Did advanced analytics ruin baseball? Could it also ruin your business?
Workplace tech consultant Phil Simon believes so. He lists unhealthy analytics as one of the tectonic forces reshaping the workplace in his book The Nine. And while most of the other eight trends are well-known – such as employee empowerment and generative AI – useless data stands out as unexpected in an era when we worship analytics.
In baseball, it started with the story from the book and movie Moneyball. General Manager Bill Beane catapulted his poorly financed Oakland Athletics to success by leveraging unexpected statistical findings in choosing players and playing the game. But Mr. Simon argues that all the other teams eventually became data-obsessed and the result is boring, by-the-numbers baseball, games stripped of their human character and for many years overly long.
He is not anti-analytics, having celebrated companies like Google, Netflix and Amazon in previous books for intelligently using data, gleaning fascinating insights that powered their success. But now he warns: “Using data to make any decisions guarantees precisely nothing. The list of firms that made regrettable decisions despite using it is long and distinguished.”
In the workplace, unhealthy analytics trace back to Frederick Winslow Taylor, who studied factory workers performing manual jobs with a stopwatch and determined the single, optimal way to complete each task. The result was a dehumanization of work. Over the years we have tended to flip between times when analytics hold sway and when the ideas of humanists like Abraham Maslow, Stephen Covey and Henry Mintzberg galvanized managers. Today, arguably, both approaches are happening simultaneously, albeit usually in different workplaces. Some managers strive for soft leadership while, as Mr. Simon notes, Amazon puts so much pressure on its warehouse employees that they are like NFL players who have to use their days off to recover from the physical poundings of the job. He cites an internal study that found unrelenting computerized surveillance of workers’ every move is leading to comparatively high injury rates and churn so that Amazon has to hire the equivalent of its work force every year and could run out of possibilities soon.
Then there’s performance reviews. Love ‘em or hate ‘em … well, the fact is most people hate them but the evaluations continue, supposedly effective. The academic literature, however, questions them. “In most organizations, traditional performance reviews are so bad they do more harm than good,” says Stanford University professor Robert Sutton.
Mr. Simon serves up two rules for you to consider. The first is Goodhart’s Law, named after British economist Charles Goodhart: “When a measure becomes a target, it ceases to be a good measure.” Campbell’s Law is similar but more explanatory, formulated by psychologist Donald Campbell: “The more any quantitative social indicator is used for social decision-making, the more subject it will be to corruption or pressures and the more apt it will be to distort and corrupt the social processes it is intended to monitor.”
Together the two laws tell us that if you define success in terms of a specific measure, you reduce its ability to accurately measure that success. They apparently run counter to what I will call Drucker’s Dictum, after legendary management guru Peter Drucker: “If you can’t measure it, you can’t manage it.”
Indeed, we might add Martin’s Musings: Former Rotman School of Management Dean Roger Martin in his Medium blog warned of a dangerous schism faced by MBA students and by implication executives in general: “Your strategy course is utterly inconsistent with your statistics course.” He was highlighting the fact that 100 per cent of the data that you need to make decisions based on rigorous analysis comes from the past and there is no data about the future.
There is truth in all those observations. That’s important to keep in mind as managers, following Drucker’s Dictum, are deepening their desire to improve their diversity programs through better measuring. Entrepreneur Randal Pinkett offers a thorough guide in his new book Data-Driven DEI and corporate director Lee Jourdan selected seven metrics to measure your organization’s DEI progress for Harvard Business Review.
He tries to cover the entire employee life cycle: Attrition, performance, promotions, leadership pipeline, employment pipeline, pay equity and inclusion. “It is rare to find any organization currently utilizing all these metrics, but all should aspire to,” he writes.
Within each metric, he expects companies to track various employee cohorts, paying attention to different regions and cultures. For attrition, you want to study voluntary and involuntary. You need to know whether you are disproportionately losing or letting go of people from under-represented groups. This should include not only collecting hard data but also conducting exit interviews with all departing employees to figure out whether bias or not feeling included had anything to do with their departure. If you conduct performance reviews, you need to ensure equal distribution of high and low ratings is occurring across all groups.
He observes that “transparent data provides one version of the truth and helps organizations determine priorities.” I’d stress “one version” and “helps” – if data is taken too narrowly, we can be in the world of unhealthy analytics.
Cannonballs
- Companies love to protect, notes Basecamp chief executive officer Jason Fried: They protect brands with trademarks and lawsuit, data and trade secrets with rules, policies, and nondisclosure agreements and they protect money with budgets, CFOs and investments. But often they fail to protect what’s most vulnerable and precious: Their employees’ time and attention.
- When Lorraine Marchand, author of The Innovation Mindset, was general manager of the life sciences division of IBM’s Watson Health, she gathered her direct reports at the end of the week for a stand-up meeting she called Fail-safe Fridays, in which they each recounted one thing they had done that week that hadn’t worked, so the group could learn from it.
- New research by Bain & Co finds companies rely on past sales data for assigning sales representatives when it would be better to rebalance sales accounts based on customers’ expected future spending.
Harvey Schachter is a Kingston-based writer specializing in management issues. He, along with Sheelagh Whittaker, former CEO of both EDS Canada and Cancom, are the authors of When Harvey Didn’t Meet Sheelagh: Emails on Leadership.