Balto Workforce Survey


At Balto, I was part of the Conversation Excellence Lab, which conducts academic research to generate insights for contact centers.

For this project, we partnered with Centiment to survey over 2,000 contact center agents in order to find out more about the typical agent profile and experience. We asked a number of basic demographic questions, as well as more in-depth questions related to their job satisfaction and enjoyment, their career aspirations, their motivations for joining the industry, and much more.

One of the reasons for conducting this survey was that at the time of this research, America was in the midst of the “Great Resignation”, when many employees were quitting in droves. Many of our customers in the contact center industry were hit hard by these resignations, so we performed this analysis to help them reduce attrition by giving them insight into their workers’ motivations.


Here is a link to a subset of the raw data, as well as all of the analyses performed in individual sheets. I used Google Sheets for this project. Each sheet measures the effect of a different independent variable (such as gender, education level, or salary) on a wide variety of dependent variables (such as satisfaction with pay or where they see themselves in 5 years).

Some variables were used as both an independent variable and a dependent variable. For example, we asked about the agent’s fear about automation/AI replacing them in the workforce- this fear of automation was examined as both a predictor of job enjoyment (in which case fear of automation was the independent variable) and as a result of demographic factors (in which case it was the dependent variable). Many other variables were treated similarly.


The deliverable of this analysis was a series of articles, published on our website, which summarized our findings and gave suggestions to contact center managers about how to apply them. I did not write these articles, but I did perform all the analyses and create all of the graphs found in the articles using a JavaScript library called Chart.js.

All of the published articles are linked below: