Today’s marketing environment is characterized by a surge in multichannel shopping and ever more choice in advertising channels. This requires firms to understand how both digital and traditional advertising drive sales within the same channel (e.g., digital advertising affecting online sales) and across channels (e.g., digital advertising affecting offline sales). We develop a Dynamic Linear Model (DLM) to measure these effects. The model addresses: (1) the endogeneity of advertising, (2) dynamic advertising effects, (3) a multivariate dependent variable (4) heterogeneity across markets, and (5) competitive advertising effects. It decomposes advertising’s impact into customer counts and spend. We calibrate the model using data from a large, upscale retailer. We estimate elasticities for traditional (offline), online display (banner) and online search advertising on online and offline sales. Further, we develop and test hypotheses on how advertising impacts own- and cross-channel sales. We find that cross-channel effects exist and are important. For example, much of the impact of digital advertising can be attributed to its effect on the offline channel, primarily because of the impact on customer count.
Note: Research papers posted on SSRN, including any findings, may differ from the final version chosen for publication in academic journals.