By Duke Perrucci, Chief Revenue Officer at Cambridge Analytica
In the 1960s, the advertising industry started to rapidly evolve through the use of focus groups. These sessions were typically led by psychologists and were designed to explore a consumer’s emotions, attitudes, and beliefs in relation to products and product categories. Unfortunately, the data played a very small role in the design of creative as the omnipotent creative minds wielded a very powerful position in those days. They built advertising from their own creative genius many times in direct conflict with the research findings. This unique dynamic continued into the 1990s.
Being a brand manager in the consumer packaged goods industry in the late nineties, I can tell you that even the most revered agencies were still allowing creatives to overrule the insight derived from planners. Most folks were more focused on winning awards rather than producing practical advertising built off of consumer insight. People wanted Clio Awards rather than advertising that drove awareness, interest, and ultimately sales.
A Glacial Shift in the Mid-Nineties
With the adoption of marketing mix modeling in the mid-nineties, some industries began to realize the folly of chasing advertising awards at the expense of effectively driving their brand’s growth. Now marketers could quantify the volumetric impact of their advertising. With this transparency, a glacial shift began: the use of market research played an increasingly significant role in informing the creative approach. Qualitative was being used to explore the broad issues that existed around the brand – how it made consumers feel, how it fulfilled their needs, and the role the brand played in their lives.
Data is just the beginning of the journey towards better advertising. Once assembled, the data must be properly studied with the help of a skilled data scientist.
These insights, along with the brand’s strategy, were taken into account when building the advertising approach. Before advertising made it to market, it was tested quantitatively through companies like ASI and Millward Brown to verify its effectiveness with consumers. For the first time, the creative process was held accountable against measurable data to get results. Many companies began implementing thresholds on these advertising tests to ensure that ineffective ads didn’t make it to market.
The Limits of Traditional Market Research
This period of advancement was still hamstrung by the limitations inherent in traditional market research. Both qualitative and quantitative suffered from a fundamental issue – neither could understand consumer behavior at its core — how does the consumer truly make decisions? Some experts would say that decision-making takes place in the brainstem – the part of the brain responsible for survival. While this part of the brain is very effective at keeping us alive, it makes decisions driven by emotion rather than framed by logic. Discussing a consumer’s connection with a brand in focus groups is simply not going to give a researcher a glimpse into that subconscious process.
Not Just Any Data
Enter big data. An overused term that generally means vast amounts of data that require computational methods to understand. When analyzed effectively, large and diverse data sets can reveal unique relationships across thousands of data variables. Big data grew out of the confluence of tech advances in data storage and processing power. Layer in the ubiquity of the Internet and now you have a collection vehicle that allows companies to gather petabytes of data with ease. Many data scientists consider these vast streams of information to be passive in nature — they are collected without the biases that present themselves in active market research. And when these data are both broad and deep, a very rich and insightful window into human decision-making can be formed. But the devil is in the details.
First, broad data. It involves stitching together both internal data streams as well as appending external data sources. Take an automobile manufacturer for instance. The breadth of the data set would need to include a consumer’s geo-demographics, financial data, lifestyle/lifestage, media consumption, spending habits, religious and civic affiliations, and attitudes, among many other data points. With these data, you have a very broad but still very shallow view of the individual consumer.
This is where depth becomes important. The automobile manufacturer could then layer in a deep pool of data regarding their current customers – model purchased, price paid, options, warranties, service visits, customer satisfaction scores, as well as the trove of data now available in most modern cars – speed, deceleration, acceleration, passenger weight, journey length, seat position, temperature, etc. This deep data cache, when combined with broad data sets, yields a holistic view of the current consumer and how they engage with the brand.
However, data is just the beginning of the journey towards better advertising. Once assembled, the data must be properly studied with the help of a skilled data scientist. These data specialists are adept at navigating a complex process that is both art and science. The advent of machine learning has allowed researchers to identify highly complex patterns that would never have been possible through traditional market research. In the example of our automobile manufacturer, the modeling might reveal that the combination of a missing father figure and conservative religious beliefs coupled with education level and driving behavior is a strong predictor for buying a particular model. This kind of unintuitive, deeper insight has already started to radically impact the effectiveness of advertising. You can now study consumer behavior across thousands of data points, using sophisticated models to discern the strongest predictors of behavior – both conscious and unconscious.
Whether traditional agencies will admit it or not, the future of advertising will reflect a quantum shift. No longer will brands be satisfied with the vagaries of a creative-dictated process. Traditional agencies will either evolve and transform their approach or will get left behind.