14 min read
New: see the most important themes in transcripts instantly
June 5, 2020
3 min read
Call transcripts are one of the most powerful unstructured datasets to analyze for signals on the state of companies’ strategies and outlooks. But with the average earnings transcript running over 2,500 words, it’s time consuming to parse each one for the most salient information. Whether you’re eyeing a competitor, looking for an inflection point for a portfolio company, or preparing an earnings script, capturing the most meaningful information from a large volume of transcripts is critical, yet often an inefficient time suck.
Good news: AlphaSense now automatically extracts key themes from transcripts in seconds as transcripts appear in the platform and presents them in a dynamic, sortable format to connect you to the most important concepts within the transcript as they load in the platform. With powerful topic clustering and the added context of metrics like number of mentions and sentiment, you’ll drill straight into the most important details.
Dynamic, NLP-driven theme extraction is live in AlphaSense across all transcripts.
How it works:
Now, clicking into any transcript in AlphaSense activates an Insights panel in the center of the screen.
Within the Insights panel, the document topics module shows the key themes extracted from that transcript. AlphaSense has a unique ability to parse and understand financial language, generating a list of key themes.
By default, the topics will be listed by ‘score’, our measurement of overall relevance. You can sort themes by the metric that you are most interested in, such as:
- Overall number of mentions of a topic
- Negative sentiment
- Positive sentiment
- Increase/decrease number of mentions since last quarter (available for earnings transcripts only)
Further adding to the relevance of theme extraction, topic clustering ensures that related concepts are captured under the umbrella of a single theme, making it easier for users to understand the various factors contributing to a single theme.
For example, inside of the Production theme on a recent Nissan transcript, we see terms like ‘manufacturing expenses’ and ‘production capacity’ are among the many concepts associated with that theme.
Why it matters:
Theme extraction is the latest expression of our best-in-class tech stack that drives our powerful synonym recognition (NLP), sentiment analysis, and provides our users with ever-increasing robust and granular analysis into text-based documents.
If you’re a current client/trial user and ready to dig into Themes, you can
- Log in to get started
- Check out our Help Center for more tips
- See theme extraction analysis applied to recent CPG earnings transcripts
Not a current user? Start your trial now.
More like this
10 min read
5 min read