Abstract
Conversion funnels represent the journey of a consumer through a marketing campaign resulting in a sale. The described techniques incorporate marketer inputs and apply machine learning to optimize conversion funnels in online advertising. The techniques incorporate initial inputs from marketers and apply machine learning techniques to optimize conversion funnels in the context of advertising. The initial inputs include definitions of states and optimal actions to move consumers to next states of a funnel. With user permission, a machine learning model, e.g., that uses reinforced learning is randomly applied to try other actions defined by marketers (e.g., for other states) to collect training data on various actions. The training data is used to train the model and eliminates the requirement of a large amount of training data to be provided at the beginning of the process. The trained model thus obtained predicts optimal actions to move the consumer through the conversion funnel. Predicted actions using the model for a given state, along with supporting evidence, are provided to marketers for review and approval. Marketers can modify the funnel parameters, e.g., states and actions, based on such evidence.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Wang, Gang and Mao, Yiran, "Conversion funnel optimization using machine learning", Technical Disclosure Commons, (February 21, 2018)
https://www.tdcommons.org/dpubs_series/1061