Redefining Employee Satisfaction Surveys in the Age of GPT
Technological ignorance can manifest in two distinct ways. The first involves underestimating the far-reaching impact of innovative breakthroughs on our daily lives. The second, laced with a touch of irony, arises when we recognize the significance of a groundbreaking invention but fail to capitalize on its full potential. This is because our focus remains narrowly centered on addressing issues that are made obsolete by this very invention. I have recently encountered an intriguing example of this latter form of ignorance: the proposal to employ GPT for the creation of employee satisfaction surveys featuring closed-ended questions.
To understand why this approach is illogical, we must explore the historical origins of closed-ended questionnaires, which date back to 1838. The development of these questionnaires was propelled by the lack of techniques for automatically interpreting text. As a result, closed-ended questions with a finite set of response options, such as "Yes/No" or a numerical scale, enabled the transformation of unstructured data into a quantifiable format. Regrettably, this technique is not without its flaws, one of which is commonly referred to as survey fatigue. Put simply, the more information we attempt to get from respondents via surveys, the less we actually acquire. This paradox stems from the fact that to cover all conceivable scenarios, a questionnaire must, by its very nature, encompass each of them. As a result, the survey becomes overly burdensome for respondents, who often disengage and resort to autopilot mode midway through. This problem is particularly acute in longitudinal studies, wherein identical questionnaires are administered repeatedly to the same participants at regular intervals.
Fortunately, the advent of GPT has rendered traditional employee satisfaction surveys obsolete. We now possess the capability to gather employee feedback in the form of unstructured, free-flowing text. Rather than posing questions directly to individuals, we can analyze their textual expressions to extract the insights we desire, thereby revolutionizing the manner in which we obtain and interpret employee viewpoints. This can be thought of as surveying the text itself, rather than the humans behind it.
Some observant readers may express reservations: "We need to inquire about crucial topics, or employees won't address them." I respectfully challenge this notion. We are not conducting a survey on whether the office should invest in an espresso machine or an orange juice extractor. Instead, we are collecting information on employee satisfaction, and it is the employee's prerogative to determine if an issue is significant enough to discuss. Moreover, this "forgetful employee" mentality prevalent in HR often leads to surveys starting with questions like "Do you believe your salary is competitive?" which predictably elicits negative responses from a vast majority of employees. Conversely, by allowing staff to articulate their concerns unprompted, we may witness a decline in this percentage.
HR should be more concerned about omitting a topic in a survey that is important to employees but unknown to HR. The conventional method of closed-ended questions can accidentally lead to overlooking legitimate issues. After all, if there's no question, there's no problem. In contrast, open-ended questions coupled with GPT-based analysis yield a more comprehensive understanding of a company's strengths and weaknesses, including problems that may have escaped HR's notice but are brought up by employees.
My particular interest in employee satisfaction feedback stems from Culturama, a system I had a hand in developing that embraces a contemporary, GPT-based approach. We pose 2-3 open-ended questions and then employ GPT to analyze responses, extracting insights on the company's strengths and weaknesses as well as recommendations for improvement. By doing so, Culturama offers a more sophisticated and insightful method for obtaining employee perspectives on the organization.
As we continue to navigate a rapidly changing workplace landscape, it's crucial that we adapt our methods for gathering feedback to ensure that we are truly capturing the needs and concerns of our employees. GPT renders traditional employee satisfaction surveys with closed-ended questions obsolete, thereby eliminating the necessity to automate their creation.