Cost fluctuations are a natural and expected part of the Pinterest Performance+ experience, especially during the first two weeks of a campaign. This period is known as the Learning Phase. Algorithmic Training: During the early stages, Pinterest's AI is "exploring" the platform to find the users most likely to engage or convert. The delivery algorithm tests different placements (Home Feed vs. Search) and different audience segments, which leads to varying costs as it gathers data on what works best. Auction Dynamics: Pinterest uses a real-time auction system. Costs (CPMs and CPCs) fluctuate based on how many other advertisers are bidding for the same audience at any given moment. During high-traffic times or seasonal peaks, costs may spike; during slower periods, the algorithm may find "cheaper" inventory. Budget Flexibility: Performance+ campaigns use flexible budgeting, allowing the system to spend up to 25% more than your average daily budget on days when it identifies high-value opportunities. This results in daily cost variations but is designed to improve overall campaign efficiency over a 7-day rolling average. Likelihood of Action: The algorithm prioritizes "Likelihood of Action." In the beginning, because the system has less historical data for your specific ad, it may bid more aggressively to secure the initial conversions needed to "learn" your target customer profile.