Have Your Cake and Expand the Base Too: Repeat Growth With New Logos

1Introduction

For anyone who parses earnings calls, the dual-engine story - expansion revenue growing inside the existing base while new customer logos keep landing - is one of the most repeated claims in software and subscription businesses. It signals durable demand on two fronts, and management teams lean on it heavily in prepared remarks. But how often does it actually appear on calls, what does the surrounding language look like, and how have those calls traded afterward? Using Artul.ai's research library of AI-read fields across 2015-2024, this study examines the 186 calls that answered YES to the hypothesis 'Repeat expansion inside the base while new logos still arrive' and profiles their candor, confidence, guidance behavior, and subsequent returns.

2Data & methodology

The corpus comprises 499 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 "Repeat expansion inside the base while new logos still arrive" (n = 186; 37.3% of the reference set, 95% Wilson interval 33.1%–41.6%). 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 186 qualifying calls (37.3% of the 499-call corpus) read measurably more confident than the rest: confidence 7.54 vs 7.33, promotion 5.41 vs 5.11, and stress 2.23 vs 2.39, with candor essentially flat (6.85 vs 6.91). Topic lifts reinforce the growth framing - 'Early Products Growing Fast' appears at 1.41x the base rate (54.8% vs 38.9%), 'Founder-Led Companies' at 1.41x, and 'A Tiny Fraction of the Market' at 1.39x - while 'Pricing Recovering' runs below base (0.70x). Guidance behavior barely differs: 25.3% raised versus 21.8% in the base, and 12.9% lowered versus 13.2%. Post-call returns tell a mixed story: median -5.7% versus -10.5% in the base, but a mean of -8.5% and only 42.9% beating their benchmark.

4Discussion

A careful reader should conclude that this dual-expansion pattern is common - roughly a third of the corpus - and that the calls carrying it use more confident, promotional language and lean on growth topics like early product traction. The median post-call return is less negative than the base's, but the gap between median (-5.7%) and mean (-8.5%) shows a skewed distribution, so averages alone are misleading. What should not be concluded: that this pattern causes better outcomes, predicts returns, or offers a trading edge. The 42.9% beat rate has a confidence interval (32.4%-54.0%) that comfortably includes a coin flip, and all figures here are descriptive, not causal.

5Limitations

All fields in this study are AI-read and therefore noisy; language scores like candor and confidence inherit model error. The returns sample is only 77 of 186 calls, drawn from a larger universe of 22,449 calls skewed toward liquid names, so selection effects loom large. Our own forward tests falsified directional prediction from these features - none of the observed differences constitutes an edge. Additionally, LLMs partially remember famous stocks' histories, which can contaminate any apparent backtest relationship between call language and later returns. Treat every number as a descriptive snapshot of a noisy measurement process, not as evidence of a repeatable signal. 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.