Research › Management Behavior
Artul.ai Research LibraryStudy No. 74Management BehaviorUpdated 2026-08-28

37.3% of earnings calls score as promotional - and 26.1% raised guidance

By Artul.ai Research Group · n = 61,671 earnings calls · First published 2026-08-28
Abstract

This study examines which earnings calls read as promotional, defined as calls scoring 6 or higher on a 0-9 promotion meter. Of 165,182 calls from 1990 to 2026, 61,671 (37.3%, 95% CI 37.1%-37.6%) meet the threshold. Promotional calls skew more confident (7.68 vs 7.21) and more evasive (2.86 vs 2.70), but slightly less candid (6.57 vs 6.86). Guidance mixes differ: 26.1% of promotional calls raised guidance versus 21.1% of others, while 7.1% lowered guidance versus 11.6%. Phrases like 'Scale-Dependent Advantage Claims' appear 1.8x more often in promotional calls. In a 7,706-call subset with returns data, the median next-period return was -9.3% versus -7.2% for the 22,449-call base.

Key findings
  • 37.3% of the 165,182 calls in the 1990-2026 corpus score 6 or higher on the 0-9 promotion meter (95% CI 37.1%-37.6%).
  • Promotional calls show higher confidence (7.68 vs 7.21) and higher evasion (2.86 vs 2.70), but lower candor (6.57 vs 6.86) than other calls.
  • 26.1% of promotional calls raised guidance versus 21.1% of non-promotional calls, while 11.6% of non-promotional calls lowered guidance versus 7.1% of promotional calls.
  • Among 7,706 promotional calls with return data, the median return was -9.3% versus -7.2% for the 22,449-call base, with 38.2% beating versus 39.5% in the base.

1Introduction

Earnings calls are the single most-watched corporate communication event, and how promotional a call sounds - versus how candid - shapes how analysts and investors hear the quarter. Because promotion scoring is now automated at scale across 165,182 calls spanning 1990 to 2026, it is possible to ask what actually distinguishes calls that read as promotional: their tone profiles, their guidance behavior, and the language patterns they contain. This study describes calls scoring 6 or higher on the 0-9 promotion meter, compares them with all other calls, and reports how their share has shifted over time and how their subsequent returns compare with the broader sample.

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 scoring 6 or higher on the 0–9 promotion meter (n = 61,671; 37.3% of the reference set, 95% Wilson interval 37.1%–37.6%). 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

Promotional calls are more confident (7.68 vs 7.21) and more evasive (2.86 vs 2.70), yet slightly less candid (6.57 vs 6.86) and less specific (7.47 vs 7.56). Their guidance mix tilts positive: 26.1% raised guidance versus 21.1% elsewhere, and only 7.1% lowered guidance versus 11.6%. Language lifts are led by 'Scale-Dependent Advantage Claims' at 1.8x, with 'A Tiny Fraction of the Market' at 1.66x and 'Founder-Led Companies' at 1.51x; 'When the CFO Dominates' appears 0.53x as often. The annual share of promotional calls jumped from 34.96% in 2020 to 48.44% in 2021, easing to 39.17% by 2024. In the returns subset, the median return was -9.3% versus -7.2% in the base, and 38.2% beat versus 39.5%.

Table 1. Mean behavioral scores (0–9 scale), study group versus baseline
MeterStudy groupBaselineΔ
Candor6.576.86-0.30
Evasion2.862.70+0.16
Specificity7.477.56-0.09
Stress2.352.43-0.08
Promotion6.445.05+1.38
Confidence7.687.21+0.47
Table 2. Guidance actions, study group versus baseline
ActionStudy groupBaseline
Raised26.1%21.1%
Maintained45.1%48.8%
Lowered7.1%11.6%
Withdrawn1.7%2.7%
Table 3. Co-occurring battery signals ranked by lift (group prevalence ÷ baseline prevalence)
SignalLiftIn groupBaseline
Scale-Dependent Advantage Claims1.80×20.0%11.1%
A Tiny Fraction of the Market1.66×49.9%30.0%
Founder-Led Companies1.51×30.3%20.0%
Early Products Growing Fast1.45×55.9%38.5%
Deferred Revenue Growing1.45×12.8%8.9%
When the CFO Dominates0.53×7.5%14.1%
201532.10%
201631.25%
201733.41%
201834.95%
201933.88%
202034.96%
202148.44%
202241.15%
202339.10%
202439.17%
202533.25%
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-9.3%-7.2%
Interquartile range-30.9% to +12.9%
Share beating SPY38.2% (95% CI 37%–39%)39.5%
Observations7,70622,449
Table 5. Most recent calls matching the study definition
TickerQuarterCall dateCall grade
SBFGQ2 20252025-07-25A
AONQ2 20252025-07-25C
FLGQ2 20252025-07-25B
FRSTQ2 20252025-07-25A
CHTRQ2 20252025-07-25C+
BAHQ1 20262025-07-25C+
ASPSQ2 20252025-07-25D
AAFRFQ1 20262025-07-25B

4Discussion

A careful reader should conclude that promotional-sounding calls, as measured here, co-occur with higher confidence, more raised guidance, and distinctive language - and that their subsequent median return in this sample was somewhat lower than the base. None of these patterns establishes that promotion causes outcomes, that scoring predicts returns, or that any category of call offers a trading edge. The returns comparison is descriptive only, the samples differ in composition, and differences of this size can arise from which companies and periods are included. Treat every figure as a description of this corpus, not a forecast.

5Limitations

Promotion, candor, and confidence scores are AI-read fields and are inherently noisy, so category boundaries should not be over-interpreted. The returns sample covers only 7,706 of the promotional calls against a 22,449-call base skewed toward liquid names, limiting generalizability. Our own forward tests falsified directional prediction from these scores. Additionally, large language models partially remember famous stocks' histories, which can contaminate any backtest built on AI-read fields like these. 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). “37.3% of earnings calls score as promotional - and 26.1% raised guidance.” Artul.ai Earnings-Call Research Library, Study No. 74. https://artul.ai/research/promotional-hype-heavy-calls

Related studies

Specificity Is Next to Godliness: 161,829 Earnings CThe Truth Won't Set You Free: High Candor Is the Nor95.5% of earnings calls since 1990 clear the 6-pointStraight Talk, Stress Less: Low-Evasion Earnings Cal39% of earnings calls trip the complexity meter - anOnly 11.3% of earnings calls score high on uncertain
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.