Artificial Intelligence in Marketing Science:
Marketing Mix Modeling and Optimization with Bayesian Networks and BayesiaLab
The first part of this half-day program focuses on fundamental causal questions, such as how to perform causal identification and estimate causal effects from observational data. In the second half of the program, we apply these causal concepts to the field of marketing science. We utilize recent advances in Artificial Intelligence to generate a model of a high-dimensional problem domain, of which we have data available for sales, multiple advertising channels, competitive advertising, and seasonal variables. More specifically, we use BayesiaLab to machine-learn a Bayesian network model from the available historical data. On the basis of the network, plus the causal identification criteria presented in the morning, we can analyze the nonlinear response of sales to the input of all marketing drivers. This allows us to isolate the (mostly nonlinear) direct effects of each marketing driver. We can now proceed to marketing mix optimization and utilize BayesiaLab's genetic optimization algorithm. To conclude the seminar, we present an innovative method for estimating contributions and synergies with counterfactuals, which is conveniently implemented in BayesiaLab.
The Program in Detail
Motivation & Background
“Half the money I spend on advertising is wasted; The trouble is I don’t know which half.”
Over the last century, various versions of this quote have been attributed to John Wanamaker, Henry Ford, and William Procter, among others. Yet 100 years after these marketing pioneers, in this day and age of big data and advanced analytics, the quote still rings true among marketing executives. The ideal composition of advertising and marketing efforts remains the industry's Holy Grail. The current practice remains “more art than science.” The lack of a well-established marketing mix methodology has little to do with the domain itself. Rather, it reflects the fact that marketing is yet another domain that typically has to rely on non-experimental data for decision support.
It's a Causal Question!
The single most important thing we need to recognize about marketing mix modeling is that it is a causal question. This means we are not looking for a prediction of an outcome variable based on the observation of marketing variables. Rather, we are attempting to manipulate the available marketing variables to optimize the outcome. Thus, we must simulate interventions, not observations, and we must switch from observational inference to causal inference. This brings us to the Holy Grail of statistics, i.e. deriving causal inference from observational data. Is this even possible?
Graphical Models for Causal Identification
We introduce the fundamental concepts of probabilistic graphical models and how they can help us perform causal identification, i.e. determine whether or not it is possible to estimate a causal effect from observational data. For this, we require causal assumptions about the domain (from expert knowledge) plus a decision criterion, such as the well-known Adjustment Criterion. While it is straightforward in theory, the complexity of the marketing domain limits the practical application of this criterion. As an alternative, we introduce the Disjunctive Cause Criterion (Shpitser and VanderWeele, 2011) that significantly reduces the number of assumptions required for causal identification and, consequently, confounder selection. In theory, we now have all we need to estimate causal effects. In practice, it is only half the battle.
Artificial Intelligence, Bayesian Networks, and BayesiaLab
To go from causal identification to causal effect estimation, we require an "inference engine." In the simplest case, we could use a regression. However, with dozens of interacting variables, that is no longer feasible. This is where Artificial Intelligence comes into play: we employ the machine-learning algorithms of the BayesiaLab software platform, which can build a high-dimensional statistical model that represents the joint probability distribution of all marketing variables. As a result, we obtain a Bayesian network that represents a multitude of relationships between drivers and the outcome variable. Using BayesiaLab’s visualization tools, we compare the machine-learned graph to our understanding of the domain. Furthermore, we can examine the (mostly nonlinear) response curves of the outcome variable as a function of the marketing drivers. Most importantly, we use BayesiaLab to perform Likelihood Matching on all confounders to establish the causal response of the outcome variable.
Resources and Optimization
With all causal response curves computed, we introduce cost functions for each marketing driver via BayesiaLab’s Function Node. On that basis, we proceed to Target Optimization, which, by means of a genetic algorithm, searches for an optimal combination of all marketing drivers, while being subject to constraints of individual variables and an overall marketing budget constraint. The optimization report shows feasible solutions along with the degree of achievement.
Counterfactuals and Contribution
The final step in developing our marketing mix model is the question of attribution and contribution. While it is easy to understand the meaning of "contribution," calculating a numerical value that represents this concept is not. We introduce counterfactuals as a necessary concept to compute the contributions of individual marketing drivers. This allows us to decompose an observed effect into its its average contributions and into time-interval-specific contributions. As a result, we can answer questions like, "had it not been for the $100,000 we actually spent on TV advertising in August, how much would we have sold without it?"
When we speak of the benefits of synergies, we typically mean that a combined effort produces an effect that is greater than what would be the sum of the effects from separate efforts. This is expressed in the popular quote, “the whole is greater than the sum of its parts.” This is what we attempt to quantify as part of the contribution analysis process. For this purpsose, we compare the joint contributions of two drivers to the sum of their individual contributions, which provides us with synergy amounts for each pair of drivers.
Frequently Asked Questions About This Seminar
Who should attend?
Data scientists, decision scientists, knowledge managers, management scientists, market researchers, marketing scientists, marketing manager, policy analysts, predictive modelers, statisticians, plus students and teachers in related fields.
Who is presenting?
Stefan Conrady has over 15 years of experience in decision analysis, analytics, market research, and product strategy with Fortune 100 companies in North America, Europe, and Asia. Today, in his role as Managing Partner of Bayesia USA and Bayesia Singapore, he is recognized as a thought leader in applying Bayesian networks for research, analytics, and reasoning. In this context, Stefan has recently co-authored a new book, Bayesian Networks & BayesiaLab - A Practical Introduction for Researchers.
What's the format of the event?
We will be presenting plenty of PowerPoint slides (available for download afterward), and also illustrate the introductory examples and the marketing mix model in a live software demo. Even though the event is set up as a lecture, we encourage you to ask questions at any time. Also, based on the interest of the audience, we can explore specific topics in more detail. We typically have between 15 and 25 participants at our seminars, which allows us to keep the presentation fairly informal and, more importantly, interactive. Also, you can actively participate in the discussion or simply sit back and listen to the presentation. If the presentation is not what you expected, you can grab a cookie and leave anytime.
What background is required?
To benefit from this seminar, you should be comfortable with basic statistical concepts and techniques. For instance, you should know what a regression is, or what it means to estimate a parameter. While we won't use statistics per se in this seminar, we'll often compare Bayesian networks to traditional statistical models.
What do I need to prepare or bring?
You do not need to bring a computer to this seminar or prepare yourself in any other way. If you wish, you can download a free copy of Bayesian Networks & BayesiaLab — A Practical Introduction for Researchers. Reading the first three chapters can provide a helpful context, but it is not mandatory. After the seminar, you'll have the opportunity to download a trial version of BayesiaLab and try out the examples from the seminar on your own.
What does it cost?
The seminar is free. All you need to do is register on Eventbrite.