Cassie Kozyrkov, Google's Chief Decision Scientist, has devoted her career to demystifying terms such as AI and Machine Learning. She actively advocates for a widespread, shared understanding of data science, decision intelligence, and statistics, particularly for those outside these specific fields.
I had the privilege of taking the "Making Friends with Machine Learning" course that Cassie developed at Google in 2017, and I still benefit from her straightforward explanation of everything Machine Learning.
For those who are unfamiliar, machine learning is an approach to making many decisions that involves algorithmically finding patterns in old data and using these patterns to create models (recipes) for dealing correctly with brand new data. In a nutshell: ML is all about finding and using patterns in data.
The phrase "Data is the new oil" was first coined in 2006, drawing a parallel between the latent value of raw data and crude oil, as well as the substantial effort needed to refine them into a useful form. Today, we have access to the tools to transform information into actionable decisions on any scale, and therefore it is crucial for decision-makers, including marketers, to recognize that data analysis provides control rather than absolute certainty; simply having more data does not equate to making 'right-er' decisions.
In her article, "How Should Data Analysis Impact Your Decision-Making Process?", Cassie dives into what statistics and data can and can’t tell you. If I could only share one thing I learned from the article it would be this: Do not ask "Am I making the right decision?" but rather "Am I making the decision intelligently?"
Here are some other key takeaways:
- Your statistician and data scientist are not miracle-workers. The elimination of uncertainty is impossible without possessing all possible data, which is often unattainable. They aim to guide the decision-making process informed by data, allowing for better risk management.
- Hypothesis testing involves determining a default action (the course you would take if forced to decide) and an alternative action (the path you would take if data analysis disproved your initial choice). This methodology is critical when deciding amidst uncertainty.
- The true strength of statistical analysis lies not in its ability to guarantee certainty but in its capacity to minimize the probability of wrong decisions. It enables informed decision-making based on the available data, rather than guaranteeing an inherently correct decision.
- Going back to my main learning, in the face of uncertainty, the question "Am I making the decision intelligently?" is a lot more useful than "Am I making the right decision?". Statistical methods and data analysis offer control when certainty is unattainable, and they should be used as a control panel for managing risks.
While I strive to remember the right question to ask myself the next time, I have to make a tough or snap decision, I’m also motivated by the conviction that while certainty is not always achievable, intelligent, data-informed decision-making at scale is.
Lin Classon, SVP, Global Product Management
Ready to move to 1:1 customer engagement? Let’s chat! Our suite of AI-native customer intelligence and activation software can help you build deeper relationships with your customers and grow a loyal customer base.
- Track, predict, and optimize your customers’ lifecycles with Kognitiv Pulse. Learn more.
- Enable 1:1 personalization at scale with Kognitiv Ignite. Learn more.
- Launch and manage a successful loyalty program with Kognitiv Inspire. Learn more.
- Intelligently acquire and engage customers across paid channels with Kognitiv Amplify. Learn more.