5 steps toward data-driven leadership

A manager is flying across the desert in a hot air balloon when he realizes he is lost. He calls down to a man riding a camel below him and asks where he is.

The man replies, “You’re 42 degrees and 12 minutes, 21.2 seconds north, 122 degrees, 10 minutes west, 212 meters above sea level, heading due east by north east.”

“Thanks,” replies the balloonist. “By the way, are you a data analyst?”

“Yes,” replies the man, “how did you know?”

“Everything you told me was totally accurate, you gave me way more information than I needed, and I still have no idea what I need to do.”

“I’m sorry,” replied the camel-riding analyst. “By the way, are you a company manager?”

“Yes,” said the balloonist, “how did you know?”

“Well,” replied the analyst, “You’ve got no idea where you are, no idea what direction you’re heading in, you got yourself into this fix by blowing a load of hot air, and now you expect me to get you out of it.”

Data-driven leadership sounds so simple.  With the quantum leaps in technology over the last decade or so, shouldn’t leaders be able to effortlessly shift their decision-making to include better and better information?

As it turns out – nope.

As a flood of big data and pricey analysis (including an increasing number of outputs from machine learning/AI) rushes toward them, leaders are frozen – defaulting back to flying by the seat of their pants when the instruments in their leadership “cockpit” have never been more abundant.

Speak to business heads across industries, and you’ll hear a common theme: “We want our leaders to operate on a data-driven basis…but we’re not sure how to get them there.”

At Korn Ferry, we’ve been devoting considerable time and attention to helping organizations re-shape their leaders for a digital- and data-driven future. Here are what we recommend as the five key steps any leader can take toward becoming a data-driven leader:

1)   Understand cognitive biases, know your own, and work to manage them.  Cognitive bias – when the human brain operates irrationally in certain consistent ways, time after time – prevents data-driven insight from properly taking hold.

Consider, for instance, the anchoring effect – a bias in which the first number someone sees in a given context sets their mental framework for the rest of the numbers they see in that context.

Let’s say an executive sees a customer survey from several years ago, where product satisfaction was at 86% after the introduction of a rare and groundbreaking new product.

If they fall victim to he anchoring effect – and believe that 86% satisfaction is the norm going forward – they are unable to rationally evaluate all future surveys, where satisfaction will seem low. Data – even the best data – doesn’t properly impact their decision-making.

Accordingly, it’s critical for executives to understand which cognitive biases tend to affect them. Know your biases: there’s a good short list here, a good medium-length list here, and a nice long list here.

Mitigating cognitive bias doesn’t have to be anything fancy – sometimes it can be as easy as reminding yourself that you suffer from a certain bias. On a deeper level, well-documented strategies exist to combat cognitive bias and effectively return the brain to a more rational state.

2)   Become a savvy critic of data quality.

Once a leader has worked to understand and mitigate his or her own biases, the next place to turn a critical eye is toward the data itself.  Leaders would do well to always ask a series of basic questions about how data was collected, from whom, via what systems, etc.

Truly capable data-driven leaders care deeply where particular data came from, as well as how it was collected and analyzed, because data quality is absolutely critical to making data-driven decisions effectively. Consider the poor guy who had his house torn down due to a Google Maps error!

However, data-driven leaders also care about data quality because they want to tell stories using data. This critical influencing tactic hinges on being able to discuss where the information you’re leveraging came from in the first place. Without context, data driven stories can become purely academic exercises…void of impact for an audience.

3)   Constantly generate strategic questions you want data to answer.

When we looked at what makes for great leadership in the digital age, curiosity was identified as a mission-critical trait.  In the realm of data, curiosity indeed serves as a powerful tool.

Leaders who actively frame strategic questions to be answered by data – as opposed to passively taking in whatever analysis is offered to them – consistently make more thoughtful data-driven decisions.  Using strategy as a guiding light, and pushing data exploration rather than regurgitation, they are leading with data…rather than letting data lead them. (Don’t be the leader from the joke that opens this blog! Which was sourced here)

4)   Build strong relationships with data scientists and other providers of data-driven insight.

There’s a bit of what could be called a “pizza myth” around data in organizations – the notion that data can just be delivered to a leader’s door, ready for consumption.

Effective data-driven leaders get into the kitchen, so to speak – they forge true connections with the data scientists and other folks who drive the organization’s analytical journey.  By better understanding what’s possible – and the logic and limitations driving data-driven explorations in their particular organization – they help generate better information, and then use it better, in turn.

5)   Create a forum where data is discussed and debated by a diverse array of voices.

The dangers of a lack of diverse thought when data is generated and analyzed have been well documented. Limited samples – and algorithms framed by homogeneous (and accordingly mono-thematic) populations – generate often alarmingly bad results.

But the onus for diverse thought should not just fall onto data production and analysis – users of data need to make sure people representing varied and even conflicting viewpoints receive and react to the data, as well.

Data patterns that may seem opaque to one group may tell a compelling story to another group with a different frame of reference. (For instance, it might be obvious to parents of small children that a surge in sales of sunflower butter correlates to increasing nut bans in schools – context that childless twentysomethings might well lack.)

The above five steps represent a strong, concrete start toward becoming a better data-driven leader. Leaders who understand their own thinking, the data they’re looking at, the questions they’re trying to answer, the folks providing the data, and ultimately the right group to help them make sense of that data are at a tremendous advantage to their peers who operate by instinct alone.

Is shaping better data-driven leaders important to your organization? Here at Korn Ferry, we’re happy to continue the conversation.