By Dylan Walsh
In 2013, Sheryl Sandberg made reference to what “may well be the most important document ever to come out of [Silicon] Valley.” It was not a printout of revolutionary code, not a contract outlining the merger of two giants, not even something related to Facebook, where she was (and is) the chief operating officer. Sandberg was referring instead to an unadorned deck of 127 PowerPoint slides, titled “Netflix Culture: Freedom & Responsibility.”
Organizational culture is increasingly seen as a key contributor to a company’s success: Leaders should articulate the cultural principles that define their work, then use this vision to guide actions. And often placed at the vanguard of building and maintaining culture are those doing the hiring. It’s up to them to find like minds.
“Both academics and practitioners have long thought of cultural matching as a process that should happen at the point of entry — some people fit, some don’t, and both employers and employees should look for matches,” says Amir Goldberg, an associate professor of organizational behaviour at Stanford Graduate School of Business. “But our research suggests another ingredient, or dimension, that’s overlooked.”
And that, he says, is adaptability.
People display not only varying degrees of cultural fit, but also varying degrees of adaptability, and this second feature, Goldberg argues, is one that companies should probe more deeply.
In a new paper, he and three co-authors — Sameer Srivastava of the University of California, Berkeley, and Govind Manian and Christopher Potts of Stanford — gathered more than 10 million internal emails from a technology firm sent between 2009 and 2014. They used linguistic analysis to monitor cultural fit over time among employees. (Language use is intrinsically related to how individuals fit, or fail to fit, within social environments.) Individuals were measured against those with whom they had the most frequent communication.
While an employee’s cultural fit at the time of entry was loosely connected with outcomes — those who fit well from the outset tended to perform well — a much more powerful predictor of success was an employee’s ability to recognize and internalize standards. “We find that what predicts who will stay, who will leave, and who will be fired is not so much initial level of cultural fit as much as their trajectory, the degree to which they adapt,” Goldberg says. “There are important differences between individuals insofar as they are capable of reading cultural code and shifting behaviours accordingly.” The authors refer to this malleability as “enculturability.”
For human resources departments, then, hiring questions should perhaps revolve around how adaptable people are as much as they do around how much their beliefs align with the company’s beliefs. Have applicants lived in other countries or environments? Have they readily moved between multiple and varied work environments? Have they smoothly adapted to each of these environments?
But the value of this insight spills beyond the single act of hiring and into the realm of employee retention. Goldberg and his colleagues parsed the data on employees who left the firm to explore who went voluntarily and who was asked to leave. They found that employees who struggled with enculturation from the outset were often fired; these were classic cases of cultural mismatch. But a second group started out as mismatches, quickly learned to fit in, and then, over time, their attachment to the firm began to weaken. Ultimately, this drifting interest often led to volitional departure.
Both academics and practitioners have long thought of cultural matching as a process that should happen at the point of entry — some people fit, some don’t, and both employers and employees should look for matches. But our research suggests another ingredient, or dimension, that’s overlooked.
For Goldberg, this presents an opportunity. It’s possible that language diagnostics — and, in the future, other tools, like observing body language or how people dress — could help leaders “keep a finger on the cultural pulse of their organization,” he says. Companies invest significant sums in hiring the right people and providing orientations or “onboarding,” but cultural integration often stops soon after that.
“We don’t currently have the right tools” to do more, he says. Engagement surveys are costly and infrequent, and besides, people lie to themselves or misunderstand questions, and the resulting data is questionable, at best. “But can we use these linguistic tools to see which parts of the organization are functioning harmoniously or discordantly from a cultural point of view?” Goldberg wonders. “Or maybe the tools can even provide individuals clearer insight about themselves, about how they’re doing with regard to the culture.”
Goldberg’s excitement over these possibilities — the potential for more effective hiring screens, the ways in which companies might monitor cultural health — is tempered by the prospect of misuse.
While his research has generated an interesting twist on conventional wisdom, it isn’t definitive. “I wouldn’t want the fate of someone’s employment based on what is only a burgeoning research program,” he says. “It’s important to communicate that point because questions about hiring and firing are critical questions, and we want them to be answered fairly.”
In time, he believes, this research will mature in a way that can better avoid false positives or negatives, and that will valuably supplement the conventional work of recruiters and managers as they undertake employee assessments. Goldberg sees a future in which diagnostic tools that analyze language, or other expressive features, can cheaply and accurately predict a person’s cultural openness. But unanswered questions, both technical and ethical, stand between today and that future.
“We want to make the running of organizations more efficient and cost effective but without creating a Blade Runner kind of world,” he says.