Choose Wisely: Data Scientist Cassie Kozyrkov Reveals How to Make Smart Decisions

UPDATED: June 4, 2024
PUBLISHED: June 4, 2024
cassie kozyrkov decision intelligence

As the world becomes more and more data-driven and artificial intelligence increasingly permeates daily life, our power as human beings lies in our ability to make informed decisions. Cassie Kozyrkov is on a mission to increase the quality of those decisions. Kozyrkov was Google’s first chief decision scientist until she left the company to start Data Scientific. This decision intelligence agency helps world leaders and executives optimize their most important decisions.

Kozyrkov’s goal is to lead humanity toward better decision-making on a global scale. In her recent LinkedIn course, she presents 18 tried-and-true lessons that every business leader can implement to ensure their organizations thrive in a data-driven world full of complex decisions. Outside of that course, she’s also given countless talks and written numerous in-depth blogs on improving people’s decision-making skills.

The importance of good decision-making

Cassie Kozyrkov sees good decision-making as the steering wheel of people’s lives. As human beings, everything that happens to us falls into one of two categories: things we can control and those that we can’t. One thing we can always control is the quality of our decisions. Making even the slightest improvements to our decision-making skills has a compounding effect over time and gives us more control over what happens in our lives. But for many decision-makers, it isn’t just their lives that their choices impact. As people get more and more responsibility, their decisions have a massive effect on their families, communities, organizations and, in some cases, even entire countries.

Start with one simple question

Kozyrkov believes that people can become better decision-makers if they focus on one question: “What would it take to change your mind?” Whether you’re asking your internal dialogue or your spouse, coworker or anyone else, that one inquiry holds a lot of power. She says you can’t answer that question without first knowing if your mind is set or not. “If it isn’t, your procedure will be to go out, explore and go with whatever looks like the potential best option. This is true in the household decision-making sense. And it’s also true in the data science and statistical sense.”

SUCCESS Speakers Bureau offer

Overcome biases

Decision-making is riddled with bias. Cassie Kozyrkov says that two of the biggest to be aware of are outcome bias and confirmation bias. Recognizing and overcoming these biases is critical to becoming a better decision-maker. Doing so means being self-aware, challenging assumptions and seeking diverse perspectives to counteract internal biases.

Outcome bias

To understand outcome bias, we must first know the difference between a decision and an outcome. A decision is a choice made after considering all options. It’s limited to the data available to you at the time and the technique you used to seek out additional information and ultimately make that choice. An outcome is your decision plus luck—or all those things you can’t control. Outcome bias is when you forget the luck and analyze the quality of your decision based solely on what happened rather than how you made the decision.

Confirmation bias

Confirmation bias occurs when people interpret information in a way that confirms what they think they already know. To counteract confirmation bias, Kozyrkov recommends getting into the habit of setting the criteria before you have all the information. For example, decide on the minimum salary you’re willing to accept for a job before you receive an offer. That way, when you get the offer, you can quickly and effectively decide based on the criteria you set rather than emotion.

Practice decision intelligence

Decision intelligence is making sound decisions using principles from statistics, data science and psychology, among other fields. Similar to creativity, decision intelligence is like a muscle. The more you practice and train it, the stronger it becomes. That said, not every decision requires a systematic approach. And that’s where Cassie Kozyrkov stresses the absolute importance of prioritization. What you put on your sandwich for lunch shouldn’t have the same level of prioritization as the go-to-market strategy for a new product.

Stand up to AI

AI systems are increasingly becoming more prominent in our daily lives. Generally speaking, they make life easier. Whether predicting market trends or optimizing resource allocation, AI empowers decision-makers to leverage data effectively to reach better outcomes. But Kozyrkov wants decision-makers to balance AI insights and their own human judgment. She also wants people to know more about how these systems are created before unquestioningly trusting them. At their core, AI systems are simply the amplification of three decisions made by the people who built them. Those decisions come down to three questions:

What is the objective?

What is the system trying to optimize, and what does success look like? There’s no one correct answer to any of those questions. The team of people building the system has to make a judgment call as to what the purpose of the system is.

What data does the system learn from?

AI systems rely on complex data sets to function. The type and quality of that data dictates the system’s reliability, bias and objectivity. To be successful, the data needs to represent the people using the system. Suppose it’s intended for use by people in sub-Saharan Africa but relies solely on a data set from a demographic like Silicon Valley. In that case, the system won’t be helpful to its intended audience.

How do you test the system?

What does “good enough” mean? How and what needs to be measured, and in what context? If the decision-makers building the system can’t align on this, it’s best not to launch it.

Cassie Kozyrkov has repeatedly said in her work that people need to “stand up to AI.” What she means by that is, society as a whole needs to stand up to the quality of the decision-making behind AI and ensure that good decision-makers are, at the very least, coaching those building the system. Because good decision-making isn’t about the outcomes but the technique used to reach those outcomes.

This article originally appeared in the July/August 2024 issue of SUCCESS magazine. Photo by Kevin Scanlon.