Analytics news and expert advice every Thursday.
4 Principles Of Marketing As A Science
What is the future of marketing? You can almost hear “Science!” as intoned by a popular 80’s song by Thomas Dolby.
Across our profession, more and more people are talking about marketing science, scientific marketing, marketing as a science (in contrast to an “art”), and so on. In our digitized and data-deluged world of modern marketing, these phrases resonate. “Yes, marketing is more of a science today.” That just feels like a true statement, doesn’t it?
But what do we really mean by it?
What makes an approach to marketing scientific? In what ways does it differ from marketing as an art? And how far should we take this? Is science driving out the art of marketing completely? Or are they complementary worldviews?
After reading many articles on this subject and talking with numerous marketing practitioners, I’ve come to believe that “marketing as a science” can be distilled into four principles — with caveats.
1. Data-Driven Decision Making
Data is at the core of the marketing-as-a-science movement.
Colloquially, the “art” of marketing management in the past can be characterized primarily as decisions that were made from the gut (intuition) based on experience.
In contrast, marketing as a science favors data-driven decision making. When facing a marketing choice — Should we buy top-of-the-funnel keywords? Should we offer a discount in our ads? When is retargeting effective and when is it annoying? — the more scientific approach is to seek data to help answer these questions.
Because the digital environment gives us access to a prodigious amount of data, and because there is a plethora of marketing technologies that can help us analyze and leverage such data, this approach is increasingly practical across a wide range of marketing decisions.
It’s a huge step forward in marketing management and culture.
However, we have to be careful not to overreach, reading more into the data than is actually there. A Harvard Business Review blog post on The Hidden Biases in Big Data cautions managers to resist “data fundamentalism,” the belief that data has all the answers and that techniques such as predictive analytics always reflect the objective truth.
What you’re doing with data is also important, whether it’s mining data for forward-looking exploration or reviewing data for confirmation of past performance:
The real essence of data-driven decision making isn’t merely using data, though; it’s striving to use data objectively. And, that’s harder than you might think, thanks to a psychological quirk known as confirmation bias.
Therefore, ironically, this is one of the places where “art” still remains in marketing management. Are we using the right data? What other factors should we consider that aren’t in the data? How much weight should we give to the insight from certain data? Are we asking the right questions in the first place?
The more experience you have with data-driven decision making, the better you become at answering those questions.
2. Empirical Pattern Recognition & Model Building
The second principle of marketing as a science builds upon data-driven decision making, but aims to use data to organize customers and marketing activities in a much more quantitative — and automated — fashion than ever before.
Marketing science, to quote another 80’s song, has “orders to identify, to clarify and classify” customers, prospects, and influencers. This goes beyond traditional demographics and a handful of bulky customer segmentations. Instead, we drill down deeply into more specific personas and micro-segments with a granularity that has only recently become technically feasible.
Marketers as “scientists” are principal investigators into the phenomena of the marketplace and all the varied species of customers within it. This really emphasizes the exploratory side of data. Marketers are increasingly building models — mental models, data models, software models — the way that scientists would, seeking patterns and empirically validating them.
In addition to being able to target ever narrower customer segments with greater precision — asymptotically approaching a “segment of one” — these models, when embodied in data and software, enable things like marketing automation. Either explicitly or implicitly, we can use rules and heuristics to personalize messaging and experiences for individual customers.
The caveat, however, is that we must remember that the model is not reality. At best, it’s an approximation that can be very helpful to us and our target audiences. But, to paraphrase a famous survival guide, when the model and reality disagree, always listen to reality. Keep a vigilant watch for data that requires us to modify (or outright discard) a model and create a better one.
3. Controlled Experimentation: Hypothesize, Test, Refine
Of course, the real workhorse of a scientific approach to marketing is running good, controlled experiments:
- Create a hypothesis.
- Test the hypothesis.
- Accept or reject (or refine) the hypothesis.
You can certainly run tests without a hypothesis. (For example, Google’s infamous 41 shades of blue experiment.) But having meaningful hypotheses generally helps to direct experiments toward useful business goals and helps build validated learning.
Great hypotheses can come from anywhere, but increasingly, they’re emerging from the analysis of big data. It’s the smart use of hypotheses that helps drive data-driven decision making and validates the kinds of models discussed above.
Marketing as a science also recognizes that the number of opportunities for controlled experiments is large, spanning almost every facet of modern marketing. Therefore, businesses that value a scientific approach to marketing will encourage broader experimentation throughout their organization à la massively parallel marketing.
4. Cross-Pollination Of Ideas From Other Scientific Disciplines
The fourth principal of marketing as a science is simply embracing ideas from other scientific disciplines that are relevant to marketing, such as psychology, biology, sociology, neuroscience, economics, computer science and so on.
Scientific marketers are usually interested in science more broadly. They’re curious and open to new learning. As they read about new findings, theories, and frameworks from other disciplines, they consider how they may be applicable to their own work and are eager to try such cross-pollination — which is another excellent source of hypotheses for marketing experimentation.
What does marketing as a science mean to you?
Some opinions expressed in this article may be those of a guest author and not necessarily Search Engine Land. Staff authors are listed here.