In the foresight world, experts often distinguish between forecasting and prediction.
The conventional views on this topic are well explained by Eric Siegal, author of Predictive Analytics. In the following paragraph, he discusses the difference between predictions, as they relate to prediction analytics (or PA), and forecasts:
Predictions drive how organizations treat and serve an individual, across the operations that define a functional society. In this way, PA is completely different from forecasting. Forecasting makes aggregate predictions on a macroscopic level…Whereas forecasting estimates the total number of ice cream cones to be purchased next month in Nebraska, predictive technology tells you which individual Nebraskans are most likely to be seen with a cone in hand.
That’s all generally true, but I think we should beware notion that these are completely different pursuits. For one thing, the words forecast and predict (or forecast and prediction) are synonyms in the public’s mind. Yes, experts can constantly strive to “correct” the general view that these are virtually the same thing, but that will always be an uphill battle.
More importantly, though, when we delve into the details of forecasting and PA, a lot of the distinctions start to blur. Let’s begin with one of the key methodologies used in PA: regression analysis.
“Linear models and linear regression techniques are the most fundamental methods available to the analyst for predictive modeling,” write Michele Chambers and Thomas W Dinsmore in Advanced Analytic Methodologies. Yet, regression analysis can also be used for forecasting. In his book Introduction to Financial Forecasting in Investment Analysis, John B. Guerard Jr. writes:
A forecast is merely a prediction about the future values of data. However, most
extrapolative model forecasts assume that the past is a proxy for the future…Regression analysis is a statistical technique to analyze quantitative data to estimate model parameters and make forecasts.
So, if the statistical methodologies are often similar, is the distinction just about macro versus micro viewpoints? Well, not really. PA is often used to predict how individuals with certain characteristics will react to, for example, a specific advertisement. But that prediction is not usually intended to be accurate at the individual level. Rather, in order to pay off, PA only needs a specific group of such individuals to be a little more likely to do something such as vote for a politician, buy a product, exercise at a gym and so on. In other words, like forecasting, PA tends to focus on the macro, even if its techniques for getting there are based on breakdowns of groups of people.
In short, it’s a stretch to say that forecasting and prediction have nothing in common. The truth is, they are synonyms in the popular mind, use some of the same statistical methods, and can’t be so easily segregated into “group” versus “individual.” Far from being strangers, they are more like siblings or first cousins who, though they look and act differently, share a lot of DNA as well as a family resemblance.