I’m fascinated by this chart that was included in a recent report by Eloqua and BtoB Magazine, Defining the Modern Marketer: From Ideal to Real. Based on a survey of 556 B2B marketing professionals from companies of all sizes and industries, the report reveals current perceptions about what “modern marketing” is today — and should be tomorrow.
At first glance, the tall blue and red peaks in the middle pop out, showing that the largest percentage of marketers believe that modern marketing is an equal blend of science and art. As both a humanist and a technologist — I know, strange mix — I smile because I do believe such a balanced and integrated worldview is the basis of modern marketing.
The blue line represents where marketers felt branding and messaging fell along that science vs. art continuum. The red line represents their assessment of that for campaign creation and deployment.
But then there’s the green line.
The green line represents how those same marketers characterize marketing measurement and reporting, which is heavily right-skewed towards the science end of the spectrum. To quote the report: “Marketers are strongly of the opinion that science rules metrics, with the intuitive marketing qualities playing little part.”
In some ways, that feels intuitively natural. Branding, messaging, and campaigns are predominantly about creation, while metrics and reporting are predominantly about analysis. Those are qualitatively different activities, right?
But let’s probe a bit on how we think they’re different. Metrics use numbers? Creative uses words, images, concepts? Analysis is left-brain, creative is right-brain? Of course, modern marketers use analysis to inform creative — which presumedly is why the red and blue lines peak in the middle. Is the green line just a component of the other two activities, their left half? The science half? The distinctions start to get fuzzy.
Which leads me to the question: what do we mean by “science?”
Science is more than math, less than magic
I’m not proposing a deep philosophical debate, but I do want to examine the expectations and assumptions behind “marketing as a science” because they affect what we do and how we think about it.
Do marketers equate scientific marketing with quantitative measurement? True, science cares deeply about accurate measurement of empirical evidence. But for most scientists, measurement is simply a means to an end — a mechanism for determining if a particular experiment gives evidence for or against a hypothesis. The scientific method, what I would argue is the heart of science, is about a much broader process of discovery. That pursuit of discovery is a spectacularly creative endeavor.
Just counting the numbers isn’t science, it’s bookkeeping.
More concerning, however, is when “science” in marketing is interpreted as formulaically predictable phenomena, like physics equations for customer creation. I’ve already argued why that isn’t so at the beginning of my article on the big data bubble in marketing, so I won’t rehash that here. The short version is: human psychology and behavior — and organizations and markets that entwine many humans together in complex ways — defy being boiled down into a set of deterministic algorithms.
It’s common sense. If you had a machine that could perfectly predict human behavior, why waste your time doing marketing? Just point it at the stock market and become the world’s first trillionaire. Or systematically achieve world peace, if you’re feeling more charitable.
That’s not to say that computational marketing, or algorithmic marketing as some call it, isn’t valuable. It can be. But its scope is usually quite limited to specific scenarios — for instance, retargeting in display advertising, where you’re really just seeking a probabilistic edge in your guess about which ad will work better for a particular prospect on a particular page at a particular moment in time. Hey, a probabilistic edge in display advertising is nothing to sneeze at. It makes a difference in your marketing ROI.
But problems arise if you develop delusions of grandeur believing that such probabilistic improvements imply actual knowledge of customer thinking. The models used in these computations are not reality — at best, they’re highly simplified approximations within a narrow context. If you overreach and venerate them as a magic black box of customer creation, you will be disappointed (and probably out of a job).
4 principles of good marketing science
So if science in marketing is more than math — but less than magic — what exactly is it?
Ironically, the answer to such a question is inherently subjective, as marketing isn’t a well-defined science, much less a “hard” science with objective natural laws. But the metaphor is useful in driving management and culture. So with that caveat, here are what I believe are the 4 principles of good marketing science:
1. Marketing as a science is about objectively using data to support decision making. It’s about understanding how to interpret data — and how to identify common errors or misrepresentations in interpretations. It’s also about critically questioning data, its source, its quality, the context in which it was collected, and alternative interpretations that it supports. It’s not just taking pre-canned metrics as a given. It’s about recognizing what data is missing. It’s about trying as much as possible to minimize biases to make good judgments with imperfect information in an uncertain world.
