A Tree Falls in the Earnings Forest: Substance Without an Audience
This study profiles earnings calls that answer YES to the hypothesis "Substance without an audience" — substantive calls that appear to lack an audience. Of 979 calls scored from 2015 to 2025, 97 (9.9%, 95% CI 8.2%–11.9%) fit. These calls show higher confidence (7.7 vs 7.36) and specificity (7.67 vs 7.65), lower stress (1.94 vs 2.35) and evasion (2.36 vs 2.67). They raised guidance 24.7% of the time and lowered it only 3.1%, versus 12.9% in the base. Theme lifts range from 1.35x down to 0.57x, and the share of such calls peaked near 12% in 2022 before falling to 0% in 2025.
- Calls in this group show higher confidence (7.7 vs 7.36) and lower stress (1.94 vs 2.35) than the base corpus.
- They lowered guidance only 3.1% of the time versus 12.9% for the base corpus, while raising it 24.7%.
- The theme "Volume About to Step Up" over-indexes at 1.35x, while "The Hidden Segment" under-indexes at 0.57x.
- The share of these calls peaked around 0.12 in 2022 and fell to 0.0 in 2025.
1Introduction
Most attention on earnings calls goes to the ones everyone watches — the crowded calls, the viral quotes, the meme-worthy exchanges. But a smaller population of calls appears substantive without drawing much attention: management speaks candidly, guidance is firm, and yet the call seems to lack the audience its quality might merit. Whether quality and attention actually travel together is an open question for anyone who screens transcripts for signal. This study examines 97 such calls out of 979 scored between 2015 and 2025, profiling their behavioral tone, guidance actions, thematic lifts, and year-by-year frequency.
2Data & methodology
The corpus comprises 979 earnings-call transcripts published between 2015 and 2025, each scored independently by a large language model on an identical 37-field battery: seven categorical business verdicts, eight 0–9 behavioral meters, and twenty yes/no judgments. The study group is defined as calls that answered YES to the research hypothesis "Substance without an audience" (n = 97; 9.9% of the reference set, 95% Wilson interval 8.2%–11.9%). Baseline figures use the set of calls on which this question was tested. Market outcomes join a fixed sample of 22,449 calls with twelve-month total returns in excess of SPY, measured from the first close after each call; this sample skews toward liquid U.S. names and is reported as descriptive history only.
3Results
Behaviorally, these calls read as composed: candor is slightly higher (6.79 vs 6.94 baseline shows the gap is small), evasion is lower (2.36 vs 2.67), and stress is notably lower (1.94 vs 2.35), while promotion (5.42 vs 5.08) and confidence (7.7 vs 7.36) run higher. Guidance actions skew positive: 24.7% raised and only 3.1% lowered, against 23.7% and 12.9% respectively in the base. Thematically, "Volume About to Step Up" (1.35x), "A Tiny Fraction of the Market" (1.33x), and "Calls That Read Rehearsed" (1.31x) over-index, while "The Hidden Segment" (0.57x) and "When the CFO Dominates" (0.61x) under-index. The trend series is volatile, peaking near 0.12 in 2022 and reading 0.0 in 2020 and 2025.
| Meter | Study group | Baseline | Δ |
|---|---|---|---|
| Candor | 6.79 | 6.94 | -0.14 |
| Evasion | 2.36 | 2.67 | -0.31 |
| Specificity | 7.67 | 7.65 | +0.02 |
| Stress | 1.94 | 2.35 | -0.41 |
| Promotion | 5.42 | 5.08 | +0.34 |
| Confidence | 7.70 | 7.36 | +0.34 |
| Action | Study group | Baseline |
|---|---|---|
| Raised | 24.7% | 23.7% |
| Maintained | 54.6% | 51.6% |
| Lowered | 3.1% | 12.9% |
| Withdrawn | 1.0% | 1.1% |
| Signal | Lift | In group | Baseline |
|---|---|---|---|
| Volume About to Step Up | 1.35× | 34.0% | 25.2% |
| A Tiny Fraction of the Market | 1.33× | 41.2% | 30.9% |
| Calls That Read Rehearsed | 1.31× | 50.5% | 38.6% |
| The Hidden Segment | 0.57× | 12.4% | 21.7% |
| Results Worse Than Direction | 0.59× | 25.8% | 43.6% |
| When the CFO Dominates | 0.61× | 9.3% | 15.1% |
| Ticker | Quarter | Call date | Call grade |
|---|---|---|---|
| CRGO | Q1 2024 | 2024-05-20 | C+ |
| SNCR | Q1 2024 | 2024-05-07 | C+ |
| LINC | Q1 2024 | 2024-05-06 | B+ |
| LUNG | Q1 2024 | 2024-05-01 | B+ |
| CERS | Q4 2023 | 2024-03-05 | B+ |
| RDCM | Q4 2023 | 2024-01-31 | A |
| CHT | Q4 2023 | 2024-01-30 | C |
| CSPI | Q4 2023 | 2023-12-12 | D |
4Discussion
A careful reader should treat this as a descriptive profile, not a verdict. These calls are measurably calmer, more confident, and less likely to cut guidance than the average call in the corpus — but the study measures association within a scored dataset, not causes or outcomes. Nothing here establishes that substance without an audience predicts returns, that low attention is exploitable, or that the 2022 peak and 2025 trough mean anything structural. The thematic lifts describe which framings co-occur with the label, not why. Read the numbers as a map of what these calls look like, nothing more.
5Limitations
All fields are AI-read from transcripts and carry labeling noise; the hypothesis label itself is a model judgment, not ground truth. Behavioral deltas of 0.2–0.4 points are small relative to scoring variance. The trend series is based on varying yearly call counts and a partial 2025 sample. Any returns-linked reading is further constrained because our own forward tests falsified directional prediction, and LLMs partially remember famous stocks' histories, contaminating backtests. No causal or predictive claim is supported by this data. See the full methodology, including the C1 pattern’s forward-test failure and the LLM-memorization finding.
Companion page: every company matching this hypothesis is listed at the question’s own page.