Add the Power of Expert Insights to Your Research. Learn More

UPS: Negative NLP Sentiment, ML Highlights Unfavorable Mix Amid the Covid Crisis; China Growing in March

For the next few days, we will review Q1 calls across a variety of industries using our machine learning and natural language call transcript application, Smart SummaryTM. Today we will focus on UPS.

We released the first version of Smart Summary last July to rave reviews. This was followed up by an in-depth expansion of many of the machine learning classifications and improved visuals (word clouds, heat maps, and keyword extraction) in our March 2020 Sentieo 4.1 product release. 

Smart Summary fundamentally changes the way our clients read transcripts. While it’s important to listen to high priority calls live, Smart Summary ensures that you do not miss anything in your high priority list, whether you listen live, or not. Further, with Smart Summary you no longer need to read transcripts from your broader coverage universe chronologically. 

Today we will look at the most recent UPS transcript (Q1 2020 was reported on April 28, 2020). We can see that the Smart Summary email came in shortly after the “old way” transcript (don’t worry, we have the full transcript in the Smart Summary too, with our NLP sentiment scoring applied).

The Smart Summary PDF starts off with the word cloud through NLP keyword surfacing on top of the new tearsheet. 

We can see that Covid-related terms are prominently featured, as is Asia (recovering!), recession, and healthcare (big increase in mentions there due to a number of corporate initiatives). We can also see Headwinds and B2C: this immediately alerts us to the product mix headwinds that the company is experiencing (shift to the less profitable B2C). 

The overall sentiment of the call was decidedly negative versus prior quarters. 

We can also see this reflected in the darker color in the ML-based Categories Heatmap. Notice the particularly negative change in the Economic category, as well as the Deflection category, indicating considerable uncertainty. 

The top keywords in the transcript (along with the top keywords in the prior two years) are next: these mirror the word cloud. We can see the mix mentioned above (SMB and B2C), we can see Healthcare, Asia, and of course, Covid.

In the next section, we highlight several categories from the ML output, along with their sentiment scoring. The top category is Covid. In the top five Covid comments, we can immediately see the guidance withdrawal and the unfavorable mix shift language.

In the next category, Business, we see a mix of positive and negative in the top five. While ROIC and the initiatives are positive, we see the decline in U.S. operating income and mix shift (again!) as negatives. 

In the video below, we do a more detailed walkthrough of the emailed Smart Summary along with the full version inside our product.

Stay tuned for more highlights over the next week, and register below for our Q1 2020 Earnings Season Recap webinar on Thursday, May 7.

If you would like to learn more about Sentieo’s ML and NLP capabilities, please get in touch.