Skip to content
Select Your Location
View available investments and insights in your market
Europe and the Middle East
Asia Pacific
Data Science At Work

“Why do I not want to hear the words ‘very optimistic’ on an earnings call?”

Data Source Use Case: Conference Call Transcripts and Filings
Hypothesis Categories: Competitive Threat, Customer Mix, Customer Loyalty

Key Data Source
Key Data Source
Conference call transcripts and filings
  • Earnings and acquisition announcement calls
  • Monthly update calls (e.g. sales)
  • Company conference presentations
  • 1,500+ U.S. public companies from 2010 to present
  • All corporate SEC filings
What Were We Looking For?
What Were We Looking For?
Indications of management sentiment from earnings call transcripts
Light Bulb 
What Did We Learn?
What Did We Learn?
The AT&T conference call for Q2 2019 revealed genuine optimism from management, which was justified by subsequent company performance.

At Neuberger Berman we have developed a natural language processing model that can identify more than 5,000 phrases of two or more words, which, when they appear in the transcripts of quarterly earnings calls, it considers to carry positive or negative implications about the company.

Sometimes it is intuitive that these phrases should be construed as positive or negative. Let’s take a few “bigrams,” or two-word combinations, as examples.

It will come as no surprise that “raise guidance” is seen as positive and “reduce guidance” is seen as negative. Similarly, “ahead of schedule” is better than “slower than.” We would rather see “repurchase shares” than “equity offering.”

However, is “very optimistic” a good or a bad sign? Correlation between this bigram and subsequent stock performance, picked up from thousands upon thousands of transcripts, indicates that it is a bad sign, on average. Management tends to say it when results have been poor and they want to persuade investors that a turnaround is on the way. Bigrams such as “closer look,” “briefly review” or “quick update” may seem completely neutral, but analysis shows that the first tends to be positive and the other two negative—management likes to get into details that reflect well on the business and skirt over the problems.


Extract from the AT&T Inc., Q2 2019 earnings call, July 24, 2019

Highlighted Text 

Source: S&P Capital IQ. For illustrative purposes only.


The earnings call extract above shows our model in action, picking out positive bigrams in green and negative bigrams in red.

Again, there are intuitive findings: “paid off,” “expanding margins” and “EBITDA growth” all tend to be positive. And there are apparently neutral phrases that contain hidden information: “will get” and “other operating” are interpreted negatively because they are associated with the non-specific language management uses to gloss over challenges, whereas “segment operating” is associated with a speaker who is laying out the details of how different parts of the business have been performing; a bigram such as “straight quarter” is a positive sign because few CEOs like to boast about their “third straight quarter of declining margins.”

One can immediately see the preponderance of green over red bigrams in this extract. Our model classified this transcript in the top 5% of more than 800 that it processed in July.

These examples are a good reminder that alternative data do not always come in the form of numbers, and that the techniques applied to interpret the data are as important as the data themselves.

White Paper
The Data Science Revolution
Download White Paper
More Big Data Case Studies
Credit/Debit Card and Bank Account Transaction Data
“Are men really going to buy clothes from Lululemon?"
Online Transaction Data
“Are Amazon’s new batteries eating into Energizer’s market?”
Internet Content and Search Data
“Are employees buying into the big changes at Procter & Gamble?”
Online Job Post Data
“How can I tell if companies are investing in the post-capex era?”
Big Data
The Data Science Revolution
How the new richness and accessibility of data, and advances in data science, are enhancing both quantitative and traditional fundamental investment research—and sparking a revolution in active management.
Learn More