For data scientists, the data and business objectives may seem to exist on entirely different planets—and they must learn how to successfully live on both.

Data scientists need to learn how to communicate insights in a way that the C-suite and other business leaders can understand, explains Megan Dixon, vice president of data science for Assurance IQ, a marketplace for insurance and financial products.

“It's really hard to build momentum and make decisions off of that data if people don't understand what the data is trying to say,” Dixon said.

When data scientists are unable to communicate insights, they miss business opportunities. Sometimes they may work hard to solve the wrong business problem, according to Dixon.

To help bridge the gap between data science and business, here are some strategies to consider.

Acquire Core Data Science and Analytical Skills

If you’re a data scientist, you must learn core skills for data science like SQL and Python. But to speak both business and data science, you will also need analytical skills.

“It’s not just being able to build a dashboard or create a machine learning model or deep learning or NLP but to understand why you're doing it, what the decisions are and how it could drive business,” Dixon said. “That’s the analytical thinking capability we're really looking for, and that ability to understand business is critical.”

While using large language models (LLMs) and generative AI are popular tech tools now, data scientists must learn the right time to use these tools for their business. Data scientists should become business savvy by learning to understand a business problem first before applying large language model (LLM) techniques.

“It puts the business problem at the core of what our data scientists are doing so business partners understand what the data science team is working on and how we're driving change,” Dixon said.

Master Business Value Translation

To speak both data science and business, learn to communicate the financial value of using ML algorithms, advised Eswar Nagireddy, senior product manager of data science at Exasol, an in-memory analytics database company.

He explained how data scientists must learn to communicate that degree of accuracy an algorithm provides. For example, if a business leader chooses to invest additional funds in the algorithm, it could boost the accuracy of that algorithm. A data scientist must learn how to ask a business leader for additional funds to improve a data science model and explain how the technology could save the company money each year.

Tech pros should learn how to correlate financial figures into ML KPIs. They should also know how to translate artificial intelligence (AI), cloud computing and GPUs into business KPIs, Nagireddy said.

“The skill set that is required is to have a transparency on the KPIs of every data science project and make it transparent to the data science team from the business,” he said.

Nagireddy recommended a course on Udemy called The Analytics Translator. It explains the analytics translator’s role in linking business, analytics and IT. The course also covers problem-solving, product management, software design and process modeling.

Align Data and Analytics Using AI

AI is a key skill to help data scientists solve business problems in a more efficient way using data, according to Dixon.

That includes not only ML and generative AI but “explainable AI.” Data and analytics engineering build trust in explainable AI, which is important for highly regulated industries.

“Explainable AI really means you as a data scientist understand the output of what you're building,” Dixon said. “Oftentimes data scientists focus on the results and the accuracy of a model as opposed to the explainability of the model.”

A data scientist may know that a model is accurate but not understand why it’s accurate. A recent IBM Institute for Business Value (IBV) Chief Data Officer study revealed that 63 percent of CDOs were aligned with business strategy, compared with 48 percent of CDOs.

“To get from data to valuable, actionable insights, you need to understand what drives the business, connecting your data and analytics strategy to business objectives,” wrote Kip Yego, program director for data and AI product marketing at IBM, in a blog post.

Going forward, companies should organize workshops either quarterly or twice a quarter to bring both the data science and business teams together as well as educate C-suite leaders, according to Nagireddy: “These design-thinking workshops, starting small, would actually impact a rapid transformation within the organization to bridge the gap between the business side and the tech side.”