From a need to a requirement: an example
Imagine a technical service company that identifies a need: to have better control over service requests. As it stands, this is an objective, and different proposals could address it in very different ways. By turning this need into requirements, the company can specify, for example, that each request must have a status and an assigned person, that technicians must be able to record work on their mobile phones even without an internet connection, and that customers must receive a notification whenever the status of their request changes. Information recorded offline must be synchronised as soon as the connection is restored.
It is equally important to define how to verify whether each requirement has been met. For example, if every request must have an assigned person, the system could be required to prevent a new request from being fully registered until someone has been assigned to it. This is a specific behaviour that can be tested before the solution is accepted. If the requirements remain as vague as "managing requests efficiently", three companies could submit three proposals with very different scopes and costs, and the differences might only become apparent after the project has started.