Systems behave differently based on its environment. With differing inputs, the system can have a different set of outputs. Sensitivity analysis is a way by which we can analyze the effect of individual inputs on a system. Sensitivity analysis helps us understand which inputs have more effect on the system than others. This concept is used in system identification, control theory and even in marketing. Ads are usually A/B tested in a smaller group to understand what features of an ad appeals better to users before it is rolled out.
A heuristic way of applying this is, if all else are equal a car that travels a longer distance will burn more fuel than a car that travels a shorter distance. This gives us a relation between distance and fuel burnt. However, there are other factors to consider in real life like if the car has the different top speeds, drag, load/mass etc. But what we are trying to do effectively is to abstract all that out of the relation to understand the relation between fuel burnt and distance travelled.