Case Study — 10 Jul, 2023
An Economics Professor Unlocks New Insights from Earnings Calls
Highlights
THE CLIENT: An Asia-based university
USERS: Professor of economic management
NLP has revolutionized the speed at which textual data can be analyzed and linked to other datasets to uncover new insights, making it a ‘must have’ for this professor to take his research to a new level.
Public companies share their financial information each quarter in their earnings calls that also provide an opportunity for executive teams to discuss their plans and answer questions from analysts, investors and media personnel. By transcribing these earnings calls, these earnings calls transcripts contain highly informative content for investors and researchers, but the unstructured nature of the calls makes researching these events difficult. The advent of machine-learning (ML) capabilities, such as natural language programming (NLP), is making it possible to easily assess large volumes of unstructured textual data to uncover new insights. In addition, NLP can help users of transcripts dissect the tone, complexity and overall level of engagement with analysts as indicators of earnings sentiment.
A professor of economic management at a large university in Asia wanted to harness unstructured data from corporate earnings calls to support his research into whether these transcripts provided additional stock selection power. He wanted access to an offering that used NLP in order to maximize the number of transcripts he could evaluate and perform research as efficient as possible.
Pain Points
The professor had been using a third-party data provider to access earnings call transcripts, but the capabilities were limited. The former transcript data source required extensive manual work to extract datapoints and try to integrate them with other datasets. He wanted a more powerful solution that could:
- Aggregate data from earnings calls and deliver this in a machine-readable format for NLP applications with metadata tagging.
- Leverage the information to keep track of events for specific companies, including dates, times, dial-in and replay numbers and investor relations contact information.
- Easily combine data from earnings, company conference presentations and special calls with more traditional datasets to develop proprietary analytics.
- Quickly identify the rationale of sell-side analysts' revisions.
- The professor was a current user of S&P Global Market Intelligence's ("Market Intelligence") Capital IQ Pro desktop platform to access and analyze several datasets. He reached out to learn more about what the company offered with respect to transcripts.
The Solution
The Machine Readable Transcripts offering that would provide the professor with the ability to:
|
Easily analyze earnings calls and more |
Users can access textual transcript data in a machine-readable format with metadata tagging. Historical data goes back to 2004, and coverage includes 10,600+ companies (and growing), with 100% coverage of the S&P 500, Russell 1000 and FTSE 100, plus 95% coverage of the S&P Euro 350.[1] |
|
Quickly link information with traditional datasets |
Users can easily combine transcripts data from earnings, M&A, guidance, shareholder, company conference presentations and special calls with traditional datasets to formulate unique, proprietary analytics. They can seamlessly link the Speaker ID to the S&P Capital IQ Estimates and Professionals databases to help identify sell-side analysts’ revisions.
|
|
Keep track of event information for specific companies |
Users can sync with events data either via feed or through streaming XML messages for details and alerts on upcoming calls scheduled for transcription. This includes dates, times, dial-in and replay numbers and investor relations contact information. |
|
Access sentiment and behavioral-based metrics |
The Textual Data Analytics (TDA) dataset takes earnings call transcripts one step further with sentiment and behavioral-based metrics rigorously researched and tested against frequently used quantitative strategies. Users can implement signals from 800+ predictive and descriptive metrics derived from NLP in combination with data from the Professionals and Estimates datasets, with differentiated and additive characteristics. |
Key Benefits
Machine Readable Transcripts addressed all the professor's requirements, and he added the offering to his ongoing subscription. He is now able to:
- Eliminate the previous manual processes and quickly and efficiently evaluate a wide range of transcripts for companies of interest, including 11,600+ entities with history going back up to 2004.
- Rely on a full history of data, with the accuracy and timeliness of a trusted provider.
- Quickly link data to the company ID and professional ID to embellish the available information and uncover additional relationships and insights.
- Easily keep track of upcoming events so as not to miss important announcements.
Click here to learn more about S&P Global Market Intelligence's Machine Readable Transcripts.
[1] Coverage numbers as of December 2022.