In Dan Ariely’s book, “Predictably Irrational”, he shines light on a problem that has stumped marketers and salespeople for generations. Namely, the fact that human beings don’t like to play nice inside their pre-scripted roles as consumers. In fact, no matter how realistic standard assumptions about consumer behavior seem to be, there will always be plenty of outliers who do the irrational thing instead.
Common sense and anecdotal evidence seem to support concrete black-and-white views about “rational” buying behaviors like:
- When demand goes up, but supply stays the same, you can raise the price and customers will pay.
- When the economy improves and there's more money to spend, consumers will spend it on more frivolous things.
- When a consumer is 25 years old, they spend more money on eating out than on groceries, but when they turn 35, the numbers reverse.
These assumptions pass the common sense test. Plus, there's no denying that they're true...
Other times, buying decisions don't fit simple, predictable rules established by economists. After all, humans aren’t 100% consistent. At any point, a variable may arise that affects the consumer's behavior and causes them to do something outside the realm of what's considered “rational.”
How Rational Behavior Goes Awry
The idea of “rational” consumer behavior originated in the mid 20th century with the development of neoclassical economics. A central tenet of this framework was that rational economic factors are focused solely on maximizing their personal utility. While the concept of utility was intended to encompass the entire well-being of an individual, “well-being” is in fact difficult to measure and quantify.
Thus, monetary consumption arose as a common proxy for utility, and in turn, basic models were developed to predict how measurable variables would impact rational behavior. For example, as price decreases, customers will buy more.
However, limitations existed within the framework, including:
- What predictive data could be observed, measured, and recorded.
- What consumers valued, and the challenge of stated vs. revealed preferences.
- The complexity of models that could be reasonably developed by humans.
- The inherent nature of human beings to not all behave similarly.
When a consumer engages in “irrational buying behavior,” or acting as a human, the neat predictability of classic economic theory doesn't hold up. It doesn’t leave room for the unpredictable impact of friends, life experiences and emotions.
“Irrational” Does Not Mean Unpredictable
In the WSJ article Making Sense of Irrational Customer Behavior, John Lucker does a great job laying out the potential for “Big Data” and Analytics in explaining seemingly irrational customer behavior. A clear example of this type of behavior happens when a customer buys two bottles of salad dressing at the grocery store because he has a coupon to save $1 on two bottles, even though buying two costs more, and his pantry has room for only one.
This is where the potential for “Big Data” makes an impact. Each of us generates digital exhaust and electronic bread crumbs on a daily basis. From the websites we browse to the brands we follow on Facebook, data is available to gain insight on all buying decisions.
The entire buyer's journey - from initial research to final purchase - creates data points as sales representatives, technical support reps, and customer service professionals interact with a consumer.
Before machine learning, a human expert was required to formulate and test a hypothesis about what variables would drive this customer to purchase two bottles of salad dressing. This type of data analysis is feasible in a world with few variables and conceivable outcomes, but the analysis breaks down in a multi-dimensional world where relationships are inherently unexpected (i.e. irrational). The sheer number of variables involved in understanding and predicting “irrational” actions are far beyond the scope of what the human brain can analyze.
This is where Machine Learning - essentially allowing machines to sift through all this data and locate patterns no human mind could identify - is enabling marketers to develop the analytics models to exploit Big Data and enhance our understanding of human consumer behavior.
With machine learning, marketers can capture and monitor every signal emitted by a consumer to understand exactly when, where and why a purchase is being made, even if the consumer uses coupons. Machine learning provides the texture and nuance to better predict aspects of consumer behavior, removing the limiting “if x then y” rules required in classical economic models.
It no longer matters that human beings aren't “rational” enough to play 100% by a list of predetermined rules. Machine learning can make seemingly irrational behavior predictable and profitable.
To learn more about how to predict your customer’s behavior across their life cycle, and use machine learning to optimize customer service, download our guide.