Sensitivity Analysis

Categories: Metrics, Trading

You gently smack little Jane with a pillow and she laughs, coming back at you with a pillow-thumping of her own. The game has begun. You smack little Joseph with a pillow to get him in on the game...but he’s less stoked about feathery pillowy violence. More sensitive, one could say.

What factors could have led to these different sensitivities? Maybe Jane went to a summer camp in the Alaskan wilderness, while Joseph went to a coding summer camp. This is the beginning of a new game: sensitivity analysis.

Sensitivity analysis, in its broadest form, just means you’re trying to figure out how certain independent variables (like type of summer camp) affect a certain dependent variable (child sensitivity to pillow fights). There aren't always going to be correlations...and where there are correlations, that doesn’t mean there’s causation, i.e how sensitive the overall equation is to a change in a given input. And mathematically, that runs like x + y + 37z = whatever. The equiation is extremely sensitive to changes in whatever the z variable represents.

Still, sensitivity analysis is useful for just about everyone. Businesses can use it to figure out which factors give them the most bang for their buck. For instance, they may find that, the faster they ship their goods, the more business they get...say, a 2% increase in sales with every day cut from shipping time. Dig even further, and that business may find that sales also correlates with customer service with even greater sensitivity than the faster shipping...say, a 10% increase in sales for every salesperson who completes the new salesperson training program. Booyah.

Naturally, traders use sensitivity analysis (often called technical analysis) to figure out which factors affect the prices of securities. Economists use sensitivity analysis via econometrics, figuring out which factors affect the dependent variable, and by how much, depending on the coefficients.

Sensitivity analysis is ballooning along with the insane amount of data big tech companies are collecting, with no signs of slowing.



Find other enlightening terms in Shmoop Finance Genius Bar(f)