Predicting Future Behaviors

In the late 1970s, fresh off of their success with “The Book Of Lists” and “The People’s Almanac,” writers David Wallechinsky, Amy Wallace, and Irving Wallace wrote and published “The Book Of Predictions” in 1981. The book was inspired by reader interest from their previous books regarding future trends in politics, science, technology, and social events. As a result, they went out to talk to experts, psychics, and the futurists of that time in an attempt to get a handle on what we could expect to take place over the next 50 years.

In most cases, these predictions weren’t even close.

That’s not to say that none of the predictions came true. In some areas of science and research, variations of these predictions eventually came to pass. In other cases, such as those predicting futures filled with world peace, an absence of disease, and an abundance of flying cars, many of the predictions seem almost laughably silly 30 years later.

When predicting the future, the future has the tendency to be non-cooperative. A quick look at technological advances and world events from the past decade offers plenty of evidence that even with the best minds working on it, it’s nearly impossible to accurately identify what’s going to happen next.

I bring this up because we’re hearing a lot about the uses of predictive targeting as part of how we plan future media buys and reach select audiences. Which begs the question: “How accurately can we predict the future behaviors of online consumers?”

The answer is limited by what we want to know about the future. As complex as human beings and society are, most of us exhibit fairly similar behaviors when put into similar circumstances. For example, my behaviors as a consumer before I had children were quite different from those after I had children. At the time my wife and I decided to start a family, we unknowingly set into motion a series of behaviors that would be self-perpetuating even two decades later. As a fairly naïve new dad, I had very little insight into what changes were in store for me. However, any marketer who had been studying the behaviors of new parents for any length of time would have been able to easily identify upcoming milestones in my life that would require access to different products and solutions. As a veteran dad, I now look back and most of these behavior points seem patently obvious.

Starting a family is only one area where predictive targeting could do a pretty good job of accurately identifying the probability for a consumer to behave in a certain way (like selling that little red sports car and buying a minivan instead). Other easy to track behavioral milestones include events such as graduating from college, retiring, career changes, buying a new home, and getting a new pet.

However, these are generally macro behaviors that have an extended range of time for their completion. What about predictive technologies for advertisers looking for people who are thinking of buying a pizza for dinner tonight or thinking that they should get their car washed soon?

In these cases, advertisers need to identify different points of criteria that they can use in order to understand the current motivations of prospects and customers. For example, the recent cold snap in the southern part of the U.S. resulted in an increase of pizza sales and delivery of 10 percent to 50 percent over normal rates due to the fact that few people wanted to leave their homes to visit a restaurant. Also, as anybody in the pizza business can tell you, the Super Bowl is the biggest pizza selling day of the year in the U.S. Other factors such as day of the week or current economic mood may also play a significant role in ongoing pizza sales.

Regarding car washes, for people who live in the Northeastern United States, like I do, during this time of year most people look like they’re driving around a “salt mobile” due to the efforts of road crews to keep snow-covered highways safe and clear. For car wash operators paying attention, having an understanding of how their customers behave in the days following a snowstorm can provide them with great insights into sales patterns and how to best position offers to drive more traffic into car washes.

In his recent book, “Reading Virtual Minds,” author (and behavioral scientist) Joseph Carrabis, explores how computers are starting to be able to identify future behaviors based upon recognizing the overt and covert cues provided by their consumers. Characteristics such as mood, cognitive wiring (such as being a visual, auditory, or kinesthetic learner), and comfort levels with Web sites can often be successfully used to help determine specific behavioral probabilities for those visiting consumers.

As of right now, predictive targeting is not an exact science because it requires many outside variables to be effective. However, marketers who inherently understand the needs and motivations of their customers can already take advantage of ways to address those needs when they are most relevant by simply paying attention to those events that drive behavior and using that insight to help plan out their media targeting.

Related reading