RDEL #7: How does language signal cultural fit?
A look at how an engineering team's shared language can offer clues for cultural fit.
Welcome back to RDEL! Each Monday, we pose an interesting topic in engineering leadership, and apply the latest research in the field to drive to an answer.
This week, we look at an interesting aspect of culture in an engineering team - the team’s shared language. How does language predict the success and tenure of teammates in engineering?
The context:
Organizational researchers have long-acknowledged the importance of teammates culturally fitting in with their team - achieving cultural fit leads to both higher performance and longer tenure on teams. Culture, and an individual’s fit to that culture, evolves over time. This means that teammates can achieve cultural fit and lose it multiple times during their tenure (for example, during periods of organizational change).
It turns out that language is a great signal to model cultural fit, as it is less susceptible to reporting bias and more granular to measure. The question we ask this week is - how does language signal cultural fit, and how do we apply this to engineering teams?
The research:
A team of researchers at Berkeley and Stanford developed a language-based model of cultural fit using over 10 million emails at a mid-sized technology firm. They mapped each email to a set of semantic categories using the Linguistic Inquiry and Word County (LIWC) lexicon, and created distributions of word usage to determine the “shared language” of the team. They then used demographic information of the 601 employees that sent these emails over a five-year period to map language to outcomes.
The key insights the researchers found were:
Cultural fit has the highest positive slope in the first few months of a company as a new hire is ramping up.
Cultural fit never reaching net positive is a strong predictor for involuntary departure. The researchers interpret this lack of cultural adaptability is likely due to negative reception by colleagues.
Cultural fit can drop later in a career, and can be a strong leading signal for a voluntary departure. In this case, researchers speculate that it is an indication of a decline of an individual’s attachment to the organization.
In summary, cultural fit is a continual process, not an end state.
The application:
Using language as a primary measurement tool opens up many unique opportunities for engineering leaders to catch signals of changes in cultural adaptation.
A couple of examples include:
Changing remote communication: In remote teams, it’s often hard to gauge team engagement. If an engineer is disengaging with the team’s communication channels (ie via slack), this could be an early signal of a cultural fit change. That would be a good opportunity to connect 1:1 and understand how to better support the employee - retention is always significantly less costly than replacement.
Onboarding: If a new engineer has not adapted the shared “language” of the team, this could be a signal of a challenging ramp-up experience. In this case, it would be important to understand whether the individual or process is contributing more to that lack of adaptation. In both cases, stepping in early to offer support will increase their odds of cultural adaptation.
In summary, this paper shows that cultural adaptation is not a linear journey. Culture is something that evolves with both the team and broader organization, and language is a useful tool to signal changes in cultural fit on the team.
Thanks for reading, and happy Research Monday!
Lizzie
From the Quotient team