Hypothesis Tested: Earned edge, pressed harder

Some companies claim an edge they've earned — a durable advantage they can point to — and this study asks what happens when management presses that claim harder on the call. Among 485 calls in our 165,000-transcript corpus where an AI reader detected this pattern, 183 (about 38%) also showed the intensified version. Anyone who follows earnings calls knows the "our moat is widening" cadence; here we can finally check what it tends to look like, and what followed.

The intensified group sounded more promotional (5.47 vs 5.09) and more confident (7.58 vs 7.32), with slightly higher specificity (7.71 vs 7.65) and similar candor. These calls co-occurred with "Early Products Growing Fast" (1.45x lift) and "A Tiny Fraction of the Market" (1.44x). Guidance tells a sharper story: 28.4% raised guidance versus 22.1% in the base, and only 8.2% lowered it versus 13.4%. Yet the 84 calls with return data showed a median return of -15.4%, worse than the -8.6% baseline.

These AI-read fields are noisy, and the return figures are descriptive history from a 22,449-call sample, not a trading signal — our own forward tests falsified directional prediction. Language models also partially remember famous stocks' histories, which contaminates any backtest. Treat the co-occurrences and profile gaps as patterns in language, not causes of outcomes.

Behavior profile

Guidance actions in this group

What travels with it

Share of calls by year

Recent examples

This hypothesis has its own page with every matching company: see the full question.