如果一部分的轉換無法找出與廣告互動的關係，行銷人該怎麼做？

When some conversions cannot be directly attributed to ad interactions, using conversion modeling is the most effective solution. Conversion modeling leverages machine learning to estimate the impact of ad interactions that cannot be directly observed due to privacy restrictions (like cookie consent or browser limitations). This provides a more comprehensive and accurate understanding of overall ad performance, allowing marketers to make informed optimization decisions. Asking partners to use digital fingerprinting or alternative IDs (Option A) is problematic because these methods often violate user privacy and are increasingly restricted by regulations and browser policies. Avoiding ad buys on certain browsers or device types (Option C) is a reactive and limiting approach that sacrifices potential reach and valuable audience segments. Constantly modifying Smart Bidding campaigns daily (Option D) is counterproductive; Smart Bidding algorithms require time to learn and optimize effectively, and frequent, drastic changes hinder their performance.