This document describes a technique of dynamically throttling a promotion or content item placement to reduce marketing spending using machine-learning. A data processing system can determine a click score or an auction score based on various factors. The data processing system can further determine a threshold, for example, by predicting an annoyance effect of showing the promotion or the content item to a user. If the click score or the auction score is below the threshold, the data processing system can throttle the promotion or the content item placement such that the promotion or the content item is not shown to the user.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Fu, Bo; Bansod, Sourabh; Jain, Kunal; Price, Thomas; Chandrasekaran, Deepak Ramamurthi Sivaramapuram; Gupta, Prachi; Singh, Sarvjeet; Chew, Sue Yi; and Xie, Jierui, "Dynamic Throttling of In-App Promotions to Reduce Marketing Spend Based on Machine-Learning", Technical Disclosure Commons, (March 01, 2017)