Unlocking Hidden Insights In Employee Comments

More HR professionals would rather mow the lawn than analyze employee survey comments. This is anecdotal, and I have no statistics to support this claim, but I have seen a lot of new leaf blowers is my neighbourhood. In fact, open-end survey questions are often avoided because free-form responses can be challenging to analyze and interpret. And we have leaves to blow. Although there is truth to this statement, it’s not insurmountable, and in this course, Unlocking Hidden Insights in Employee Comments, I will guide HR professionals through the essential skills needed to analyze open-end survey questions. 

We start with a scenario; a change to employee benefits and a sample of responses to that change. Using Excel, I’ll walk you through foundational text analysis techniques, including document-term matrices and keyword extraction with add-ins like MeaningCloud. As the course progresses, you'll gain skills in sentiment analysis, cosine similarity, and more—all designed to reveal deeper insights from employee comments.

By the end of this course, you’ll be equipped to handle open-ended feedback confidently in Excel, transforming raw text into actionable insights. We’ll even quantify our findings with a little introduction to Bayesian probability.

Course Modules:

Word Frequencies in Excel: In this module, we’ll start by calculating word frequencies across the 50 employee comments, providing an overview of the most common terms related to the benefits change. You’ll learn how to use Excel’s basic functions to build a word frequency table, highlighting recurring themes and concerns.

Calculating Cosine Similarity: Using Excel, we’ll calculate cosine similarity to identify comments that share similar language and topics. This technique will help us cluster comments with similar sentiments or viewpoints, revealing patterns in employee reactions to the benefits update.

Term Frequency/Inverse Document Frequency (TF/IDF): Explore how TF-IDF in Excel highlights words that are significant yet unique across different comments. You’ll learn how this technique moves beyond frequency to emphasize words with specific importance, helping to indentify less frequent but highly relevant terms in employee feedback.

MeaningCloud and Sentiment Add-ins: We’ll load the MeaningCloud and Sentiment add-ins to quickly categorize employee comments as positive, neutral, or negative. This automated sentiment analysis provides an initial layer of interpretation, helping you quantify the emotional responses to the benefits change.

MeaningCloud Extract Topics: In this module, we’ll use MeaningCloud to extract key topics from the employee comments, offering a high-level view of the main issues and themes. This will help isolate specific points of interest, such as concerns about prescription coverage or praise for new vision benefits.

MeaningCloud Cluster Texts: Using MeaningCloud’s clustering feature, we’ll group comments with similar topics or sentiments, making it easier to analyze common themes. This clustering approach allows us to identify distinct categories within employee feedback, enabling more organized and targeted insights.

Summarizing the Findings: Finally, we’ll bring together all of our findings in a clear, structured summary. You’ll learn how to create a concise report in Excel, effectively communicating the key themes, sentiments, and insights from the employee comments on the benefits change—and with the assistance of an interative generative AI approach.