Using data from the SEC’s EDGAR server log, we examine the consumption of financial information in filings from 2003 through 2012. We document how quickly download counts for EDGAR forms decay over the year following their initial filing. We find that different types of disclosures experience very different information flows as measured by download activity. After examining a variety of the most common forms, we focus on annual reports (10-K) and registration statements (S-1). The magnitude of the daily EDGAR requests for 10-Ks is relatively low and shows a surprisingly small difference between firms with and without publicly-traded equity. (Coauthor B. McDonald.)