Ask a friend to check the price of a drink at the shop. They go, look, come back and tell you the number, and then you decide what to do with it. The friend is the function, your request is the call, and the number they bring back is the return value. In code: def double(x): return x * 2. The line result = double(5) runs the function with x = 5, the function sends back 10, and the call double(5) is replaced by 10, so result holds 10.
Because a call turns into its value, a call can sit anywhere a value can: double(5) + 1 is 11, and double(double(3)) is double(6), which is 12. return also ends the function at once. In a function sign(n) whose body is if n < 0: return 'negative' followed by return 'not negative', a negative input never reaches the second line, because the first return has already left the function.
The look-alike is print. A function show_double(x) whose body is print(x * 2) puts 10 on the screen, which seems like the same result. But the caller receives nothing: in Python, result = show_double(5) stores None, and result + 1 then crashes. Ask who needs the value. If a person needs to see it, print it; if the program needs to keep calculating with it, return it.
Think of it like this. A return value is the receipt handed to you at a shop counter: the work happens behind the counter, but the receipt is the one thing you carry away and can use later, for example to add up your spending. Once it is handed over, that visit is finished. Unlike a shop, a function hands back exactly one value per call, though that value can be a list.
Core idea. Return values let functions communicate results back to their callers. A function without a return value just performs an action.