I recently rebuilt how I pull the executive roster (CEO plus named officers, with titles) for US public companies straight from SEC data. Coverage went from about 6,400 officer rows across 3,242 companies to roughly 62,500 rows across 24,358 companies, so close to 10x the rows and 7.5x the company coverage. Here is the journey, because the naive approach fails in interesting ways. Attempt 1: the proxy statement (DEF 14A) The intuitive source is the annual proxy. Since fiscal 2022 the SEC standardized the "Pay versus Performance" disclosure as inline XBRL, and there is a tag literally called ecd:PeoName (Principal Executive Officer Name). Perfect, right? Not really. A lot of large filers tag the compensation numbers but never tag ecd:PeoName. Microsoft and Alphabet both returned exactly 0 officers for me this way.…