important feature that a hypothesis must have is that it should be falsifiable--meaning that your intelligent guess, when tested, can be proven wrong or incorrect. For example, claiming extraterrestrial life exists can be proven correct at some point. All that is needed in the future is evidence. Although it can be proven right, proving it's false can be nearly impossible as human beings have no methods of examining the confounds of space to ensure that it is completely devoid of outside life.
Practical Importance of Testability and Falsifiability
Both features are very important in order to create a scientific hypothesis that can lead to a logical and applicable solution. The scientific method--typically used to solve problems in laboratories--needs to have a testable and falsifiable solution to make sure the solution is correct. If not, the solution applied can still be problematic and may often provide no proof of the hypothesis, rendering the experiment void.
Other Considerations
A hypothesis should be based on data, information and measurements, all of which are common to scientific standards today. For example, guessing the weight of an object if there was no method to measure weights would make for a poor hypothesis. A scientific hypothesis should also be objective, allowing all factors to be measured or quantified, as opposed to only a few, which can quickly render an experiment inaccurate.