The path each company walked: revenue vs. headcount
Twenty years of revenue against headcount for eight companies, each plotted against its own pre-Covid trend line. For fifteen years every software company walked a straight line on log-log axes: more revenue, proportionally more people. After 2022 every one of those paths bends below its own line, while Delta, the non-tech contrast, stays on its trend. The bend is the story: growth continued, hiring did not follow.
What does it show?
On a log-log plot of revenue against headcount, every software company tracked its own straight line for fifteen years — and every one bends below that line after 2022. Delta, the non-tech contrast, does not.
Methodology
Each panel plots one company's annual whole-company revenue (x, log scale) against fiscal year-end headcount (y, log scale), 2005 to its latest filed year. The dashed reference is that company's own OLS fit of log headcount on log revenue over its pre-Covid years (through 2019), extended across the panel; a path on the line means hiring tracked revenue at a stable elasticity. Dots are coloured by era. Revenue is from SEC XBRL company facts and the dashboard's audited Big Tech dataset; headcount is from 10-K workforce disclosures, with values not already in the dashboard's audited files extracted from filing text and committed with their filing URLs in research/employment-growth-panel.json. Delta Air Lines is included as a non-tech contrast (its headcount is XBRL-tagged). M&A moves employees without matching revenue timing (Oracle/Sun 2010, Microsoft/Nokia 2014), and fiscal calendars differ across companies. Descriptive: a bend shows the relationship changed, not why.