Skip to main content

Earnings Updates & Sentiment Scores

explanation on how we're doing it

Admin avatar
Written by Admin
Updated this week

Objective

The goal of this sentiment scoring method is to translate complex financial language into a simple, standardized signal that captures how a company is performing across five key areas. The AI acts like a trained analyst, interpreting the transcript through a structured lens that focuses on: ​

  • Direction of Change: Is performance getting better, worse, or staying the same compared to the previous year? ​

  • Strength of Signal: Are the comments and figures clearly pointing in one direction, or are they mixed or uncertain? ​

  • Evidence-Based Reasoning: Only assigns a score if there’s supporting information—either in the form of concrete YoY data (like “+15% revenue”) or strong directional commentary from management(like “we’re raising our sales forecast”). ​

This approach removes subjectivity and vague language, so that each score:

  • Reflects how a typical analyst would interpret the section ​

  • Can be compared across companies and time periods ​

  • Is always backed by a brief, understandable justification ​​

Scoring: Positive Green circle

  • Clear YoY improvement in financial or strategic performance ​

  • Optimistic tone backed by data or concrete management statements ​

  • Guidance or commentary signals confidence and momentum ​

    Electric light bulb Example: “Margins improved 3pp YoY driven by operational efficiencies and pricing power”

Scoring: Neutral / Mixed Yellow circle

  • Performance is flat, mixed, or inconclusive ​

  • No clear guidance or directional signal is provided ​

  • Positive and negative elements offset one another ​

    Electric light bulb Example: “Revenue was flat YoY; price increases were offset by lower volumes”

Scoring: Negative Red circle

  • Performance deteriorated year-over-year (e.g., drop inrevenue, margins, or outlook) ​

  • Management expresses concerns, flags risks, or signals worsening trends ​

  • Language includes terms like “declined”, “under pressure”, “weaker”

Electric light bulb Example: “Sales declined 12% YoY due to reduced demand in key markets”

How AI applies this logic

  • It relies only on the earnings transcript content. ​

  • Evaluates year-over-year comparisons and direct statements from leadership.​

  • Requires either data (e.g., “+18% net sales”) or strong directional language (“we’re raising guidance”).​

  • Uses a topic-specific lens (e.g., outlook is judged only on forward-looking commentary).

Topic covered

  • Revenue development
    How sales changed versus prior periods (e.g., year-over-year and quarter-over-quarter), plus what drove it: price/mix, volumes, new products, acquisitions/divestments, and currency effects.

  • Profit development
    How profitability moved (operating profit/EBIT/EBITDA, net income) and why—cost inflation/deflation, operating leverage, productivity, pricing, one-offs.

  • Market conditions
    The external backdrop affecting results: demand trends, customer/end-market health, competitive dynamics, input costs, supply chain, regulation, and macro indicators.

  • Revenue outlook
    Management’s view on future sales (often guidance): expected growth rates or ranges, key assumptions (demand, pricing, FX), and risks/upside factors.

  • Profit outlook
    Management’s expectations for margins and earnings (EBIT/EBITDA/net income): target ranges, cost and efficiency plans, investment impacts, and major sensitivities.

Did this answer your question?