Conversationally, the terms “accuracy” and “precision” are often used interchangeably. Scientifically speaking, however, the terms refer to markedly different conceptual measures. “Accuracy” refers to the proximity of an estimate to the actual value, while “precision” refers to the closeness of individual estimates to each other. For example, throwing darts close to the bullseye would be accurate, while throwing darts very close to each other would be precise even if they’re nowhere near the bullseye!
Companies often make the mistake of utilizing methods that seem powerful by virtue of their precision. At first, declaring that third-quarter sales will hit precisely $496.83M sounds more impressive and compelling than an estimated range of $462M to $465M – that is, until actual sales come in at $463M.
In this case, the precise estimate was precise…ly incorrect! At Spinnaker, we always value accuracy in predictive analytics solutions.
About the author:
Nirav Dagli is the CEO of a predictive analytics and artificial intelligence company based out of Boston, and a former partner at Oliver Wyman. He specializes in financial modelling, algorithm development and predictive analytics with an expertise in banking, insurance, asset management and consumer goods. Other specialties include provision of business engineering support for senior executives in top spanning strategic planning, merger integration and performance improvement. Nirav is also a board member at the Boston Children’s Museum and Better Business Bureau.
Mad Over Data's Blog 