Loading…
Loading…
Tellvest turns earnings calls into structured, comparable intelligence: sentiment scores, analyst dynamics, macro signals, and returns data for every company, every quarter.
Every quarter, public company executives get on earnings calls and talk to analysts about their business: results, outlook, strategy, and risks. These calls contain signals about individual companies, entire sectors, and the broader economy. But there are thousands of them, and reading even a fraction isn't realistic.
Tellvest reads every call and extracts three layers of structured intelligence: management sentiment (a 1-10 score for every quarter), analyst dynamics (who pushes hardest and where management deflects), and macro signals (what hundreds of CEOs are reporting about the same themes). The result is a comparable view across companies, sectors, and time.
You can browse all 524 companies, explore the economy and sector dashboards, see how score consistency maps to returns, dig into analyst and firm profiles, or check the earnings calendar to see when the next scores drop.
Tellvest analyzes the full earnings call transcript each quarter to produce a single Sentiment Score (1–10) capturing the overall tone of the call. The score is not opinion — it's built from a transparent rubric anchored on the company's reported GAAP financials, then adjusted for what management actually said.
Every score is the sum of four components:
base score + transcript adjustment + EPS adjustment + guidance adjustment
We start from year-over-year GAAP revenue growth (sourced from SEC filings) and look up the base score in a fixed table:
| Revenue YoY | Base score |
|---|---|
| ≥ 15% | 8 |
| 8 – 14% | 7 |
| 3 – 7% | 6 |
| 0 – 2% | 5 |
| −1 to −5% | 4 |
| −5 to −15% | 3 |
| < −15% | 2 |
For most companies GAAP revenue is the right number. But for some sectors it isn't — and the rubric makes a narrow, deliberate exception. A ±1 Tier 1 adjustment can apply to any company when a single-quarter event (an impairment, an acquisition, an FX swing, a divestiture) makes GAAP revenue unrepresentative. A larger ±2 Tier 2 adjustment is reserved for sectors where GAAP revenue is structurally the wrong metric. That list is closed and maintained internally:
Examples include energy companies (where revenue moves with oil prices rather than operational execution), banks (where GAAP "total revenue" differs from managed net revenue), and P&C insurers (where GAAP EPS can swing hundreds of percent on catastrophe timing alone). In each case, the rubric substitutes the metric that analysts actually use to evaluate the business — net written premiums, operating income, FFO, core EPS, and so on.
The list of sectors eligible for Tier 2 adjustments only grows through manual review when a new structural pattern is identified. Outside of recognized sectors, no Tier 2 adjustment is applied — even when GAAP looks distorted to the eye. This prevents the rubric from being talked into a higher number on a case-by-case basis.
We compare GAAP EPS growth to revenue growth. If EPS outruns revenue by 5 percentage points or more, that's margin expansion and the score gets +1. If EPS trails revenue by 5pp or more, that's margin compression and the score gets −1. Inside ±5pp the EPS line is neutral.
When GAAP EPS is distorted by something below the operating line — an equity-method investment loss, a pension mark-to-market, a tax-rate swing, fresh acquisition interest expense — we cross-check against operating income or against management's own adjusted/core EPS figure and use that comparable instead. When this happens, the score carries an override flagso the reasoning is auditable: the GAAP spread, the adjusted spread, and the resulting adjustment all appear in the writeup.
Forward guidance is the most direct signal management gives. A raise of 3% or more at the midpoint of the guided range is +2; a smaller raise (or new annual-cycle guidance introduced for the first time) is +1; a maintain is neutral; a small cut is −1; a cut of 3% or more is −2; withdrawing guidance entirely is −3.
A few narrative conditions put a ceiling on the final score, regardless of what the waterfall adds up to. These rules exist to stop a strong transcript or guidance raise from papering over a weak underlying quarter:
| Condition | Max final score |
|---|---|
| Majority of themes are negative or mixed | 6 |
| GAAP revenue declined year-over-year | 7 |
| Forward guidance was lowered | 7 |
Caps are only shown on the score breakdown when they actually bind — that is, when the calculated waterfall total would have exceeded the cap. A quarter where revenue declined but the waterfall still landed at 5 has the cap recorded in the audit trail, but no “Hard cap” chip appears because the cap didn't change the outcome.
