Hypothesis Tested: Substance without an audience

Some earnings calls deliver real substance — specific numbers, candid discussion, concrete guidance — but almost nobody seems to be listening: low analyst engagement, thin coverage, no audience. This study isolates that pattern, "substance without an audience," among 97 calls out of a corpus of 165,000 earnings-call transcripts that an AI read and scored against the same battery of questions. For anyone who follows calls closely, it asks a simple question: does good disclosure matter if the market isn't watching?

The group's calls score high on specificity (7.67 vs 7.65) and candor (6.79 vs 6.94), yet they co-occur strongly with calls that read rehearsed (1.31x lift, 50.5% vs 38.6%) and with "volume about to step up" (1.35x). They rarely appear alongside "results worse than direction" (0.59x) or "the hidden segment" (0.57x). Guidance tells a clearer story: only 3.1% lowered guidance versus 12.9% in the base. The pattern is also rare and spiky by year — 0.03 in 2015, near zero in 2020, a peak of 0.12 in 2022, and 0.02 in 2024.

Honest limits apply. The AI-read fields are noisy single-model judgments, and the co-occurrence lifts are descriptive, not causal. Any return statistics here are historical descriptions of a 22,449-call sample, not a trading signal — our own forward tests falsified directional prediction outright. Language models also partially remember famous stocks' histories, which contaminates any backtest built on their outputs. Treat this as a pattern observation, not an edge.

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.