Custom audiences let advertisers target people who already have a relationship with the business, such as past website visitors or an uploaded customer list, while lookalike (or 'advantage+') audiences ask the platform to find new users who statistically resemble that existing group. Since Apple's App Tracking Transparency policy launched in 2021, requiring apps to ask permission before tracking users across other apps and websites, platforms like Meta have had less precise visibility into off-platform conversions, which pushed advertisers toward first-party data (like a Conversions API connection) and broader, algorithm-driven targeting rather than narrow manual targeting. A/B testing a paid social ad means changing one variable at a time -- such as the image, headline, or call-to-action -- while keeping everything else constant, so that any difference in performance can be attributed to that one change rather than noise. Because paid social platforms reward ads that get engagement quickly, testing several creative variations early and reallocating budget to the winner is usually more effective than committing a full budget to a single untested ad.