Methodology, EarningsCallAI, EarningsCallAI

Data Sources

Our primary data source is SEC EDGAR, the U.S. Securities and Exchange Commission's public filing system. Specifically, we extract structured numbers from XBRL-tagged 10-Q (quarterly) and 10-K (annual) filings, which use a standardized financial taxonomy so revenue means revenue across every filer. EDGAR data is in the public domain.

Earnings call transcripts and management commentary are sourced from SEC Form 8-K furnishings (typically the earnings release exhibit) and contemporaneous public commentary. These let us cross-reference structured financial trends against management framing on demand, pricing, competitive dynamics, and forward outlook.

How We Calculate the Operator Signal Score

Every company receives an Operator Signal Score classified as one of four signals:

TAILWIND, Revenue accelerating quarter-over-quarter, guidance raised, and management commentary positive on demand. Constructive signal for adjacent businesses and competitors, the market is expanding.
HEADWIND, Revenue decelerating, guidance lowered or cautious, and management flagging demand weakness or pricing pressure. Warning signal for businesses in the same space.
MIXED, Conflicting signals, strong revenue but cautious guidance, or positive commentary with declining margins. Requires nuance to interpret.
NEUTRAL, Results in line with expectations, no significant changes in trajectory. Business as usual.

The classification is based on revenue growth trajectory (accelerating vs. decelerating across the most recent four to eight quarters), guidance revisions (raised, maintained, or lowered), margin trends, and qualitative commentary analysis. Confidence (HIGH/MEDIUM/LOW) drops when the quantitative trend and qualitative commentary diverge.

Investment Themes

We aggregate individual company signals into sector-level Investment Themes, cross-cutting trends that multiple companies are confirming or denying in their earnings calls. Themes are tagged with their signal direction and the number of companies confirming them. Examples include AI infrastructure, reshoring, energy transition, and consumer discretionary pressure.

Data Collection Process

We query SEC EDGAR for quarterly filings, extract XBRL-tagged financial data, calculate quarter-over-quarter and year-over-year growth rates, and classify each company's signal based on the quantitative and qualitative factors described above. Signal classification is updated after each earnings season. Theme membership is reviewed and adjusted at the same cadence.

Update Frequency

Earnings season occurs four times per year (January, April, July, October). We update company data as each company reports, with the bulk of S&P 500 companies reporting within 4-6 weeks of quarter end. Theme analysis is refreshed at the end of each earnings season. The page footer shows the most recent data refresh date.

Known Limitations

Signal classification involves judgment calls, particularly for the MIXED category. Different analysts may interpret the same earnings differently.
Coverage is limited to S&P 500 companies. Smaller public companies and private companies are not included.
Earnings calls are backward-looking, they describe what happened last quarter. Forward guidance is management's best estimate, not a guarantee.
This is not investment advice. Operator Signal Scores are designed for business operators evaluating market trends, not for stock trading decisions.

Frequently Asked Questions

What does TAILWIND mean in EarningsCallAI?

TAILWIND means revenue is accelerating across recent quarters and management commentary trends positive on demand and pricing. For operators in adjacent markets, the read is constructive: budgets in this space are likely healthy and customer demand is expanding.

What does HEADWIND mean?

HEADWIND means revenue is decelerating or contracting and management commentary trends cautious on demand or pricing. Operators selling into the same space should expect longer sales cycles and tighter buyer budgets, defensive moves are typically warranted.

What does the confidence level (HIGH/MEDIUM/LOW) mean?

Confidence reflects whether the quantitative trend (revenue growth direction) and the qualitative commentary (CEO/CFO language) agree. When both inputs point the same way, confidence is HIGH. When they diverge, confidence drops to MEDIUM or LOW. LOW-confidence signals warrant deeper segment-level reading.

How current is the data?

Earnings season runs four times a year (January, April, July, October). EarningsCallAI updates company data as each company reports, with the bulk of S&P 500 companies reporting within four to six weeks of quarter end. Theme analysis is refreshed at the end of each earnings season.

Is this investment advice?

No. EarningsCallAI is built for business operators evaluating market trends and competitive dynamics, not for stock trading decisions. Operator signals are not buy/sell recommendations. Always consult a licensed financial advisor before making investment decisions.

How can I cite EarningsCallAI?

Cite the specific URL and access date. Underlying SEC EDGAR data is in the public domain and can be re-cited directly to the SEC. Operator signal classifications are EarningsCallAI methodology and should be attributed to this site if quoted.

How to Cite This Data

If you use data from EarningsCallAI, please cite:

EarningsCallAI. "[Company Name] Earnings Analysis." earningscallai.com, 2026. Accessed [date].

Underlying financial data is sourced from SEC EDGAR XBRL filings and is in the public domain. Cite SEC filings directly when quoting raw financials. The Operator Signal classification is EarningsCallAI methodology, attribute to this site when quoted.

EarningsCallAI Methodology