Leigh Caldwell, Founding Partner at Irrational Agency has been quoted alongside a few others in an article by Audiense about personalized and data-driven (or data-informed?) marketing. The questions that prompted the piece were interesting, and deserve a more expansive answer.
Much of the insight world still looks at marketing as a mostly offline activity - even when it guides digital campaigns, the insight is one-way: do a research project and then design your campaign based on the results. But perhaps we could link research with the digital process in a more iterative loop: where the performance of our marketing assets generate new insight, and insights generate new marketing ideas. Ever-improving marketing and ever-deeper insights. Here are some ideas on how that might happen.
We think the real question here is whether brands have the ability to tell compelling stories that are customized to the individual. Simply sending promotional emails or offers based on previous purchase history is not really enough to motivate customers in a noisy digital world. Brands need to tap into consumers' psychology, how they think and feel, and influence it by creating new narratives.
If a brand has sufficient insight in-depth on the motivations, beliefs, and culture of individuals, and the technology to create stories that will move those customers, then absolutely - they should go for individualized marketing. But we suspect that most brands will not have enough insight or the required level of investment in communications platforms and technology. So, for the majority of companies, persona-based marketing will allow them to create four to six key stories that are resonant, motivating, and relevant. A personal model will only ever fit around 70% of the customer base, but that is better than trying and failing with a clumsy personalized approach that doesn't truly connect with anybody.
These sound like the same thing, depending on whether it's being described by a programmer or a creative! All creativity is strongly influenced by data. Sometimes the data is visible, for example, if you are algorithmically designing digital ads. At other times, the data is invisible, because it's filtered through the creative's brain. Every idea that a creative person produces is influenced by their experiences, their knowledge, and their training: all these are forms of data, just not in a conventional way.
We believe the real question should be: should creativity be implemented by a well-understood algorithm and business process, or should it be run by black-box neural networks (which could be a piece of AI technology or the brain of a graphic artist)? The answer depends on the maturity and competitiveness of the marketing discipline in your industry sector. If you have invented a unique product or work in an industry with unsophisticated marketing, a data-driven process is probably fine. If you are in a mature market where brands are really important, simple processes won't cut it. That's when the creative process needs the richness and unpredictability of real or artificial brains.
There is one possible resolution to this dilemma on the horizon. Cognitive technologies and cognitive analytics apply the thinking processes and imagination of the human mind to the scale of big data. These are in their infancy, but they just might gain the capability to achieve the creative brilliance of human beings - at least in the marketing discipline - while calibrating their outputs to the latest and most specific data. Maybe this is the breakthrough we've been waiting for.
It will force insight teams to find new ways of combining passive behavioral data with primary research into customer needs and desires. At the moment, the analytics team in a company is often separate from the market research team: they compete to answer the business questions of their marketing stakeholders. With fewer data points from cookies, data scientists might know what customers are doing, but they will have even less of an idea of the why. To bridge this gap, we will see more brands doing research projects to overlay a new layer of richness and understanding on top of existing passive data.
Passive data has major strengths: it covers all customers with (little or no) sampling bias. Primary market research also has advantages: it goes into more depth and allows you to ask about hypothetical situations such as new product launches, not just the past. Combining the two can give a powerful new way to get real insights into the future.