Big Data’s legend has reached mythical proportions. Increasingly, it is described as a powerful solution that could resolve some of the world’s most complex challenges, from terrorist strikes to curing cancer.
It is still early days for Big Data implementation, however. Every industry and category is still working through how to best harness it to deliver on its inherent potential. This is especially true in the advertising world.
“Big Data is changing business strategy, but marketers find it difficult to handle the information,” notes a recent eMarketer report. The real issue stems from not knowing how to both capture and rein in the vast amount of data available. Unfortunately, in the absence of established best practices, myths emerge that exacerbate challenges, increase confusion, and lead marketers to focus attention in the wrong place.
Myth No. 1: The more data, the better.
It is not about the amount of data you are able to collect. Instead, effective marketing requires the ability to recognize patterns within a critical mass of data, then draw actionable insight from that. The only benefit to a larger volume of data is if it serves to improve the ratio of signal to noise.
The industry is struggling with too much information, so amount of data available is not the issue. The execution — using the data to solve problems — remains the challenge. Rather than focus on the number of rows in your dataset, it is important to zero in on the information you most desire, so you can access the meaningful patterns and signals that will help attract and reach a target audience. The rest of what you’ve collected is just noise.
Myth No. 2: Investing entirely in hardware will help.
As marketing depends more on automatic transactions, Big Data is only getting more advanced. An enormous amount of data and the velocity of its accumulation defy traditional techniques to gather information (i.e., surveys, polls). This is why technology is needed to filter and analyze the vast amount of data.
Increasingly, the advertising industry is attempting to address Big Data by putting all its eggs in one basket and depending entirely on hardware. Nearly half are simply increasing their capacity for dealing with hardware, according to a recent KPMG survey of CFOs and CIOs on Big Data strategy. This is due to a misguided attempt to speed up the data analysis process.
However, it is important to remember that marketers need to invest in in-house software technology to help them discern treasures from trash — or in other words, to do a better job of granularly capturing and addressing human behavior rather than simply adopting hardware to hastily speed up the process.
All of this said, a good story was never told by a machine. An effective marketing campaign using Big Data will always need a human to translate the data into a meaningful message to deliver results.
Myth No. 3: Data is most effective when automated.
On the contrary, data reminds us that we are working with people. Our brains are quick to notice patterns, to chunk information, and to treat it as a simplistic object. This seems like an odd feat to herald, but it is one that technology is not advanced enough to handle. When reducing people to what Big Data can measure, emotion — the most salient human attribute — is left out.
It takes people to design the systems that collect and organize data — and people are needed to understand the limitations and biases of the systems. It takes people to ensure that the data focuses on the right questions that can lead to meaningful and actionable insight.
As the saying goes, “it takes one to know one.” Insights driven by people, not machines, are essential to making data actionable.
Myth No. 4: Data is infallible.
Behavior patterns are constantly changing, so Big Data cannot be used to predict the future. From changes in shopping plans during the holidays to the members of a person’s family, data doesn’t remain constant.
To fully understand data, context is needed. Without context, it’s impossible to understand human behavior. Marketers rely on demographic data to target their audience, but this approach is only beneficial and effective to a certain extent. Marketers need to go beyond age, gender and ethnicity to discover contextual clues and distinct characteristics to gain a more rounded, granular understanding of their target audience.
The same action, even by the same person, can mean wildly different things. A person purchasing a children’s toy at a supermarket or drugstore often indicates a child is present — unless it is December, when the holidays wreak havoc on shopping patterns.
The same product purchased online usually is bought by an adult without children being present. And if the toy is purchased in a store outside a consumer’s home area, there is likely to be a parent traveling alone at the register, feeling regrets about being away.
Myth No. 5: Data is most telling when it comes from consumers themselves.
Passive observations are often the best way to gather data. Humans are masters of self-deception. Unlike weather patterns or traffic data, information that people volunteer is always biased in one way or another.
People often distort the truth about all kinds of things — sometimes directionally, as in how much they earn, and sometimes in unpredictable ways, such as feelings about a product they know others like.
This is when human judgment comes into play to decipher the data in order to reach the target audience.