Transforming Automotive Sales And Marketing: The Impact Of Data Engineering And Machine Learning On Consumer Behavior

Abstract

The impact of data engineering and machine learning is analyzed on the automotive sales and marketing performance. Specifically, the effect of the implementations from the automotive industry on consumer behavior is investigated. A regression analysis is conducted on the monthly sales of over 300 car models in the Chinese automotive market with respect to various marketing strategies, marketing network measures and geographic network measures, together with other auxiliary variables. Topological features of the marketing and geographic networks are used to predict their corresponding networks in the following month with a generalized linear model approach across this large-scale dataset. The marketing network seems to have much stronger explanatory power for the marketing scale compared to that of the geographic network and its other sales and network measures. With extensive experimental results, it is demonstrated that a boost in the model due to the recent hype in marketing scale can be a source of massive data leakage in the automotive market, resulting in potentially misleading actionable insights, their experimental results vanishing endeavour such patterns. In such a case, marketing scale cannot be causally attributed to success, instead it argues for paying more heed to network structure and suggests utilizing the recent methodology to mitigate this pervasive issue. In recent years, predictive machine learning techniques have been evolved and applied to analyze subsequent consumer behavior and build tailored marketing strategies. A huge cohort of startups have arisen to enable companies to target their consumers in a more personalized and direct way with the utilization of social media and other customer interaction and engagement data. These companies can gain a better understanding of consumers by creating predictive marketing algorithms that analyze consumer behavior and preferences, thereby delivering the right incentivized message at the right moment with the desired product. Nonetheless, the efficacy of this newly spawned industry has constantly been a subject for debate. From a consumer perspective, even a slight deviation in the strategy can revamp the recommendation, while from a marketing point of view, it is crucial to analyze the implications of each order. Given this contentious backdrop, the execution of veritable randomized controlled trials (RCTs) in a marketing context is handicapped by stringent operational issues. In light of this, a statistical model is proposed to irrefutably gauge the efficacy of the garnered results in marketing campaigns. Time is inherently intertwined in networks, so it is crucial to consider the time difference of edges .