Research › Call Signals
Artul.ai Research LibraryStudy No. 1Call SignalsUpdated 2026-08-28

A Drop in the Bucket: Calls That Claim a Tiny Slice of a Big Market

By Artul.ai Research Group · n = 49,623 earnings calls · First published 2026-08-28
Abstract

We studied earnings calls where the model answered YES to the battery item "A Tiny Fraction of the Market" — 49,623 of 165,182 calls (30.04%) across 1990–2026. These calls skew promotional: promotion scores 5.83 versus a 5.05 baseline (+0.78), and 22.31% raised guidance versus 21.05% baseline. Management is also more confident (7.38 vs 7.21) and slightly more evasive (2.84 vs 2.70). The claim co-occurs strongly with scale-dependent advantage framing (2.48x lift) and founder-led narratives (1.60x). For 5,508 calls with returns, the median next-day return was -6.57% versus -7.16% for 22,449 baseline calls, with 41.96% beating versus 39.47% baseline — differences that are suggestive, not predictive.

Key findings
  • The "tiny fraction of the market" framing appears in 49,623 of 165,182 calls (30.04%), with a 95% CI of 29.82%–30.26%.
  • These calls score +0.78 higher on promotion (5.83 vs 5.05) and +0.16 on confidence (7.38 vs 7.21) than the corpus baseline.
  • "Scale-Dependent Advantage Claims" co-occur at 2.48x the baseline rate, the strongest over-representation in the study.
  • In the 5,508-call returns subsample, the median next-day return was -6.57% and 41.96% beat, versus 39.47% for the 22,449-call baseline.

1Introduction

Every pitch deck and every earnings call needs a way to say "we have room to grow." The most common version is the market-size claim: our product could serve a vast market, and today we hold only a tiny fraction of it. It is a framing device that promises scale without asserting much about the present. Because it is so common, it is worth asking whether it travels with a particular communication style — more promotion, more confidence, certain narrative companions — and whether it co-occurs with guidance behavior or market reactions. This study examines 165,182 earnings calls from 1990 to 2026, isolating the 49,623 calls where our model answered YES to "A Tiny Fraction of the Market" and profiling their language, guidance actions, and subsequent returns.

2Data & methodology

The corpus comprises 165,182 earnings-call transcripts published between 1990 and 2026, 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 where the model answered YES to the battery item "A Tiny Fraction of the Market" (n = 49,623; 30.0% of the reference set, 95% Wilson interval 29.8%–30.3%). Baseline figures use all scored calls. 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

The language profile is the clearest signal: calls using this framing score 5.83 on promotion versus 5.05 baseline (+0.78) and 7.38 on confidence versus 7.21 (+0.16), while candor runs slightly lower (6.58 vs 6.86). Companion narratives matter too: scale-dependent advantage claims appear at 2.48x baseline, early-products-growing-fast at 1.84x, and founder-led framing at 1.60x. Meanwhile, CFO-dominated calls are under-represented (0.64x) and "finished-story" calls at just 0.71x. Guidance is modestly more bullish (22.31% raised vs 21.05%). The trend shows a spike to 35.49% of calls in 2021, easing to 27.35% by 2025. Returns are nearly indistinguishable: median -6.57% vs -7.16%.

Table 1. Mean behavioral scores (0–9 scale), study group versus baseline
MeterStudy groupBaselineΔ
Candor6.586.86-0.28
Evasion2.842.70+0.15
Specificity7.357.56-0.21
Stress2.502.43+0.07
Promotion5.835.05+0.78
Confidence7.387.21+0.16
Table 2. Guidance actions, study group versus baseline
ActionStudy groupBaseline
Raised22.3%21.1%
Maintained44.0%48.8%
Lowered10.1%11.6%
Withdrawn2.2%2.7%
Table 3. Co-occurring battery signals ranked by lift (group prevalence ÷ baseline prevalence)
SignalLiftIn groupBaseline
Scale-Dependent Advantage Claims2.48×27.5%11.1%
Early Products Growing Fast1.84×70.8%38.5%
Founder-Led Companies1.60×32.0%20.0%
Deferred Revenue Growing1.50×13.3%8.9%
Volume About to Step Up1.42×40.3%28.5%
When the CFO Dominates0.64×9.1%14.1%
Pricing Recovering0.64×13.8%21.5%
The Finished-Story Tell0.71×3.1%4.4%
201527.85%
201625.90%
201726.30%
201828.07%
201929.61%
202026.68%
202135.49%
202234.08%
202332.29%
202431.94%
202527.35%
Figure 1. Share of all analyzed calls matching the study definition, by year.
Table 4. Twelve-month excess total returns versus SPY (descriptive history, not a signal)
StatisticStudy groupReturns sample
Median excess return-6.6%-7.2%
Interquartile range-31.5% to +17.1%
Share beating SPY42.0% (95% CI 41%–43%)39.5%
Observations5,50822,449
Table 5. Most recent calls matching the study definition
TickerQuarterCall dateCall grade
SBFGQ2 20252025-07-25A
FFICQ2 20252025-07-25B+
FRSTQ2 20252025-07-25A
ANQ2 20252025-07-25B
CHTRQ2 20252025-07-25C+
DBOEYQ2 20252025-07-25B+
LBTSFQ2 20252025-07-25C+
ASPSQ2 20252025-07-25D

4Discussion

A careful reader should conclude that the tiny-fraction framing is a stylistic signature: it clusters with promotional, confident language and growth-narrative companions, and it became notably more common during 2021's bull market. One should not conclude the framing causes anything, predicts returns, or identifies mispriced stocks. The returns differences here are small — a 41.96% beat rate versus 39.47% baseline, and a median return gap of less than one percentage point — well within the range that sampling variation and stock characteristics could produce. Treat this as a descriptive map of how management teams talk, not as a signal to trade on.

5Limitations

The battery items are AI-read classifications of noisy transcripts and may mislabel sarcasm, hedging, or transcription errors. The returns subsample covers 22,449 calls and is skewed toward liquid, heavily covered names, so results may not generalize. Our own forward tests falsified directional prediction from these features, and no result here should be read as an edge. Additionally, LLMs partially remember famous stocks' histories from training data, which can contaminate any backtest that relies on model judgments. All findings are co-occurrences measured on this corpus; they establish association, not causation, and the confidence intervals reflect sampling uncertainty only. See the full methodology, including the C1 pattern’s forward-test failure and the LLM-memorization finding.

Cite this study Artul.ai Research Group (2026). “A Drop in the Bucket: Calls That Claim a Tiny Slice of a Big Market.” Artul.ai Earnings-Call Research Library, Study No. 1. https://artul.ai/research/a-tiny-fraction-of-the-market-earnings-calls

Related studies

Ammunition, Not Smoke: Inside the Calls Where CriticKnow What You Know: Calls Where Confidence Matched tThe Question Left Hanging, Fewer and Farther BetweenThe Guidance Was Fine All Along: What Model-EndorsedTalking the Skeptic Down: 66% of Calls Leave the DouThe Admission Is Not the Confession: Calls Flagged a
Not investment advice. Artul.ai publishes AI-generated earnings-call quality grades and expected-volatility estimates — never buy or sell recommendations. We tested over 1,600 predictive hypotheses against 165,000 transcripts; the honest result, including what failed, is documented in our methodology.