Research › 20-Year Trends
Artul.ai Research LibraryStudy No. 8620-Year TrendsUpdated 2026-08-28

14,649 calls flagged 'Deferred Revenue Growing' raised guidance 33% vs 21% baseline

By Artul.ai Research Group · n = 14,649 earnings calls · First published 2026-08-28
Abstract

This study examines 14,649 earnings calls (out of 165,182) where Artul.ai's model answered YES to "Deferred Revenue Growing," spanning corpus years 1990 to 2026. These calls show a distinct communication profile: promotion framing of 5.57 vs 5.05 baseline, confidence of 7.54 vs 7.21, and stress of 2.21 vs 2.43. On guidance, 33.0% raised versus 21.1% in the base, while only 9.2% lowered versus 11.6%. The overexpressed phrase "Early Products Growing Fast" appears 1.67x more often, and "The Finished-Story Tell" is underexpressed at 0.68x. In a matched returns sample of 2,336 calls, the median next-period return was -0.46% versus -7.16% for the 22,449-call base, and 49.6% of these calls were followed by results that beat, versus 39.5% baseline. These are descriptive associations, not causal claims.

Key findings
  • 8.9% of calls in the 165,182-call corpus (14,649) were flagged YES for "Deferred Revenue Growing."
  • Guidance was raised on 33.0% of flagged calls versus 21.1% of baseline calls, and lowered on 9.2% versus 11.6%.
  • The phrase "Early Products Growing Fast" appears 1.67x more often on flagged calls; "The Finished-Story Tell" appears 0.68x as often.
  • In the returns sample of 2,336 calls, the median next-period return was -0.46% versus -7.16% for the base, with 49.6% beating versus 39.5%.

1Introduction

Deferred revenue is one of the most closely watched line items for subscription and prepayment businesses, so whether a management team emphasizes its growth on an earnings call is a meaningful signal about how they are framing the quarter. Artul.ai's research library tracks model answers to fixed questions across every call transcript, allowing large-scale comparison of calls where the model answered YES to "Deferred Revenue Growing" against the full corpus of 165,182 calls from 1990 to 2026. This study describes how those 14,649 calls differ in language profile, guidance behavior, phrase usage, and subsequent returns, and how the pattern has trended since 2015.

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 "Deferred Revenue Growing", tracked by year (n = 14,649; 8.9% of the reference set, 95% Wilson interval 8.7%–9.0%). 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

Flagged calls skew promotional and confident: promotion framing runs 5.57 versus 5.05 in the base, confidence 7.54 versus 7.21, and stress is lower at 2.21 versus 2.43. Guidance behavior mirrors the emphasis: 33.0% of flagged calls raised guidance versus 21.1% baseline, while 9.2% lowered it versus 11.6%. Phrase analysis shows "Early Products Growing Fast" at 1.67x the base rate and "Founder-Led Companies" at 1.54x, while "The Finished-Story Tell" is underexpressed at 0.68x. The annual trend rises from 7.21 in 2015 (3,483 calls) to a peak of 10.69 in 2021 (18,253 calls), easing to 8.78 in 2024. In the returns sample, the median next-period return was -0.46% versus -7.16% for the base.

Table 1. Mean behavioral scores (0–9 scale), study group versus baseline
MeterStudy groupBaselineΔ
Candor6.756.86-0.11
Evasion2.732.70+0.03
Specificity7.627.56+0.06
Stress2.212.43-0.22
Promotion5.575.05+0.52
Confidence7.547.21+0.33
Table 2. Guidance actions, study group versus baseline
ActionStudy groupBaseline
Raised33.0%21.1%
Maintained44.5%48.8%
Lowered9.2%11.6%
Withdrawn1.9%2.7%
Table 3. Co-occurring battery signals ranked by lift (group prevalence ÷ baseline prevalence)
SignalLiftIn groupBaseline
Early Products Growing Fast1.67×64.4%38.5%
Founder-Led Companies1.54×30.9%20.0%
A Tiny Fraction of the Market1.50×45.2%30.0%
Volume About to Step Up1.29×36.7%28.5%
The Finished-Story Tell0.68×3.0%4.4%
20157.21%
20167.76%
20178.32%
20188.52%
20198.04%
20208.11%
202110.69%
202210.28%
20238.91%
20248.78%
20259.58%
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-0.5%-7.2%
Interquartile range-21.5% to +21.5%
Share beating SPY49.6% (95% CI 48%–52%)39.5%
Observations2,33622,449
Table 5. Most recent calls matching the study definition
TickerQuarterCall dateCall grade
MOG.AQ3 20252025-07-25B+
DBOEYQ2 20252025-07-25B+
CSBRQ4 20252025-07-25C+
COURQ2 20252025-07-24A
DLRQ2 20252025-07-24B
ESRTQ2 20252025-07-24C+
SKYWQ2 20252025-07-24B+
BXQ2 20252025-07-24B

4Discussion

A careful reader should conclude that calls emphasizing deferred-revenue growth are, descriptively, associated with more promotional language, more raised guidance, and a smaller median drawdown afterward. That is a correlation observed in historical data, not evidence that the language caused anything or that it will persist. The returns sample covers only 2,336 of the 14,649 flagged calls, so selection matters. The gap in beat rates (49.6% versus 39.5%) and medians (-0.46% versus -7.16%) should be read as context about how these calls cluster, not as a signal to trade on.

5Limitations

The YES/NO fields are produced by an AI model reading transcripts and are noisy; misclassification of ambiguous deferred-revenue discussion is likely. The returns sample covers 22,449 calls, skewed toward liquid names, so it does not represent the full corpus. Artul.ai's own forward tests falsified directional prediction from these fields, and LLMs partially remember famous stocks' histories, contaminating any backtest. All figures here are descriptive of past data and support no causal inference or trading edge. 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). “14,649 calls flagged 'Deferred Revenue Growing' raised guidance 33% vs 21% baseline.” Artul.ai Earnings-Call Research Library, Study No. 86. https://artul.ai/research/the-rise-and-fall-of-prepaid-growth

Related studies

Give Them Nothing: The Calls Where Critics FeastYes Means Yes: Inside the Earnings Calls Where the SThe Question Was Answered. Just Not in This Room: HaRehearsal Season: The Earnings Calls That Sound a LiWhen the model flags 'Volume About to Step Up' on 28Like Father, Like Firm: Founder-Led Earnings Calls,
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.