Survey Research Analytics

Marketing researchers are leaving a lot of black gold in the ground these days. Nowadays, advanced survey analytics is often seen as conjoint analysis and nothing else. While conjoint is a very useful technique, it is just one of a broad range of sophisticated statistical methods used in marketing research. 

The first blog I ever published, Why Survey?, touched on survey research analytics, as have Why Segment?, Combining Smart Design with Smart Analytics and Taking Quantitative Marketing Research to a Higher Level.

Why the lack of familiarity with other analytics? I suspect it is because there are fewer people in MR with advanced degrees in behavioral and social science fields that historically have required a strong background in research design and statistical analysis – those dreaded “methods” courses. When I entered marketing research in the ‘80s, at least proportionately, there were more marketing researchers with this sort of academic training than there seem to be today.

Though there are significant misunderstandings about it, many marketing researchers and clients seem to grasp the practical value of conjoint quite easily. Some other methods, particularly those used by psychometricians to measure attitudes, are more abstract perhaps and fall in a grey region between qualitative and quantitative research. My impressions, at any rate, for what they’re worth…

This will be a review to veteran marketing researchers, but here are some of the main types of marketing research analytics:      

• In Consumer Segmentation consumers or customers are either 1) statistically profiled with respect to a key target variable such as purchase frequency or 2) divided into clusters (segments) based on how similar they are to each other in terms of wants and needs and/or behaviors. Segmentation can be used either for targeting purposes or simply to learn more about a product or service category. Data can include consumer surveys, customer records or both, though historically surveys have been the main source in MR.
• Marketing Mix Modeling (aka Market Response Modeling) is quite complex, but in a nutshell, it tries to find out how much bang a client is getting for their marketing buck and identify the optimal marketing mix. It can also be extended to Demand Forecasting, in which future sales or market share are forecasted under various marketing and/or economic scenarios. Consumer survey tracking data are sometimes used along with other data from other sources.
• Pricing Analysis is especially useful for products that are not yet on the market or for which historical data are lacking. Surveys are the data main source in these cases. It also can be conducted using existing price and sales data, and price is typically a variable in marketing mix modeling. Conjoint (see below) is a popular method in pricing research.
• Choice Modeling (aka “conjoint”) is a versatile tool often used in pricing research and new product development. For example, respondents in a consumer survey might be shown a series of choice tasks with product descriptions and asked which product, if any, they would choose. Their responses are statistically analyzed with latent class or Bayesian multinomial logit models and the importance (“utility”) of specific features estimated. What if? simulations can be conducted to see which combinations of features – i.e., which hypothetical new product – would garner the highest preference share. The utilities can also be used as input into segmentation.
• Key Driver Analysis is similar to choice modeling in that one application of it is to learn more about consumer priorities. It serves many kinds of objectives and there are many ways to conduct it but, in essence, we are trying to unravel the “causes” of one or more target variables such as purchase interest in a new product, liking for a TV ad, customer satisfaction or brand equity. Causal analysis is an immensely complex topic, even for scholars, and in key driver analysis we really are picking out the independent (predictor) variables most strongly associated with our target (dependent) variable. Many analytic methods are used, one being structural equation modeling.
• Image and Positioning Research does not receive the buzz it formerly did but remains a crucial part of marketing research. Typically, brand or user image data from consumer surveys are mapped with one of several methods such as correspondence analysis to help us understand how brands or users of brands are perceived by consumers. Mapping is frequently combined with key driver analysis to see which image attributes and image dimensions are the ones clients should focus on most.
• Data Mining and Predictive Analytics are well-known to data scientists but little-known to most marketing researchers until quite recently. In Data Science and Marketing Research and Data Science and Analytics Demystified I briefly touch on their history in MR and how they fit into the bigger MR picture.

Unfortunately, survey research is often analyzed with cross tabulations and little beyond that. The focus is typically on very basic results, e.g., “What percentage of consumers have ever used my brand and does this number differ by age and gender?” Deeper analyses of what might be driving differential usage of brands (or category usage) are usually not conducted, though these sorts of analyses used to be more commonplace in marketing research than they are today. We also have many more ways to adjust for non-response and response styles than before but they appear underutilized.

This is not because multivariate methods don’t work or are impossible to explain to clients. The main reasons are that people coming into marketing research in recent years, as noted, have lacked exposure to and training in advanced survey analytics. Moreover, the focus now is on “tech” of various sorts that has been designed for one purpose and one purpose alone.

The new tech (and data sources) are often useful and have opened up possibilities few dreamed of just a few years ago. The downside is that marketing researchers now often have less understanding of the fundamentals of marketing or research compared to earlier generations of marketing researchers. The fundamental distinction between primary and secondary research, for instance, is becoming blurred with increasing specialization and (I feel) over-focus on information technology.

So, what are the key statistical methods marketing researchers should have at least some familiarity with? Here are some of them:

• Descriptive Statistics
• Data Visualization
• Inferential Statistics
• Nonparametric Statistics
• Generalized Linear Model, Structural Equation Modeling
• Random Forests, AdaBoost, k-Nearest Neighbors, Support Vector Machines, Artificial Neural Networks
• Principal Components Analysis, Factor Analysis, Correspondence Analysis, Multidimensional Scaling
• K-means Clustering, Agglomerative Hierarchical Clustering, Mixture Modeling
• Time-Series Analysis, Longitudinal Modeling

An Analytics Toolbox provides some details about these and other methods I often use. The methods page of my company website lists textbooks in my library and academic journals I subscribe to. 

Caveats: These techniques should be understood as research tools to be deployed as needed, not as sales gimmicks with which to woo clients! Secondly, DIY by research execs with limited formal training in statistics – which is not a seminar offered by a software vender – is very risky.

This is a lot, even for statisticians, but some conceptual grasp of what these methods are for, and when and how they should be used, will go a long way to make us better researchers and stand out from the pack.

I hope this has been helpful!

Arrange a Conversation