First principles reasoning: How thinking like a scientist can make you a better business analyst


“We absolutely must leave room for doubt or there is no progress and there is no learning. There is no learning without having to pose a question.”

“The Role of Scientific Culture in Modern Society” - Richard Feyman

Look closely at the work of any business analyst with an outstanding record of solving complex problems, and you’ll probably find multiple examples of them challenging existing methods and conventional wisdom to unleash creative solutions.

At the core of this ability to solve problems that no one had cracked before is what we commonly refer to as first principles thinking. In the words of James Clear,

First principles thinking is a fancy way of saying “think like a scientist.” Scientists don’t assume anything. They start with questions like, What are we absolutely sure is true? What has been proven?

An anecdote I heard years ago on a podcast illustrates how not thinking from first principles can limit results. The guest of the show told the story of how as a 12-year old she had seen from a window a neighbor smoking, and resolutely walked over to tell him how bad smoking was for his health. The neighbor looked at her unfazed, then took another puff. Once she realized her warning had been in vain, she turned around and sheepishly walked back home.

In the interview, the woman used this story as an example of how common it is for us to erroneously assume that we can convince people with data. As I previously wrote, “Merely getting people to understand intellectually what needs to happen is not enough. Statistics and facts may get people to believe, but feelings are what inspire people to act.”

An analyst helping design a smoke cessation intervention could avoid this mistake by challenging any belief not supported by firm evidence. Instead of accepting conventional wisdom that says, “If you want people to give up smoking, you must tell them it kills them.”, the analyst would ask, “How do we know this is true? What are the sources?”

These questions could lead to experimental evidence that shows that the source (the 'messenger') heavily influences smokers' intention to use e-cigarettes and tobacco cigarettes. With this knowledge, the team would be able to step outside traditional thinking and devise a more effective intervention to help smokers than simply giving them information about cigarettes’ harmful effects.

Examples from my role as a data scientist

In my job as a data scientist, going back to first principles has saved me multiple times from project failure or suboptimal results.

In one of my projects, I worked on a software application that displayed the estimated location of moving assets equipped with GPS sensors, updating the asset position every 5 minutes. Before I joined the team, decision-makers had been convinced that customer trust required all location estimates to be highly accurate. By asking, “Why do we think this is true?” we learned that the customers would be happy to wait 10 minutes for an accurate position. Based on this finding, we avoided wasting large amounts of money on models to ensure the accuracy of all location readings. Instead, we redirected our efforts to a much less expensive solution that consisted of detecting and discarding the occasional incorrect GPS reading. The system would simply wait for the next reported position to update the asset position on the screen, ensuring that customers only saw accurate location estimates.

Another time, I joined a team tasked with detecting fraud on a search platform that presented advertisements along with search results. The initial assumption was that some suspicious search activity observed in the data was not fraudulent because it only involved searches without clicks on the ads displayed. Since the platform used a pay-per-click model, everyone thought the activity was from authentic users, not fraudsters trying to increase revenue from fake activity.

Rather than taking their conclusion for granted, I decided to do more research. It turns out that outside the platform, advertisers were hiring marketers to create their sponsored content. Those marketers were paid on a pay-per-impression model. The initial assumption that only clicks produced a reward and therefore suspicious traffic without clicks was not fraudulent had been wrong.

How to apply first principles to your BA work

When faced with seemingly intractable business problems or simply looking for non-traditional approaches to traditional problem-solving, ask yourself,

What are we absolutely sure is true? What has been treated as truth without being proven?

Sometimes the best solutions are right in front of us, hidden in plain sight. Get in the habit of working from first principles and you’ll find it easier to cut through preconceptions to change the business question and quickly see alternatives that you may have missed.

First Principles Thinking

To learn more about this powerful technique, read First Principles: The Building Blocks of True Knowledge

Author: Adriana Beal

Adriana Beal has been working as a data scientist since 2016. Her educational background includes graduate degrees in Electrical Engineering and Strategic Management of Information obtained from top schools in her native country, Brazil and certificates in Big Data and Data Analytics from the University of Texas and Machine Learning Specialty from AWS. Over the past five years, she has developed predictive models to improve outcomes in healthcare, mobility, IoT, customer science, human services, and agriculture. Prior to that she worked for more than a decade in business analysis and product management helping U.S. Fortune 500 companies and high tech startups make better software decisions. Adriana has two IT strategy books published in Brazil and work internationally published by IEEE and IGI Global. You can find more of her useful advice for business analysts at



Copyright 2006-2024 by Modern Analyst Media LLC