Abstract
Measuring the causal impact of an advertising campaign on sales is an essential task for advertising companies. Challenges arise when companies run advertising campaigns in multiple stores which are spatially correlated, and when the sales data have a low signal-to-noise ratio which makes the advertising effects hard to detect. This paper proposes a solution to address both of these challenges. A novel Bayesian method is proposed to detect weaker impacts and a multivariate structural time series model is used to capture the spatial correlation between stores through placing a G-Wishart prior on the precision matrix. The new method is to compare two posterior distributions of a latent variable—one obtained by using the observed data from the test stores and the other one obtained by using the data from their counterfactual potential outcomes. The counterfactual potential outcomes are estimated from the data of synthetic controls, each of which is a linear combination of sales figures at many control stores over the causal period. Control stores are selected using a revised Expectation-Maximization variable selection (EMVS) method. A two-stage algorithm is proposed to estimate the parameters of the model. To prevent the prediction intervals from being explosive, a stationarity constraint is imposed on the local linear trend of the model through a recently proposed prior. The benefit of using this prior is discussed in this paper. A detailed simulation study shows the effectiveness of using our proposed method to detect weaker causal impact. The new method is applied to measure the causal effect of an advertising campaign for a consumer product sold at stores of a large national retail chain.
Highlights
Advertising is thought to impact sales in markets
We proposed a novel causal inference method which compares the posterior distributions of the latent trend conditional on two different sets of data: one is the observed data which contain a causal effect; the other one is the data from a synthetic control
We showed that our method can detect a smaller sized causal impact more efficiently compared with the commonly used method even when the model is slightly misspecified
Summary
Advertising is thought to impact sales in markets. (MaxPoint), an online advertising company, is interested in measuring the sales increases. Bayesian Multivariate Time Series Causal Inference associated with running advertising campaigns for products distributed through brickand-mortar retail stores. The dataset provided by Maxpoint was obtained as follows: MaxPoint ran an advertising campaign at 627 test stores across the United States. Control stores were not targeted in the advertising campaign. The company collected weekly sales data from all of these stores for 36 weeks before the campaign began and for the 10 weeks in which the campaign was conducted. The test stores and the control stores were randomly selected from different economic regions across the U.S. Figure 1 shows an example of the locations of stores in the State of Texas.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