2. Marketing as a science is about looking for patterns in the market and in customer behaviors — within data, but also through more qualitative sources of information, such as conversations with prospects, peers, and third-parties. This is where data science is applied in the exploration of marketing data. It’s about whittling those patterns into models and hypotheses that can be tested. It’s about figuring out ways to operationalize good responses to those patterns with a relative degree of precision. But it’s also about recognizing the probabilities and limitations of those patterns and being vigilant to changes over time and interaction effects with other events and activities. It’s about managing the complex system dynamics of marketing as best as possible.
3. Marketing as a science is about embracing ideas from other scientific and engineering disciplines: psychology, economics, computer science, neuroscience, biology, industrial engineering, anthropology, sociology, etc. It’s about translating the work of other scientists and engineers into the context of marketing, at least as an inspiration for new models and hypotheses. It’s about keeping an open mind to new theories, while also critically examining their reliability and relevance. It’s about a constant drive for new learning on the frontier of marketing thinking.
But most of all…
4. Marketing as a science is really about running good controlled experiments to test hypotheses. This is the heart of the scientific method applied to any discipline. It’s about designing good experiments. It’s about choosing meaningful hypotheses to test. It’s about minimizing confounding variables in such tests — while acknowledging that real-world marketing experiments are inevitably subject to uncontrolled influences. It’s about deciding how much formality should be applied to a marketing experiment. It’s about intellectual honesty in the analysis of outcomes, knowing the difference between proving something versus merely providing corroborating evidence. It’s about iterating experiments in a smart way to methodically make progress in the pursuit of a theory. It’s about knowing when to stop pursuing a theory. It’s about carefully considering how the outcome of an experiment can be generalized and expanded to other contexts — or not. It’s about assessing when to reuse the learning from a past experiment and when to revalidate it. It’s about advocating that the answer to many marketing management questions is: test it.
The common theme among all these principles? Science in marketing is the systematic quest for knowledge. Not absolute truth, but effective working knowledge that can be applied to growing a business.
What does marketing as a science mean to you?
Get chiefmartec.com directly in your inbox!
Subscribe to my newsletter to get the latest insights on martech as soon as they hit the wire. I usually publish an article every week or two — aiming for quality over quantity.
9 thoughts on “What do you mean by marketing as a science?”
Thanks for this Scott. I wrote something similar a while back (http://www.digitaltonto.com/2011/marketing-numbers-math-and-science/), but came out on the other side.
In my view, marketing is not a science, but a business function. It can informed by science and be improved by implementing elements of the scientific method, but its purpose is not discovery, but profit.
That’s an important distinction. We don’t search for truth, but what’s true enough to perform a little better than the competition. Further, it has serious consequences, like the ridiculous pursuit of tortured ROI calculations for non-material costs.
So, while it’s important to be able to think scientifically, treating marketing as a science is often counterproductive.
Greg — I love your post: Marketing, Numbers, Math and Science.
I think we share a lot of the same worldview.
Marketing is definitely not a hard science like chemistry or physics. It’s also usually not a soft science the way economics or psychology are — although if you consider academic research in marketing, it comes closer. Such academic research is certainly at the crossroads of several soft science branches of study. (One of my favorite “marketing” books — Predictably Irrational — was written by a behavioral economist who teaches marketing classes at leading business schools.)
But in business, most practicing marketers are not pursuing research in that context. As you wrote so eloquently in your post: “It is not truth we seek, but competitive advantage.”
If we forget about labels for a moment and focus pragmatically on what works, the wise use of scientific thinking in marketing strategy and execution is a competitive advantage. There have now been many studies showing that data-driven organizations outperform their non-data-driven peers — a recent SAS-sponsored HBR report on the evolution of decision-making comes to mind, as well as a “datavores” report in the UK from last November.
So while it’s not technically “science,” it is useful.
I wrestled a bit with using the “science” label — that’s why I called it a “metaphor” at one point. In the end, I think I felt that if such a label helps marketers think differently and embrace these techniques, that are new to many, then a little imprecision in terminology is a small price to pay. I’m sure real scientists could object. But they’re not really my audience. 😉
P.S. What about “data science”? I find that label is used in a very broad set of instances, many of which in industry don’t seem to qualify by the strict label of science either. But data science seems to have been embraced by a number of academic circles.
