EV/EBITDA is one of many shorthand models used to try and approximate the value of a business. In this specific case it is a highly rough approximation of how many years of EBITDA it would take to cover a current approximate valuation of the company.
It is too simple to be of much use to an individual investor due to the fact that smarter and subsequently richer people have long built much more comprehensive models to analyze businesses so will give you incorrect answers vs them.
It is used widely by boards and bigger investors to justify a variety of exercises such as mergers, takeovers, splits or other largescale capital
destroying enhancing exercises as it provides an approximate benchmark vs other industries/players in a given industry.
A company with a very poor EV/EBITDA vs its peers for example, may well be a target for a takeover or restructure due to it being inefficient compared to its peers (eg if you sack the management and replace them with people/structures like their peers have you should increase this back to the industry standard EV/EBITDA multiple etc).
All of these exercises require a lot of subtext (eg the worse EV/EBITA multiple maybe just be a product of some crazy debt the structure the former owner used before dying in a nasty paragliding accident while trying to air to air board their private jet over Tahiti), so although you will hear it thrown around a lot by people, as on the first point it's not of much use without a lot more context around the companies performance.
In hyper summary: EV/EBITDA is a rough approximation of the state of a company. As the great George Box observed of all statistical models: it is thus wrong, but sometimes useful.