Approaches to Predicting Human Behavior with Prescriptive Analytics

By Greg Silverman, Aug 24, 2020

Is it possible to put a percentage on the ability to accurately predict human behavior? Well, Northeastern University attempted to do so a decade ago when the school released a report explaining how leading network scientists found human behavior to be exactly 93% predictable. The report further emphasized the lack of spontaneity humans have, preferring to stick to their tried and true routines rather than risk trying something new and not liking it.

However, if consumers are so unsurprising, why aren’t more businesses predicting human behavior to help reach their sales targets and so on? Most organizations still don’t grasp that consumers are mostly predictable, and their forecasting models only use the “rational man” notion that individuals will always act in their best interest, especially when faced with an unexpected situation.

In reality, consumer behavior is able to be forecasted to a high degree of accuracy - especially with the advancements in technology and artificial intelligence created in the past decade since Northeastern University released its initial findings. By understanding how and why consumers make decisions, businesses are able to unlock dynamic human behavior forecasting to close the gap on the unpredictability they perceive in their current models.

How do consumers make decisions?

Individuals may fret over certain decisions, but if the outcomes of their actions are favorable, they immediately believe they have made the best choice. This is especially true for consumer purchasing habits. Consider this: A consumer is thirsty; they purchase a soda instead of water or juice to quench that thirst and feel satisfied. In their mind, they made the best decision - even if the alternatives may have been better options for their needs.

This is a very basic example but shows that when consumers make a decision that satisfies their needs, they are unlikely to consider alternatives. Kahneman and Tversky, founders of the Prospect Theory, explain this phenomenon in more detail. Essentially, consumers are always thinking in terms of expected utility of a purchase relative to a reference point, rather than what the absolute outcome may be. In the consumer’s mind, choices are framed by risk, and they dislike loss more than they would enjoy an equivalent gain. Ultimately, consumers are taking risks to avoid a loss more than they would to retrieve a gain.

This is chalked up to how humans perceive risk as well - they overestimate the probability of a worst-case scenario occurring. Consider the insurance industry. Millions of consumers purchase insurance voluntarily to protect themselves from a perceived personal risk (like a car accident, flood, etc.) without truly understanding the probability that they will need it. In times of uncertainty, the same rules apply. Think about the mad dash to a grocery store when a hurricane or snowstorm is forecasted or the shelves void of cleaning supplies only a few months ago. The perceived risk of these situations ultimately drives consumer spending habits.

Despite people who believe that humans are spontaneous, this shows that Northeastern University’s findings were indeed correct. Now, a decade later, with advanced technology and a deeper understanding of human behavior, it’s even easier to accurately predict what a consumer will do

A deeper understanding of human behavior

There is a formula for success

While there’s no perfect algorithm for businesses to use to understand what consumers will do, it is possible to create mathematical models to predict their behavior with accuracy. This is accomplished with advanced machine learning and a training set of data that is used in addition to insights from economic behavior (like the Prospect Theory) to produce advanced predictions of human action.

Businesses that rely on the old adage of “past behavior predicts future behavior” aren’t incorrect, but it’s not enough to always rely on historical data science methods. Organizations must work to deconstruct past behaviors to understand what drove a consumer to purchase and how the business’s actions affected the outcome. With this information, businesses are able to create agile models that incorporate these drivers and are scalable enough to change with consumer purchasing habits.

Even as consumer behavior changes due to evolving perceptions, market trends, and environmental circumstances, the rules of losses and gain remain relatively static. Businesses need to consider how consumers change the weights of their perceived risk, but they are confident that the consumer will act to avoid a loss. With this knowledge in mind, a prediction about behavior is made with greater confidence.

Predicting human behavior with prescriptive analytics

Market simulation with prescriptive analytics at its core recreates how a market behaves in a simulated environment, including how consumers react to business decisions. As the predictive model becomes more scalable and repeatable, it is continuously tested against environmental factors and actual market outcomes to increase its accuracy. With this forecasting technology and insight, it’s actually possible to predict human behavior with 95%+ accuracy.

Prescriptive analytics has the power to account for human behavior - no matter how predictable or irrational it may seem from the outside. In a simulated environment, businesses are able to see how a social group interacts to influence consumer behavior. Additionally, competitors, alternatives in the marketplace, and other macroeconomic conditions are also included in the simulation to create a complete picture of what drives a consumer to make a purchase.

By asking different what-if questions and running the realistic simulation, business decision-makers receive information about consumer touchpoints and behavior. At this stage, they are able to identify which points in the customer journey they have power over to meet the customer and influence their behavior - ultimately encouraging them to purchase their business’s product over an alternative.

The Concentric platform creates a flexible and scalable process for businesses to predict human behavior. A decade ago, Northeastern University found human behavior to be 93% predictable, and today we have the technology to forecast it with 95% accuracy weekly. These forecasts help organizations across industries make better, faster strategic decisions to efficiently meet their goals. Contact us today to learn more about adopting our platform into your business model.

 

How Prescriptive Analytics Provides a Roadmap to Your Revenue Target

 

A Story About How Brands Deliver Value, and How Technology Augments Their People to Be at Their Best
AI vs. AI: Everything You Need to Know About Augmented Intelligence
6 Key Benefits of Utilizing Prescriptive Analytics