Skip To Content

Highlights from the History of Predictive Analytics

Reading time: 4 minutes
history of predictive analytics

By Earl Sires, Rapid Insight

Today, predictive analytics is a ubiquitous business intelligence tool that has opened to door to data-informed insights across every industry. It’s used to predict outcomes in business, politics, epidemiology, and more. However, it wasn’t always this way. The history of predictive analytics reveals that it rose to prominence mostly in the past two decades, but its foundational principles in math have been in use for centuries.

In this post, we’ll touch on how predictive analytics developed and evolved over time to become the widely-applied business intelligence tool that it is today.

The Roots of Predictive Analytics

Though modern analytics involves computers and machine learning, predictive modeling ultimately boils down to statistics. With that in mind, early predictive modeling likely took place in the form of estimates based on what would today be deemed “descriptive analytics”.

Common Types of Analytics

Descriptive analytics asks the question “What happened?”

Diagnostic analytics asks “Why did it happen?”

Predictive analytics asks “What might happen?”

Prescriptive analytics asks “What can I do to change predicted outcomes?”

To start, let’s take a look back at examples of implementations of the foundational elements of predictive analytics in order to trace its development over time.

Construction projects take significant planning; particularly large-scale projects like the pyramids in Ancient Egypt. A project of that scale incorporated statistics through a census as a means of ensuring the right labor resources were dedicated at the right levels. Planners forecast what they’d need to complete sections of the project on a timeline by analyzing past resource usage.


The same principle carries over to generations-long projects like the cathedrals built over the course of hundreds of years in medieval Europe.  Geometry, in particular, was put to extensive use in designing the complex cathedrals. But architects also made calculations using common items like barley grains, hand widths, and eventually, measuring sticks to estimate the number of stones needed to complete a wall.

While these uses of math don’t strictly fit the definition of predictive analytics, they do illustrate some early uses of the foundational concepts from which modern predictive analytics evolved. These early benchmarks fit into the history of predictive analytics because data collection, duplication of processes, and iteration on past methods form the basis upon which predictive modeling developed.

The First Use of Predictive Analytics?

The historical record is imperfect. We know that not every word or mathematical formula ever used was written down or preserved. As such, it would be impossible to accurately identify the earliest use of predictive modeling. 

The earliest known use of predictive analytics as we understand the term comes in 1689. Shipping voyages were a treacherous proposition in the 17th century. Investors factored in breakage (a certain level of loss of goods from shipwrecks and other mishaps) to the cost of doing business. 

However, Lloyd’s of London utilized a predictive model, of sorts, to estimate the risk of a particular journey to ensure the level of risk was acceptable for the potential payout. Investors signed off on a statement of risk calculated prior to a specific journey. Lloyd’s used that statement to develop insurance policies to protect the shipping companies and investors. 


This means that one of the very first identifiable applications of predictive analytics was for business. It’s no wonder that the presence of analytics in business has developed and grown over time.

Computerized Analytics and Machine Learning

Computerized calculation multiplied the power of analytics. Starting in the 1930s, it became possible to analyze massive datasets very quickly. 

Initial implementations of computerized analytic technology were military in nature. In the 1940s, predictive technology was used to guide the automated targeting systems for anti-aircraft weaponry, and the scientists involved in the Manhattan Project used computer models to build simulations of chain reactions while developing nuclear technology.


Shortly afterward, in the 1950s, this history of predictive analytics changed again as the technology reached into many other industries. Weather forecasters, the transportation and shipping industries, and government research agencies integrated computerized predictive modeling into their routine operations. From there, the technology spread outward into the business community.

Computerized modeling techniques allowed organizations to work with data scientists and mathematicians to build powerful, complicated algorithms to predict specific outcomes. Revenue, customer churn, inventory, and supply chain management all began to integrate predictive models at a high level. And through the use of prescriptive analytic principles, businesses made changes to their strategies to offset predicted negative impacts and foster positive outcomes.

The Future of Analytics

Without question, analytics will evolve at an accelerated pace as computers and machine learning get more and more advanced.

The most significant development in analytics mirrors a similar evolution of technology in general: increased accessibility. In recent years analytics has entered the domain of the “citizen data scientist”. Predictive modeling is now within reach of people who have no formal training in statistics or math. If you understand how to use a computer, you can put modern analytics solutions to work.

Drag and drop interfaces make organizing data intuitive and easy to follow. Automated data mining and one-click predictive model building mean anyone can build a predictive model to improve business decisions. And cloud-based data sharing means that insights are available to every stakeholder in an organization.

Regardless of what your organization does, you can put analytics to use to turn your data into actionable insights. Interested in learning more? Click the button below to schedule a demo with our team!


Notify of
Inline Feedbacks
View all comments