Each scenario below is a real scored quarter. Switch between tickers to watch the rubric walk from GAAP revenue all the way to the final score — including the cases where a Tier 2 sector override lifts the base, margin compression drags EPS into negative territory, or a hard cap pulls the calculated total back down.
Clean positive: strong revenue growth anchors base 8, margin expansion adds +1, fresh annual outlook adds +1.
Every earnings call contains signals about the broader economy — but they're buried in thousands of pages of transcripts. We extract macro-economic signals from each call across a standard set of categories: consumer spending, AI and technology investment, capital expenditure, supply chain conditions, trade and tariffs, inflation, employment, and more.
Each signal has a direction (improving, stable, deteriorating, or mixed). Aggregated across hundreds of companies, this creates a view of what corporate America is actually experiencing — not what economists are predicting, but what CEOs are reporting from the field.
The Economy dashboard shows the Tellvest Economic Sentiment Index: a composite score aggregated across all analyzed companies, alongside a macro signal heatmap showing which signals are improving or deteriorating and how broadly they're being discussed.
Sentiment scores and macro signals are aggregated by sector and industry. You can see which sectors are most or least optimistic, which macro signals dominate each sector, and where cross-sector divergences are emerging — for example, technology bullish on AI capital expenditure while industrials are cautious on trade and tariffs.
Sector and industry pages surface the same earnings metrics as individual companies but rolled up across all members, making it easier to distinguish company-specific signals from broader industry-wide conditions.
The Q&A section of earnings calls is where analysts push back on management and where the real signal often lives. We extract every analyst question from each call, classify it by topic using a standard taxonomy, and tag the tone as neutral or challenging. Each question links directly to the relevant moment in the full transcript, so you can read the exchange in context.
Every analyst gets a challenge rate: the percentage of their questions that push back on management's narrative, press for specifics on missed targets, or question forward guidance. The average across 46,936+ questions is roughly 8%, but the top challengers average over 35%. Analyst and firm profiles show challenge rates alongside coverage breadth, question trends over time, and the companies each analyst covers. Explore the full Firms & Analysts directory.
A single high sentiment score is noise. But when a company scores 8+ for two or more consecutive quarters, there's a pattern worth examining. The Returns page maps score consistency against actual post-earnings stock returns to see whether the signal holds up in practice.
Returns are measured from the T+2 open price: two trading days after the earnings call, which is the first price a user could realistically act on after seeing the score. This avoids the overnight gap and measures what an investor could actually capture.
The analysis also incorporates valuation risk. A company scoring 8+ but trading at an elevated P/E or EV/Sales multiple (relative to its sector) has less room for upside. The returns data shows that filtering for low or moderate valuation risk meaningfully improves the consistency of the signal.
This is not investment advice and past performance does not predict future results. The returns data is provided for research and informational purposes only.
Every metric on Tellvest is tracked quarter-over-quarter. Sparkline charts show how sentiment changes over time for each company. Economy and sector trend views show how macro signals shift across quarters — which signals are strengthening, which are fading, and where the turning points were.
Because public companies have different fiscal year-end dates, Tellvest normalizes all periods to calendar quarters. A company with a June fiscal year-end reports its fiscal Q2 in January — we map that to calendar Q1 based on the quarter-end date. This means when you compare companies within a calendar quarter on the dashboard, economy, or sector pages, you're comparing earnings calls that cover roughly the same real-world time period.
Earnings calls contain far more information than the headline numbers. Sentiment scores, economic signals, analyst focus areas, and key themes are all embedded in the transcript. But extracting and comparing them across hundreds of companies isn't feasible manually.
Tellvest does that work systematically. You get the score before you read the call. You see which analysts pushed and where management deflected. You spot the macro shift that 200 companies just confirmed. You compare any company to its sector in seconds. Whether you're tracking a single name or scanning the full market, the intelligence is structured and ready.
Tellvest is an independent research tool and is not affiliated with any company we track. Financial data is sourced from SEC filings and public financial databases. This is not investment advice.