Your Job is to Create Value

As a data scientist or data professional, you might think that your primary job is to write code or analyze data.

In some sense that’s true, but those things are not the ultimate purpose of your work.

Ultimately, your job is not to write code. Writing code is important, but that’s not really the goal (I’ll explain more about this later).

And as a data professional, your job is not to analyze things.

As a data pro, your job is to create value.

Of course, the way that data scientists create value is with code, reports, and analyses.

But whatever you do, you need to look at your work and your actions through the lens of value creation.

You always need to ask the questions, “Will this generate cash? Will this save cash? How will it do so now or in the future?”

Two broad ways to create value

The higher you climb in the data industry, the more aware you’ll become of the relationship between the low level work and how it connects to value creation.

Team managers, middle managers, and executives need to think more about money (in the form of budgets). As you move up, you’ll be asked to identify how your team’s work contributes to strategic goals.

Over time, you’ll start to see the connection between lower level tasks and higher level performance metrics.

But earlier in your career, it’s maybe a little hard to see. Most people in their first few years as an entry-level to mid-level data scientist don’t have to think about such things. Some senior level data workers still don’t think about their work in terms of value creation … they’re just so good at what they do that it doesn’t matter, and their managers think about value creation for them.

Having said that, the sooner that you can think about your work in terms of value creation, the better. People who create large amounts of value for a company are more likely to be promoted, or at least rewarded (although, it’s a little more complicated than that).

Ultimately, there are two primary ways to create value for a business:

  • generate revenue
  • reduce expenses

Without diving too much into finance and accounting, these are both really about generating positive cash flow for the business, both now and in the future.

Generate revenue

It’s probably obvious that one of the best ways to create value for a business is to generate more money for the business.

A little more specifically (and technically), you want to generate cash flow for the business, both today and in the future. Speaking from a finance perspective, the value of a business is actually equal to the sum of all future cash flows.

Reduce Expenses

The other way to generate value is to save money (i.e., reduce expenses). In the end, saving money helps generate positive cash flows.

There are a few ways to do this as a data scientist, which I’ll get to in a second.

The overall point here is that you need to think of your work in terms of how it will create value for the business (at least in the back of your mind).

How data scientists create value

So as a data-centric worker, how do you create value?

There are a few different ways.

Find insights

One of the best ways to create value is to find insights.

It’s also one of the most commonly requested on job postings. HR representative and hiring managers almost always insist that job candidates must be able to “find insights in data.”

In terms of deliverables, finding insights in data commonly means creating reports and analyses that identify opportunities to create value (by reducing costs, driving customer acquisition, increasing cross sell opportunities, etc).

In terms of skills though, finding insights really depends on two foundational data science skill areas: data visualization and data manipulation. Finding insights requires you to analyze data, and data analysis is essentially just an application of data manipulation + data visualization, for the most part.

As it turns out, “finding insights” is probably the easiest way to create value as a junior data scientist. With about 3-6 months of intensive training, most smart people could learn data manipulation and data visualization sufficiently to be able to find insights in data. Other techniques (which we’ll talk about in a moment) are harder to learn and harder to apply.

This is why I recommend that beginners relentlessly study data manipulation and data visualization. Because you can use these skills together to “find insights,” which in turn enables you to start generating value for a business.

(If you’re ready to start learning data manipulation in Python, check out our blog post The 19 Pandas Functions you Need to Memorize.)

Make predictions

Another way to create value as a data-worker is to make predictions.

This is actually quite a bit more difficult, and should be reserved for data scientists with at least a couple of years of experience.

I’ll note that I’m not necessarily talking about making forecasts about the future. In fact, I’ve very weary of trying to make forecasts about the future.

But you can make predictions about some types of things that are very narrow in scope. For example, you can use certain types of machine learning techniques to “predict” other products that a customer might like. Netflix, Amazon, and many online businesses use these types of recommender systems. These sorts of predictions increase business value by increasing transaction sizes and by finding opportunities for cross-sell and up-sell.

These sorts of predictive systems are very common in marketing departments, where finding new prospective customers or deepening customer relationship are very important. Prospecting and customer relationship management are large drivers of revenue, which in turn creates business value.

You can perform these services with machine learning techniques like logistic regression and clustering.

Build systems

System building is another way you can create value, but in many ways, this requires you to understand software engineering as much as data science.

Many times, when you create basic deliverables like reports (for delivering insights) or machine learning models, those deliverables are “one off” projects that are produced once and delivered once.

In some cases though, those reports, models, or other deliverables can be turned into systems that generate deliverables at some regular interval. For example, you might create a system that generates a report every day/week/month; or you might create system that uses machine learning to identify new prospects for a marketing campaign every month.

As I said, these projects often require you to be able to develop larger software systems using software engineering techniques, as opposed to simple data science “scripts.”

Write clean code

One way to generate value that you might not think about is writing clean code.

Code often needs to be shared between teammates and needs to be maintained over time.

When you right messy code – code with confusing names, messy structure, code that’s difficult to read, etc – you’re actually imposing a cost on your team. If your code is difficult to read and difficult to understand, that will make it harder for your team members to understand and use. In turn, this often requires them to take more time to understand end execute. As they say: time is money. When you write code that’s difficult to read and use, you’re ultimately imposing a sort of cost in the form of extra time.

Similarly, messy code is often harder to maintain (again, because it’s harder to understand). In turn, this means that maintainance takes more time. Messy code also causes more bugs, which impose costs of their own.

All of this is to say that you can create value for your business by writing cleaner code. Use clear variable names. Write functions with a limited number of parameters (even if that requires you to define more functions).

Always remember: clean code is an asset, and messy code is a liability. Said differently, clean code generates more value.

Closing thoughts

Looking at data science work through the lens of value creation is actually very important, although rarely done.

No one ever sat me down and explained this to me. Early in my career, I just thought in terms of completing tasks, and it took some independent reading to finally stumble on some of these ideas (which are best found in introductory books on finance).

You’ll progress faster in your career if you consistently think of your work in terms of value generation … in terms of how you can increase revenue or decrease expenses. Don’t just think in terms of completing tasks.

Always try to think through how your work will generate positive cash flows for your business. If you aren’t sure that it will, think about how you can modify or improve your work so that it does.

And constantly try to increase your skills so that you increasingly add more value over time.

Joshua Ebner

Joshua Ebner is the founder, CEO, and Chief Data Scientist of Sharp Sight. Prior to founding the company, Josh worked as a Data Scientist at Apple. He has a degree in Physics from Cornell University.

Leave a Comment