Companies increasingly invest in digital marketing channels to promote their products and services. While the expenditures for each marketing channel are known, the contribution of marketing channels to a successful conversion, and therefore the value they generate, is unknown, but highly relevant for strategic decision-making. In this paper, we develop a novel, context-aware additive hazard marketing attribution (CAHMA) model based on survival analysis to address this problem. In addition to channel-specific, time-decaying effects of marketing on the users’ conversion rate, we control for the effects of contextual features, such as the device or country from which users interact with marketing channels. Based on a prototypical implementation, we demonstrate the model’s applicability and evaluate it on real-world data from the industry. We find that CAHMA outperforms other models in terms of accuracy while offering unique interpretability of the results and hence, providing deep insights for practitioners into the effects of marketing.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Context-Aware Marketing Attribution Based on Survival Analysis

  • Kilian Züllig,
  • Stefan Napirata,
  • Steffen Zimmermann

摘要

Companies increasingly invest in digital marketing channels to promote their products and services. While the expenditures for each marketing channel are known, the contribution of marketing channels to a successful conversion, and therefore the value they generate, is unknown, but highly relevant for strategic decision-making. In this paper, we develop a novel, context-aware additive hazard marketing attribution (CAHMA) model based on survival analysis to address this problem. In addition to channel-specific, time-decaying effects of marketing on the users’ conversion rate, we control for the effects of contextual features, such as the device or country from which users interact with marketing channels. Based on a prototypical implementation, we demonstrate the model’s applicability and evaluate it on real-world data from the industry. We find that CAHMA outperforms other models in terms of accuracy while offering unique interpretability of the results and hence, providing deep insights for practitioners into the effects of marketing.