This document discusses the major work required to build a unified discovery index from start to being ready for sale within one year. It notes that building such an index involves acquiring content from thousands of publishers, normalizing and deduplicating over 75 million records stored across 75 terabytes of data, and developing complex indexing models to deliver relevant search results. The process of acquiring, cleaning, and indexing vast amounts of content from various sources into a single searchable index is described as a very messy job that requires strong partnerships, funding, experienced teams, and ongoing maintenance.