CausalMTA: Eliminating the User Confounding Bias for Causal Multi-touch Attribution

  title={CausalMTA: Eliminating the User Confounding Bias for Causal Multi-touch Attribution},
  author={Di Yao and Chang Gong and Lei Zhang and Sheng Chen and Jingping Bi},
  journal={Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining},
  • Di Yao, Chang Gong, Jingping Bi
  • Published 21 December 2021
  • Computer Science
  • Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining
Multi-touch attribution (MTA), aiming to estimate the contribution of each advertisement touchpoint in conversion journeys, is essential for budget allocation and automatically advertising. Existing methods first train a model to predict the conversion probability of the advertisement journeys with historical data and calculate the attribution of each touchpoint by using the results counterfactual predictions. An assumption of these works is the conversion prediction model is unbiased. It can… 

Figures and Tables from this paper



Multi-touch Attribution in Online Advertising with Survival Theory

A new data-driven attribution model based on survival theory is proposed that is able to remove the presentation biases inherit to most of the other attribution models and is also able to predict user's 'conversion' probability.

Learning Multi-touch Conversion Attribution with Dual-attention Mechanisms for Online Advertising

A Dual-attention Recurrent Neural Network (DARNN) is proposed for the multi-touch attribution problem that learns the attribution values through an attention mechanism directly from the conversion estimation objective and combines both post-view and post-click attribution patterns together for the final conversion estimation.

Additional Multi-Touch Attribution for Online Advertising

Experimental results on a large real-world advertising dataset illustrate that the proposed multi-touch attribution model (AMTA) is superior to state-of-the-art techniques in conversion rate prediction and the credit allocation based on AMTA is reasonable.

CAMTA: Causal Attention Model for Multi-touch Attribution

This work proposes CAMTA, a novel deep recurrent neural network architecture which is a causal attribution mechanism for user-personalised MTA in the context of observational data, and utilizes the users' pre-conversion actions in a principled way in order to predict per-channel attribution.

Experimental Designs and Estimation for Online Display Advertising Attribution in Marketplaces

Two novel randomized designs are proposed to estimate the overall campaign attribution without placebo ads, and disaggregate the campaign presence and ad effects, and show the ex-ante value of continuing evaluation to enhance the user selection for ad exposure mid-flight.

Causally motivated attribution for online advertising

A causally motivated methodology for conversion attribution in online advertising campaigns is presented and it is argued that in cases where causal assumptions are violated, these approximate methods can be interpreted as variable importance measures.

A Time To Event Framework For Multi-touch Attribution

The model for user conversion behavior treats conversions as occurrences in an inhomogeneous Poisson process, while the attribution algorithm is based on iteratively removing the last ad in the path.

Data-driven multi-touch attribution models

An intuitive and simple probabilistic model to directly quantify the attribution of different advertising channels and a bagged logistic regression model, which shows achieves a comparable classification accuracy as a usualLogistic regression, but a much more stable estimate of individual advertising channel contributions.

Shapley Meets Uniform: An Axiomatic Framework for Attribution in Online Advertising

This work considers a Markovian model for the user journey through the conversion funnel, in which ad actions may have disparate impacts at different stages, and proposes a novel attribution metric, that is efficiently computable and implementable, and can be interpreted as a correction to the commonly used uniform attribution scheme.

Beyond the Last Touch: Attribution in Online Advertising

This work analytically analyze the inefficiencies created by externalities and uncertainty when information is symmetric between advertisers and publishers and finds that the prices paid in the market will decrease when more sophisticated attribution methods are adopted.