He wrote to the Smithsonian Institution asking for everything they had on human spaceflight. He and his brother Orville devoured all the papers, theories, and calculations of the leading aeronautical researchers of their time. They studied the designs of Otto Lileinthal, the engineering principles of Octave Chanute, and Samuel Langley's tables on lifting and dragging.
After that they did something nice: they asked you everything.
When their gliders didn't perform as the published data predicted, the flying brothers didn't think they were wrong. They built their own wind tunnel and tested 200 wings themselves. What they found worried them. Lilleinthal's coefficients — the numbers the entire field relied on — were wrong, not just a little. Totally, totally wrong.
The Wirty brothers do not know this because of their arrogance; They used first principles thinking. And that's what made the difference between their first flight in 1903 and Langley – who had more money, fame and more – crashing his plane into the factory nine days earlier.
Today's scientists face a similar challenge. We are surrounded by playbooks, frameworks, and best practices – A/B testing guidelines, taxonomies, test selection models. These tools are very important. But like Langley's tables, they can mislead us when used blindly.
The most systematic data scientists I know do not choose between frameworks and first principles. They use both. Frame gives speed. First principles provide clarity. Together, they separate salespeople who work with organized partners.
What are the first goals to think about?
First principles Thinking means getting down to a problem from its basic facts and reconstructing it from the ground up. Aristotle defined it as finding “the first basis by which a thing is known.”
In practice, it means asking:
- What do we know to be true?
- What are we thinking?
- Which of those ideas can we challenge?
This does not mean rejecting the available information: The Wright brothers read every available study; They did not treat it as gospel.
For data scientists, the same applies. Field-proven map shortcuts for standard terrain. The first principle is the compass, to keep us oriented when the map gets a little blurry.
Why data scientists need this now
Frameworks are everywhere in data science for good reason. They help us run tests, define metrics, and build models quickly. But they can also create false confidence.
I've seen teams run flawless A/B tests that answered the wrong question. I've seen standard metrics produce dashboards that look impressive but measure nothing meaningful. This failure happened because the frameworks were flawed. They happened because no one paused to ask the first questions of the questions: What decision are we trying to inform? What value are we really trying to measure? Do we need this level of difficulty?
This is more important than ever because AI has done it to do side of data science. Gen AI can query data, generate through visualization, and use dynamic frameworks. But you can't decide if you're asking the right question.
First principles consider your separation. It's the skill that keeps the framework grounded in reality – and becomes the real strength of the strategic data scientist.
When first principles change everything
Here are three examples where the Data Science BookBook is well-written, but following it too rigidly has yielded negative results.
In addition to TextBook A / B testing
The outline is: Define your hypothesis, create random users, measure your basic metrics, test for significance. That totally works – if you ask the right question.
But the first question comes first: What decision are we trying to inform? What doubts are we trying to solve?
I once advised a group testing a new recommendation algorithm. The thinking framework said: Make random users, measure the average clicks, run for two weeks. Post the winner.
But the first principles broke and revealed something different. We weren't sure about the click – early signals suggested they would go up. We weren't sure if that click would drive real engagement or just noise.
So we changed what we measured. Instead of clicks, we focus on repeat visits, session depth, and long-term engagement. The result? The new algorithm increased clicks by 12% but decreased return visits by 8%. A standard draft would have said “send it.” The first principles speak “not yet.”
We decided that the new algorithm has been “click-baity”. The framework gave us a way. The first principles gave us the right question.
What exactly are our metrics?
MEASUREMENT OF CONNECTION MEASURES metric-North Star, Okrs, Heart-Powerful Because it provides structure. But they can also create an illusion by which we measure what is important.
First principles assume that: What is the basic behavior or value we care about? Does this metric actually capture it?
Consider getting involved. Most frameworks suggest Dau length, duration, or individual action. Proxies make sense – but are they real?
- For a meditation practice, longer sessions may seem “better,” but the basic goal is a steady habit. That would mean -short ones times over time.
- With an analytics tool, many questions for each user can signal intensive use, or it can mean that users struggle to find answers. The real value is faster, more targeted.
I once started a new job and inherited a dashboard that proudly reported active users per week as an effective metric of success. But when I dig inside, I see many “active” users were just logging in, looking around, and leaving without completing a single task. From the first principles of the lens, I decided that the real value was completed. And when I removed the metric, we found (as expected) that its use was very low under the new definition, but the new focus focused on getting reasonable acceptance.
The framework provides a menu of menus. First principles Thinking tells you if any of them actually show the value of your product. Sometimes a standard metric is perfect, but sometimes it's dangerously misleading.
Where first principles are reserved for presentation
One of the most obvious examples I encountered early in my career was when my team was tasked with creating a “user quality score” to help sales lead to sales.
The framework was clear: supervised learning, predicting possible changes, ranking by scoring. We had the data, the features, the method.
Two weeks in, as we fought our way out by adding the highest scores, someone asked: What sales decision will we actually make on that score?
We asked sales. The answer was not “give me direct opportunities.” It was: Should I take the time and custom to call this lead, or just send a quick scheduled email?
That changed everything. We don't need a complex model that takes the full possible journey. We needed a simple, adaptive classifier optimized for single thresholding.
After a while, we went from a nsenble model to a reasonable return, threw away half of our things, sent three weeks quickly – and was sent something used.
By returning to the original principles, we clarified the real problem and returned to the general framework for building a solution.
Compass and map
Here's the lesson: Strategic data scientists don't choose between frameworks and priorities. They put themselves together.
- The frameworks are map-Allowing you to move quickly and get the accumulated information.
- The first principles are these circle drawing tool-Keeping you aligned when the map clearly shows your path.
The Wiry brothers did not reject the research of their time. They build on it, but they also know when to return to the foundations.
That's the mindset shift that separates data strategies from risk. It's not about knowing a lot of techniques or working hard. It's about knowing when to follow the map and when to check your compass.
With the guidance of AI, frameworks will continue to be easy to use. But the campus – that's yours to build. And that's what will keep you relevant, strategic, and relevant for years to come.
This is one of the key themes I explore in my new book, Strategic data scientist: Level Up and thrive in the age of AI (Amazon Affiliate Link). It's about thinking about pairing foundations with proven structures to create impact, influence the streets, and position yourself as a strategic partner, not just a financial architect.
If you're wondering how your job will change as Ai becomes more capable, or simply I want to make a big impact as a data scientist and progress in promotion, Please check out this book on Amazon!
