The multi-billion-dollar search engine marketing (SEM) industry is all about gaining visibility for businesses on search engine results pages (SERPs) to bring customers to a firm’s website. On a micro level, a listing’s ability to attract attention is affected by other listings featured on the SERPs; as such, it is context specific. This paper develops a model that describes consumers’ moment-to-moment visual inspection process by incorporating their decisions about what listings to look at and whether to scroll additional search results (e.g., “below the fold”) into view. The results show that the distance to “hidden” parts of the screen, scrolling direction, and the time and percentage of listings viewed predicts consumers’ scrolling decisions. Low-level features (e.g., ad rank, bold font) are predictive of visual attention shifts at early stages of screen inspection; yet their impact is overshadowed by semantic context as inspection progresses. Transactional content discovered in the inspected listings invites repeat attention, while descriptive content discourages further inspection. The proposed model can help SEM practitioners create “visibility maps” of SERPs, which can be used to identify visual competitors, assess listing position performance, and diagnose problems in rank-bidding strategy.