Researchers have developed a novel approach to forecasting cost-per-click (CPC) in paid search auctions by incorporating competition-aware signals. Utilizing 1.66 billion Google Ads log records from a car-rental market, they constructed a weekly panel of keyword series and developed proxies based on semantic neighborhoods, CPC trajectory alignments, and geographic intent. These signals were evaluated in spatiotemporal graph forecasters, outperforming baseline models. The findings indicate that graph-based models are most effective for short-term (one-week) forecasting, while covariate-augmented foundation models excel at longer horizons (six and twelve weeks), particularly for high-CPC, high-volatility keywords. AI
IMPACT This research could improve forecasting accuracy in advertising markets, leading to more efficient ad spend and better campaign planning.
RANK_REASON This is a research paper detailing a new methodology for forecasting in a specific domain. [lever_c_demoted from research: ic=1 ai=0.7]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
