The Numbers Are Fine; Everything Else Is Pending: Flagging Unbooked Growth on Calls
We studied 992 earnings-call transcripts from 2015 to 2025 and identified the 426 calls (42.9%, 95% CI 39.9% to 46.0%) where management answered YES to the research hypothesis "Identified near-term contributors not yet in the reported numbers" — in effect, teasing growth not yet visible in reported results. These calls show modestly higher specificity (7.71 vs 7.66), promotion (5.24 vs 5.07), and confidence (7.47 vs 7.36), with slightly lower evasion (2.61 vs 2.66) and stress (2.31 vs 2.35). The standout language pattern is "Volume About to Step Up," which appears 1.58 times more often on these calls. Yet forward returns were not better: the median was -0.0947 vs -0.0697 for the base, and 40.1% beat versus 41.1%.
- 426 of 992 calls (42.9%) flagged near-term contributors not yet in reported numbers.
- These calls score higher on promotion (5.24 vs 5.07) and confidence (7.47 vs 7.36), with slightly lower evasion (2.61 vs 2.66).
- The phrase "Volume About to Step Up" is 1.58x more common on these calls than on others.
- Median forward returns were -0.0947 versus -0.0697 for the base sample, with 40.1% beating versus 41.1%.
1Introduction
Every earnings call contains a quiet genre of optimism: management pointing to revenue contributors that have not yet landed in reported numbers — new volumes ramping, pipelines converting, deals about to close. For listeners, these forward-looking teasers are seductive because they suggest the printed quarter understates the business. But how common is this framing, how does the tone of such calls differ, and does anything show up in subsequent returns? Using 992 transcripts scored by Artul.ai's language models across 2015-2025, this study examines the 426 calls that answered YES to the hypothesis "Identified near-term contributors not yet in the reported numbers."
2Data & methodology
The corpus comprises 992 earnings-call transcripts published between 2015 and 2025, 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 "Identified near-term contributors not yet in the reported numbers" (n = 426; 42.9% of the reference set, 95% Wilson interval 39.9%–46.0%). 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
Calls flagging unbooked contributors make up 42.9% of the corpus (95% CI 39.9%-46.0%). Their language profile tilts promotional: promotion rises to 5.24 from 5.07 and confidence to 7.47 from 7.36, while evasion (2.61 vs 2.66) and stress (2.31 vs 2.35) dip slightly — a picture of smoother, more assertive delivery. The most distinctive marker is "Volume About to Step Up," 1.58x more frequent (39.4% vs 25.0%). Guidance behavior is nearly identical to the base. The returns comparison is sobering: median forward return of -0.0947 vs -0.0697, mean -0.1007, and a 40.1% beat rate vs 41.1% for the base of 382 scored calls.
| Meter | Study group | Baseline | Δ |
|---|---|---|---|
| Candor | 6.95 | 6.94 | +0.01 |
| Evasion | 2.61 | 2.66 | -0.05 |
| Specificity | 7.71 | 7.66 | +0.06 |
| Stress | 2.31 | 2.35 | -0.04 |
| Promotion | 5.24 | 5.07 | +0.17 |
| Confidence | 7.47 | 7.36 | +0.11 |
| Action | Study group | Baseline |
|---|---|---|
| Raised | 25.8% | 23.4% |
| Maintained | 52.3% | 51.8% |
| Lowered | 12.7% | 12.7% |
| Withdrawn | 0.9% | 1.1% |
| Signal | Lift | In group | Baseline |
|---|---|---|---|
| Volume About to Step Up | 1.58× | 39.4% | 25.0% |
| Statistic | Study group | Returns sample |
|---|---|---|
| Median excess return | -9.5% | -7.0% |
| Interquartile range | -30.8% to +9.6% | — |
| Share beating SPY | 40.1% (95% CI 33%–48%) | 41.1% |
| Observations | 152 | 382 |
| Ticker | Quarter | Call date | Call grade |
|---|---|---|---|
| FRBA | Q2 2024 | 2024-07-26 | A |
| TPH | Q2 2024 | 2024-07-25 | A |
| MMLP | Q2 2024 | 2024-07-18 | B |
| ASO | Q1 2024 | 2024-06-11 | C+ |
| BBY | Q1 2025 | 2024-05-30 | C |
| CRGO | Q1 2024 | 2024-05-20 | C+ |
| BIOX | Q3 2024 | 2024-05-14 | C+ |
| ERO | Q1 2024 | 2024-05-10 | A |
4Discussion
The honest reading is that this flag measures a rhetorical posture, not an outcome. Calls that tout contributors not yet in the numbers do sound different — more promotion, more confidence, more volume language — but the subsequent returns and beat rates are statistically indistinguishable from, and slightly below, the base. A careful reader should conclude that the signal identifies a style of storytelling, not a predictor. Nothing here establishes that such calls cause better or worse results, nor that the flagged contributors ever materialize; the data simply show how often the pattern appears and what accompanies it.
5Limitations
All fields are AI-read and noisy; tone scores and hypothesis labels inherit model error. The returns sample covers only 152 of the flagged calls against a base of 382, drawn from a broader universe of 22,449 calls skewed toward liquid names, so it is not representative. Our own forward tests falsified directional prediction from these signals. Additionally, LLMs partially remember famous stocks' histories, contaminating any backtest with leaked knowledge of outcomes. 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.