Using NLP to Accurately Measure Employee Engagement During COVID-19
Updated: Sep 14, 2020
I was interested to read the recent Harvard Business Review article on Microsoft’s analysis of changing communication patterns as COVID causes world-wide changes to the way that we all work and collaborate. The move to remote WFH for a large portion of the workforce has been abrupt and has the appearance of being a long-term or even permanent change for many companies. As organizations strive to understand the effect on their business and employees, it is important that they have the ability to accurately measure the impact on employee engagement. A lot has changed for employees; gone is office comradery and the informal interchange of ideas, whiteboard brainstorming, and team interactions in the break room, at lunch or even chance meetings in the corridor. Gone is the all too important delineation between work and home life, with the danger that work is now 24/7. And compounding matters further are the potentially disproportionate impacts on female employees caused by the closure of schools and day care facilities.
While message volume and other meta-data can provide interesting high-level insights about employee behavior, in order to accurately understand how individuals and groups are adapting to this new work paradigm, it is critical to dive deeper. Without understanding the content of the communications, it is impossible to truly understand what is occurring. For instance, are employees fearful and distracted about the possibility of layoffs? Are they frustrated or even angry about the inefficiencies introduced by WFH? Are employees confident in the ability of their executives to manage through this change? Are there groups or job roles that are finding the transition more challenging? Are managers able to effectively support and foster DEI initiatives, collaboration and inclusiveness across their team when in-person 1:1 and team meetings are no longer feasible.
In addition to analyzing the wide range of communications meta-data that is readily available to companies, RSquared uses deep Natural Language Processing (NLP) to analyze the text of these communications. By understanding the sentiments being expressed by employees, RSquared is able to provide significant additional fidelity to a company’s understanding of their workforce, down to the level of individual users and groups. The models are trained to ‘understand’ the often-subtle nuances in employee email and IM conversations, providing detailed insights into the struggles employees face in this new environment. These advanced analytics also provide leadership with actionable insights for how they can help and support their most valuable asset. In addition to providing an understanding of sentiment, RSquared uses NLP to provide insights into user and group collaboration and inclusion, allowing a deeper understanding of the efficiency of the organizations and the efficacy of any changes and initiatives being introduced to improve the remote work experience for employees.