Anyone Can Do Marketing Research

I sometime hear people say that Marketing Research is easy now with all the new technology we have, and that anyone can do it. On the other hand, we could make the case that it has become more sophisticated and increasingly requires highly specialized skills.

From my vantage point, marketing research has been made both easier and more challenging by technology and other developments, not the least of which is a transformation of marketing itself. Very simple, repetitive surveys are often automated, such as the irritating and methodologically suspect customer satisfaction surveys we receive every time we interact with our bank. Some basic customer queries can be handled by chatbots, though the claim that they can, or already have, replaced human focus group moderators is ridiculous.  

Certain personality traits are essential to be a competent researcher in any field. They are generic and largely unaffected by technology. For example, a general curiosity and eagerness to learn are indispensable. Critical thinking skills and imagination are crucial. Good researchers often seem to have a relentless, almost obsessive quality about them – the answers they’re seeking typically don’t come easily – as well as considerable tolerance for ambiguity. At the same time, some amount of detail-orientation is required. Many personality traits that research demands, therefore, conflict with one other, and “normal” people, quite understandably, often see us as oddballs.

Personal and professional ethics are too big and important a topic to fit into this short space. Though we don’t talk about it very often, research and sales to some degree inherently conflict. I personally cannot sell something I don’t believe will work or don’t feel is needed for the particular business issues the client is facing. I also have difficulty giving a quick yes or no response because my answers frequently lie in a grey area in between. The “right” answer is not always a good answer, as I painfully learned early in my career. My formal education in statistics has encouraged me to think in terms of conditional probabilities, which makes it hard for me to make bold, unqualified statements. I am lousy at sales.

In A Dozen Essential MR Skills and 15 Essential Marketing Research Skills I’ve listed some skills and knowledge domain competencies I believe most marketing researchers should have. Here is a selection of them (admittedly, quantitatively-biased):

  • Research design, including experimental, non-experimental and quasi-experimental designs
  • Fundamentals of Sampling
  • Questionnaire design
  • Fundamentals of Psychometrics
  • Basics of data processing, tabulation and data management
  • Fundamentals of qualitative methods
  • Descriptive statistics (e.g., mean, median, standard deviation)
  • Fundamentals of Inferential statistics (e.g., T-test, chi square, F-test)
  • Basic familiarity with multivariate methods such as factor, cluster and regression analysis, as well as newer methods often called machine learning (e.g., Random Forests and Support Vector Machines).
  • Analytical thinking, as opposed to number-crunching
  • Knowing how to find and use secondary data sources, including “Big Data”

Of course, business understanding, knowledge of marketing, interpersonal skills, communication skills, management experience and many other competencies are needed, and qualitative specialists and people working on the data management side of data science would greatly expand on my list. Some background in the social and behavioral sciences beyond that covered in marketing courses and seminars is also very helpful in marketing research. 

Research also entails a way of thinking, and is not just knowing about math and methods. Data, Analytics and Decisions, What Makes a Good Analyst? and Research Thinking explain what I mean by this.

This is a lengthy list, and I’ll try to simplify my thoughts about this subject. First, domain knowledge is critical, for example understanding marketing and specific business issues important to your client or employer. Numbers or customer verbatims need context to be interpretable and are meaningless in a vacuum. Descriptive information is not insights, nor is it knowledge, and neither can be machine-generated. Abstract conceptual thinking and the human gut is required, and AI is still years away from achieving this level, if it ever does. As I’ve said elsewhere, pattern recognition is not cognition. There is no way around it – some very important things cannot be automated, at least in the foreseeable future. 

D

Secondly, an expanding array of technical skills is required. In my case, these include knowledge of statistics and machine learning, software written to perform these operations as well as other skills on my list above. Again, though, if I lack the necessary context and understanding of the important business issues, there is only so much I (or a machine) can do. Competing statistical models often provide roughly the same “fit” to the data but suggest very different courses of action to decision makers. Computer scientists use the expression “There is no free lunch” to emphasize that there is no single solution that always works best, and the same holds for statistics.

Moreover, in the past couple of decades there has been an explosion in the number of tools for data analysis, and automating analytics in many respects has become more difficult since there so many more options. See Analytics Revolution and An Analytics Toolbox for examples of methods frequently used by statisticians and data scientists.

Then there is the need to be able to design research, which also has become more complex. Statisticians do much more than crunch numbers using data they have been given. I touch on this in Preaching About Primary Research. To reiterate, irrespective of their specializations, most marketing researchers must have wide-ranging technical knowledge and software skills. I’m amazed at the software wizardry of ethnographers these days! Still, automation lags the ever-expanding needs of researchers.  

Lastly, there are communication and interpersonal skills. At all phases of a project – from research design through presentation of findings – a marketing researcher needs to be able to communicate and get along with people from diverse backgrounds, who have different skills sets and different occupational roles. Some of them will see us a threat, not a partner, and we seldom have the same reporting lines and may work for different organizations.

On the other hand, no one person can do it all. In Putting It All Together I propose a structure we can use to establish our priorities and efficiently disseminate the knowledge and skills most needed.  

In short, many tasks have become easier and can be automated, or require less-skilled human researchers. At the other extreme, however, marketing research is harder than ever. Marketing has become highly multifaceted and we now have an enormous collection of research tools to master, many of which are not simple to grasp or use appropriately. Modern Bayesian statistics and Deep Learning are just two examples of complex, rapidly-evolving technologies that are now part of the MR toolkit.

In An Elephant In Our Room I suggest what some may consider heresy – that it is actually becoming harder to find people with core marketing research competencies.

Thinking has not yet become obsolete.

Browse

Article by channel:

Read more articles tagged: Featured, Marketing Analytics

Data & Analytics