Is Business Acumen a Hard Requirement for Data Science?

by | Dec 10, 2018 | Career Skills

I recently ran across this post on LinkedIn:

A few of the follow-up comments also caught my eye (emphasis mine):

You make excellent points! Maybe part of the disconnect is not knowing how to use business acumen. It is useful in all levels, but I suppose it is more obvious how to use those skills in positions of leadership. (View original)

My point is that less people learn about blending analytics and business because not as many people teach it that way and solving business problems is not as easy as just applying algorithms, there needs to be a full understanding of the business use case and implications. (View original)

And I agree with both of the above. Show me one bootcamp (online or in person) or tutorial site that talks about business acumen and how it’s used in combination with data analytics, and I’ll send you a pizza. (No joke. You can use this form to send me the link. Don’t forget to include your address and desired toppings.)

As Elon Musk notes, “it’s important to teach problem solving … teach to the problem, not the tool.” (Musk has noted that this is a key concept at his super-secret Ad Astra, a real school registered with the State of California.)

So given all of this, I’d argue that having business acumen is not a hard requirement for data science — but the ability to think about how things fit into the constraints of a business is, and this skill is best learned from having experience in the business setting.

Why Experience Matters

The only reason companies place a lot of emphasis on business acumen is so folks have an idea of how things work in the real world. Business settings can be chaotic and riddled with gray areas. As an employee, you are of value to a company when you’re able to work through all the imperfections. The challenge, however, is that everyone has a different tolerance and definition of imperfection.

I’ve previously said that your ability to manage risk is proportional to your ability to manage gray areas. This is where experience becomes invaluable. But what kinds of experience?

  • It’s experience about using your accumulated knowledge to decide what to do in a novel situation.
  • It’s experience to look at cause and effect. If you do A, how does that affect B and operations in other parts of an organization?
  • It’s experience in knowing what to do if you are required to present one unified recommendation given that your team has dissenting opinions for what is best.
  • It’s experience in knowing how to work with all sorts of ambiguity.
  • It’s experience that tells you why Netflix paid $1,000,000 for an algorithm to improve the accuracy of its recommendation system by a mere 10%, and then why they didn’t implement the recommendation after their paid for it.

In short, it’s your ability to productively work within the specific ecosystem of the business (which includes all of the politics).

Yes, you are better off if you’ve experienced these situations before, but I’d also argue that knowing that those situations exist and knowing how’d you act in each one is also a good starting point.

Many argue that the data scientist is a rare breed of three skill sets: programming, math/statistics, and business. Lack of business experience is not a deal breaker — you can teach that. But you’re not going to go far in the data scientist industry if you don’t know how to develop context and use data to make effective decisions within the business that you’re supporting.

The Week’s Top Five

Everybody Lies: Big Data, New Data, and What the Internet Can Tell Us About Who We Really Are by Seth Stephens-Davidowitz (Book)

Stephens-Davidowitz makes a compelling case that despite what we tell people or post on social media about how they look, how we feel, and what we really think is quite different from the truth (as determined by search results).“Often, after I give a talk,” Stephens-Davidowitz writes, “people come up to me and say, ‘Seth, it’s all very interesting. But it’s so depressing.’” One of the big drivers for data-driven decision making is that, regardless of our intention, what we think will happen is very different than what actually happens. Think about that the next time you’re asked to provide an estimate of when you’ll have a project completed.

What If?: Serious Scientific Answers to Absurd Hypothetical Questions by Randall Munroe (Book)

If you’re looking for a textbook about cause and effect and using first principles, this is it. Munroe takes the ability to predict cause and effect to a whole new level as he provides serious answers to ridiculous questions by breaking down the problem to assemble an answer. My personal favorites include “Is it possible to build a jetpack using downward-facing machine guns?” and “What would happen if you tried to hit a baseball pitched at 90 percent the speed of light?”

To Sell Is Human: The Surprising Truth About Moving Others By Daniel H. Pink (Book)

If I told you that being able to sell is a major skill to have in your data science toolkit, you’d laugh me out of the room. But as strange as it sounds, being able to sell — your ideas, points of view, and recommendations — to the client is part of the day-to-day for the practicing data scientist. Pink does a fine job making the point that “we’re all in sales now” and then follows up with several strategies to help you get your points across to others.

Complications: A Surgeon’s Notes on an Imperfect Science by Atul Gawande (Book)

Complications is a must-read for anyone who is transitioning from academia to the business workforce. Gawande does an excellent job at explaining, through vivid examples, how the practice of medicine is actually imperfect, contrary to society’s belief. He writes that: “we look for medicine to be an orderly field of knowledge and procedures. But it’s not. It’s an imperfect science, an enterprise of constantly changing knowledge, uncertain information, and fallible individuals, and at the same time lives on the line. There is science in what we do, yes, but also habit, intuition, and sometimes plain old guessing.” Regardless of whether you work in data science, you need to get over this single fact: Like medicine, business is imperfect and sometimes requires “plain old guessing.”

Getting Past No: Negotiating in Difficult Situations by William Ury (Book)

The ability to resolve dissenting opinions between two or more people is invaluable, no matter what your career. Ury breaks down the five barriers that get in the way of cooperation and then outlines his five-step strategy for getting through to people: 1. Buy time to think. 2. Step to their side. 3. Reframe. 4. Build them a golden bridge. 5. Use power to educate. (A good follow-up to this book is Getting to Yes: Negotiating Agreement Without Giving In by Roger Fisher and Ury.)

Like What You See?

If this issue of Missing Data has brought you any value, it would mean the world to me if you shared it with one friend. Thanks!