Harvard Business Review: Advertising's Big Data Dilemma

By Christian Madsbjerg and Mikkel B. Rasmussen, August 7, 2013

The proposed merger of ad industry giants Omnicom and Publicis, forming the world’s largest advertising firm, promises to change the face of Madison Avenue forever.

So long, Don Draper. Hello, Hal 9000.

It’s understandable why two giants of traditional advertising would pursue such a consolidation — to better do battle against Google and Facebook, relative newcomers to the ad game but already upending the entire business dynamic by using Big Data analytics to create highly targeted ads.

But an algorithm can never truly master the art of persuasion.

Great marketing and advertising campaigns are exquisitely attuned, not to past behavior or to individualized needs and desires, but to the larger cultural zeitgeist. Great advertising speaks to our deepest fears and desires; it answers to our nascent yearnings.

Traditionally, the heart of any successful advertising agency has always been its creative department. During the heyday of Madison Avenue, in the late 1950s and into the 1960s, a wave of young art directors and writers used wit, energy, and style to usher in the era referred to as the Creative Revolution.

Consider one of the classic examples from the period: in 1955, a minor brand of cigarette aimed at women smokers tapped the Leo Burnett agency to revamp its brand. Burnett and his team might well have turned to the data available at the time on female smokers. Instead, they recognized an opportunity to tell a new story, one that tapped into the cultural zeitgeist, a sense of confusion and loss about American masculinity. The advertising campaign featured a rugged cowboy on horseback, an uncompromised man struggling not with a demoralizing bureaucracy but with the forces of the natural world. He smoked a cigarette “designed for men that women like.” The Marlboro Man was born.

This is the art of persuasion. Great marketing and advertising campaigns are exquisitely attuned, not to past behavior, nor to individualized needs and desires, but to the larger cultural zeitgeist. Great advertising speaks to our deepest fears and desires, it answers to our nascent yearnings. Perhaps most importantly, it acknowledges that the majority of our decisions are social: we do things within the context of our communities and we get swept away by the mood of our times. From Volkswagen’s “Think Small” print ads to Apple’s groundbreaking “1984” television commercial — directed by Oscar-award winner Ridley Scott — to Nike’s “Just Do It” slogan, persuasive advertising campaigns have left an indelible mark on our public imagination.

Big Data analytics simply can’t address us with this kind of depth, the full context of our lived reality. Take a recent example with Boston potholes. An app called Street Bump was designed to collect smartphone data from the city’s drivers. The idea was to collect information about pothole repair at a low cost. Unfortunately, the app had difficulty distinguishing between bumps in the road, manholes, and potholes, and, as a result, the Office of New Urban Mechanics received an overwhelming amount of false positives. Even more problematic, by relying only on feedback from the app, Boston was not receiving any information from neighborhoods where the residents didn’t own smartphones. This skewed the objectivity of the data received.

This kind of “skewing” is always happening with data. Despite what we are told, data is never objective, never free of bias. Kate Crawford, researcher at Microsoft, calls this problem “Big Data fundamentalism — the idea that with larger data sets, we get closer to objective truth.” Aggregator sites, seemingly objective, are designed with built-in assumptions. They assume that the frequency of your clicks is the same as your level of interest or the degree to which the material “moves” you. But only our fellow humans will ever really understand what we care about. Our care — our deeply felt investment in the world — is always context dependent.

What is the role of Big Data in the future of advertising? Data analytics plays a part in informing a successful marketing strategy. According to the chief executive of one of the industry’s major data marketing companies, advertisers can determine, in milliseconds, whether someone looking for a car is a “luxury” or “used car” buyer, and based on that information, they can determine whether to even display an ad or not. If your problem frame is simple and straightforward — “I want to reach a consumer who buys luxury cars in order to entice them to buy my luxury car — these types of targeted ads could do very well for you.

But what if your problem frame is not straightforward? What if, like Marlboro or Nike, you are trying to elicit a new emotional response from your consumers? What if, like Volkswagen, you feel that your product could thrive in an entirely different kind of market? What if, like Apple, you are trying to lead, not follow the decisions of your consumers?

To address a more complex problem frame, you need a more complex piece of technology. In these situations, an algorithmic business model based on Big Data analytics — if this, then that — is not going to provide you with the greater insight or perspective. It certainly isn’t going to create a strategy or a campaign. For any of the above, you are going to need the human mind.

Most of people’s decision-making is based on social, not rational, factors. Thus, in order to truly understand the complex reality of people’s social lives and what motivates them to buy something, a human mind is needed.

This article originally appeared in the Harvard Business Review.

 

Related content