It’s an exciting time to be a people practitioner: it seems like every day a new technology emerges promising to make internal processes more convenient, efficient, and data-driven.
While there is definitely value being added by these tools, there is just as much value being left on the table. Take engagement survey platforms for example: yes, they have greatly facilitated the collection of feedback data, but the translation of those data into informed decisions remains a significant challenge.
Text data in particular, such as open comments in engagement surveys, are notoriously under-utilized. A major reason this valuable information goes untapped is that it is unstructured in nature: text data do not match a pre-defined format, and therefore require transformation to be usable. That transformation often requires both time and specific technical expertise that many people practitioners simply don’t have, which motivates a focus on data that are already structured.
The problem is that, while only using structured data will indeed tell a story, that story will be fundamentally incomplete. It can be likened to reading a copy of Moby Dick with the final third of pages ripped out: sure, it will be a story, but will the conclusions drawn be the same as if one had read the whole book? Probably not. And that’s precarious in an organizational context, where the conclusions decision-makers draw have resounding effects on the company.
We once worked with a large client in the technology sector who was having trouble with employee rewards. In the numeric data from their engagement survey, the People team found employees were highly dissatisfied with rewards in the company. With the best of intentions, they interpreted this as employees saying that the rewards weren’t high enough, and thus decided to increase the value of all performance rewards by 30%.
To the People team’s great surprise, employees were livid after this change. Why? As the text data would have revealed, most employees didn’t feel that the amounts for the awards were unfair, but rather that the process through which they were distributed was. Increasing the value only exacerbated the problem, and made workers feel ignored.
This example illustrates that making decisions based solely on numeric (or structured) data, without the fuller picture provided by text (or unstructured) data, can lead one’s practice astray. Still, even if it’s best to leverage text data in an People Ops context, what about the temporal and technical limitations of doing so?
Click here to continue reading Shape’s article.