Making Ads Count: Using MMoE and Auxiliary Tasks to Better Connect Buyers & Sellers

Summary

Etsy’s Ads Search ranking model needed a stronger purchase-intent signal because clicks are noisy and purchases are sparse later in the buyer journey. It retained its shared feature-representation and interaction layers, then replaced the shared feed-forward network with Multigate Mixture of Experts (MMoE), using task-specific gates and towers for click-through rate (CTR) and post-click conversion rate (PCCVR). Add-to-cart was added as an auxiliary task, while favorites were rejected after producing negative transfer. Offline, the resulting model increased Purchase and Click PR AUCs by average 3.5% and 1%, respectively, and Purchase ROC AUC by 0.5%. Online, it improved purchases and purchase calibration, supplied better inputs to auto-bidding, kept model size flat through pruning, and reduced inference cost.

Takeaways

1. 1

   MMoE lets CTR and PCCVR use different weighted subsets of experts, allowing task-specific patterns while retaining shared representations.

2. 2

   Add-to-cart improved purchase metrics, while favorites were noisy and caused negative transfer; only add-to-cart was ramped up in production.

3. 3

   The deployed model improved purchase calibration for auto-bidding, kept overall model size flat through pruning, and made serving inference less costly.