May the Best Company Win: 94 Earnings Calls That Show Their Work

1Introduction

Every earnings season, executives claim to be taking share — from a rival's misstep, a capacity exit, a product gap. The claim is cheap; the evidence rarely is. That gap makes 'we're winning their customers, and here's why' one of the more interesting things a call can say, because management has volunteered a checkable story about demand shifting between companies rather than growing overall. For analysts, such statements mark where competitive dynamics, pricing power, and guidance risk concentrate. This study examines the 94 calls in Artul.ai's 2015–2024 library that answered yes to that hypothesis — 19.5% of the 481 flagged — measuring their language profile, trope mix, guidance behavior, trend across years, and the returns that followed.

2Data & methodology

The corpus comprises 481 earnings-call transcripts published between 2015 and 2024, 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 "Winning business away from competitors, with a stated reason why" (n = 94; 19.5% of the reference set, 95% Wilson interval 16.2%–23.3%). 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

The tone profile reads like a confident sales pitch: confidence up 0.26 (7.6 vs 7.33), promotion up 0.21 (5.3 vs 5.09), specificity up 0.12 (7.77 vs 7.64), and stress down 0.3 (2.07 vs 2.37), with candor and evasion essentially flat. Guidance skews positive — 30.9% raised versus a 22.0% base — while the flag itself registered 0.00 in 2019 and 2020 before peaking at 0.14 in 2022 and easing to 0.05 by 2024. Trope lifts reinforce the growth-story flavor: 'A Tiny Fraction of the Market' (1.3x), 'Early Products Growing Fast' (1.29x), and 'Consolidation Among Peers' (1.27x) all run hot, while 'When the CFO Dominates' (0.58x) runs cold, hinting these are CEO-fronted narratives. Returns are a wash across the 46 calls with returns: median -7.18% versus -10.54%, beat rates 37.0% versus 37.4%.

4Discussion

A careful reader should conclude that calls claiming competitive wins with stated reasons sound more confident, lean harder on growth tropes, and raise guidance more often — and that none of it coincided with better outcomes in this sample: the beat rate was 37.0% versus a 37.4% base, and the median return of -7.18% remained negative. One should not conclude that the share-win stories caused the guidance behavior, that the language caused the returns, or that these calls are more or less truthful than others. Confidence and correctness are different things; a stated reason is a story, not an audit.

5Limitations

These fields are AI-read and inherit machine judgment's noise: a flag can be missed or hallucinated, and every profile delta is only as good as that reader. The returns sample is 22,449 calls skewed to liquid names, so the comparison may not travel to the broader universe. Our own forward tests falsified directional prediction outright. And because LLMs partially remember famous stocks' history, backtests on well-known names are contaminated. Treat every figure here as descriptive, not investable. 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.