Streaming platforms are often analyzed through the visible operations of their recommender systems: personalized rows, rankings, thumbnails, search results, and "because you watched" prompts. Existing scholarship has shown that such systems participate in algorithmic taste-making, influence aesthetic choice, and shape the cultural visibility of audiovisual works. Less attention has been paid, however, to the upstream professional and institutional processes through which content becomes available for recommendation in the first place. This article introduces the concept of pre-algorithmic acquisition infrastructure to describe the human, legal, commercial, and metadata-based practices through which audiovisual works are selected, licensed, classified, localized, positioned, and made operationally available before algorithmic recommendation occurs. The term pre-algorithmic is used as an analytical distinction rather than a strict chronological claim: acquisition and commissioning may themselves be informed by platform data, but they still perform the infrastructural work of making titles legally available, metadata-enabled, and operationally actionable for future recommendation. Using a structured analytical review and conceptual framework development approach, the article proposes a five-layer model of streaming curation: acquisition curation, rights curation, metadata curation, algorithmic curation, and interface curation. It further introduces recommendation-readiness, exploitation content, exploration content, and bridge content to explain how acquisition strategy shapes what platforms can recommend, personalize, and monetize. The central argument is that algorithms personalize choice, but acquisition infrastructure produces the conditions of choice.
T. S. Kosenko (Thu,) studied this question.