Data science is a lame, abstract term. It’s another buzz word that you cannot argue for or against, as everybody has their own idealized personal definition. From a pragmatic standpoint, its hard to have rational discussions when the strawman on the table has the ability to predict anything from any data.
Plus, the general definition of a data scientist is ridiculous; it’s akin to a superhero (expert in math, statistics, databases, computer science, and a domain expert — in likely multiple domains). If you expect to hire such a “data scientist,” good luck. That’s like everybody expecting to hire a 10x software engineer. These are rare skillsets in one person. Not impossible. But rare. And incredibly unlikely.
The other, bigger problem I have is this: data is dumb. Without syntax and semantics — i.e., a proper context — it is completely useless. Once a context is properly applied and gives that data meaning, then it becomes information.
Now, information is something I can sink my teeth into. Information is relevant. Information can be smartly applied. Information can be framed for intelligent analysis and decision-making.
Scientific greats, like Einstein, didn’t just study the data. They made meaning for it. They created context for analysis. For example, the theory of general relativity provides context — and meaning — for physics data. Without that theory, that meaning, the data are useless. (BTW, it’s really hard for me to write “data are”…so I’m not data scientist.)
Instead of “Data Science,” I’d rather focus on “Information Science” – which should be focused on the creation and use information theories, or frameworks, which data can be applied to. In this way, we can provide meaningful context for the more efficient and effective data analysis.
When it comes to marketing, we need more marketing information. Not mroe marketing data
The term data science is quite controversial and, to be honest, I have no need to enter the fray. In the end, I think how it is practiced is what determines whether it is science or not.
The reason I find the term science objectionable is because it so often obscures more than it reveals. All too often, it’s used to value quantification over more qualitative approaches, masked under the authority of science.
Yet, science is always uncertain. Propositions are falsifiable, not provable. There is no such thing as “scientific proof,” only propositions that are more or less supported by empirical data. They are always subject to revision as new data comes in (btw: I’m working on a new framework that I call “Bayesian Strategy” that specifically addresses this point.
If the term “marketing science” was used in that way, I wouldn’t have a problem with it. However, it never is. It’s used to convey a false sense of certainty that never really exists.
Pingback: Where’s the science in scientific marketing? | B2BG2
The thinking of Peter Drucker informs my perspective on marketing. The purpose of business is to create a customer. As such, businesses have two essential functions: marketing and innovation. These are inextricably linked, as the purpose of marketing is to know the customer so well that a product sells itself.
That, to me, is the ideal function of marketing. The tools we employ, be they art or science, should serve to achieve that ideal.
Art and science have long been intertwined, and marketing is yet one more area where this i true. Marketing shares the two, like parents, albeit with science being the younger in this family.
Creative will always lead, as you note, branding and copywriting isn a creative endevor. Data mining and analysis merely let the artsy-fartsy types confirm that which they intuitively knew to be true, or shoot down that which they hoped to be, but wasn’t
It’s fortunate, really that the two can coexist in such wonderful synergism, and modern data analysis software can deliver powerful insight as guidence. In the end, it will always be both; art and science, that will prevail when most effectively marketing.
I think the semantic issue is definitely not what’s important here, though I would like to note that something does not have to be true (or entirely true) to be helpful, instructive or beneficial. What is important is the idea that marketing can often be approached in a scientific way and that methodical and rigorous approach can deliver incredibly valuable results. Thinking like a scientist by establishing a hypothesis and then testing it is an important concept for marketers, especially in the digital realm where testing is relatively easy, inexpensive and fast.
It’s important to know the data but a wise marketer isn’t lead by the numbers. Rather, as you’ve said, uses data to make informed decisions. Marketing research has proven that past behaviors don’t always predict what consumers are going to do next. But from a tactical perspective, we can gain insights into cause and effect. I agree with Steve that there is great power in the the ability to easily test marketing directions in the digital realm, the results of which can allow us to tailor campaigns real time, adapting to actual market response.