How (Not) to Destroy a Data-Driven Culture
*Editor's note - After writing this I realize I am getting into more complex topics. I suggest reading the background links to get a better understanding of the theories I am referencing in the newsletters from now on. Including it all will create a book vs. a newsletter. :) - Andre
When we talk about being a data-driven organization, do we know what being ‘data-driven’ even means anymore? Buzz words collect different definitions and ten people can walk away with ten different assumptions when a manager uses one. So, let us start with my definition: A data-driven organization has a culture that is obsessed with learning via a scientific process to complement direct expertise in decision making. The next big question, as a leader, how do we know that you are getting it right? Leaders set the culture which you may have heard before. But “Culture” is another word we take for granted in business.
Culture is not a written statement that HR writes, it is a set of shared assumptions that social groups form based on their experiences of survival (Edgar Schein). Leaders set the culture because your staff interprets every one of your reactions as data. Our brains add this data, event by event, and cultures spawn from this as the organization learns from these reactions collectively. Culture, believe it or not, is experienced emotionally. If you are afraid to speak the truth, that feeling of fear is the culture of the organization you are in.
Think about your approach when you are ready to show off your new models, dashboards, and metrics you're using that counter the established narrative. Being the first person to transparently describe how you operate your business, the data you use, and how performance is presented can have a deep psychological impact on those involved. If you are heralding new information that will rock the boat, it's important to realize that human beings tend to see everything through an emotional lens. It's important to include those you may impact up front, and not surprise them with what you find.
However, in my experience, leaders who follow their instinct to react negatively to this information actively sabotage any attempt to become data driven. I call this destructive leadership reaction, "the transparency tax." The transparency tax is the burden an organization carries when their leader reacts to data without a sense of constructive curiosity but negative disbelief. They engage in actively tearing down the presenter, even when no data exists to justify the prevailing opinion. Those leaders teach their staff to be equally dismissive of facts, and the culture becomes a scramble to buttress the ego of their leader, even if it is against the best interests of the organization.
领英推荐
Leaders that do not have the emotional intelligence to handle an objective representation of reality will not be successful in the long term (HBR). Hubris syndrome, coined in 2009, gives evidence to the fact that leaders can become hostile to perceived threats to their power in an organizational culture that has little to no accountability to results.
I believe the technology industry, still being relatively new in history, still has leadership lessons to learn that other industries did over a hundred years ago. For decades, several industries operate natively from data. Look at professional sports: every conversation about?sports performance includes facts, figures, and statistics based on actual events. The leaders of these organizations can review a report and interpret if the function you're responsible for is successful or not, and the onus is on you as the leader to describe how are you going to improve performance. Dr. Bengt Holstrom's Nobel Prize winning work building incentivization systems to corral the principal-agent problem, agents of the firm that optimize for their incentive packages vs. the entire firm, is a good topic to study.
Make it a point to stay humble, keep the truth tellers close to you, and remember that learning and relationships unlock true happiness in life and in the workplace. I will discuss more of that in my next edition
Happy New Year,
Andre
Principal Program Manager at Microsoft
1 年Good stuff, I was just talking with a colleague this morning about how 1:1s / smaller early focus groups on new data can lower the chance of surprises over new data. Iterative learning is key too. First looks at new data collection are almost guaranteed to include bugs, and if we approach this with curiosity about how we can help improve accuracy of new analyses that culture of safety will cultivate legions of truth miners! ??