Data scientists are equipped with some of today’s most in-demand technical skills. But to advance in the field, they also need to cultivate a less obvious area of expertise: communication.
With more seniority comes more responsibility to work across functions, and most colleagues are unlikely to understand the nuances of machine-learning models, multivariate analysis and advanced programming languages. Finding ways around these communication barriers is essential, however: The better data scientists articulate their ideas, the more likely other stakeholders are to implement them, and only then can these ideas translate into tangible results.
In numerous industry surveys, data leaders point to cultural issues like data literacy and organizational support as the most significant roadblocks to their success — and to their companies getting the most out of their data investments.
Communication is crucial in part because data insights are not confined to one department. Marketing needs data for customer segmentation and pricing, sales for forecasting, HR for workforce analytics and finance for investment decisions.
If technical teams can find a common language with these diverse audiences, opportunities for innovation and improvement will abound.
“When data teams and stakeholders are aligned, it feels like flying,” Andrew Chamberlain, Udemy’s senior director and head of product analytics, told Built In San Francisco. Everyone can move faster on projects, feel confident in their decisions and strive toward a common objective.
He encourages proactive communication and relationship-building so disparate teams have a solid foundation established before issues arise.
Boris Brown, a solutions architect at Publica, also advised using a narrative approach to ensure stakeholders understand the central conclusions first before they hear the steps it took to reach them.
Read on for Brown and Chamberlain’s recommendations for telling — and selling — data insights to non-technical audiences.
Udemy is an online learning marketplace serving more than 62 million students.
How does frequent communication with cross-functional stakeholders — especially non-technical ones — benefit your team?
Communicating clearly and frequently with cross-functional stakeholders is crucial for any data science team. When it comes to non-technical stakeholders, it’s essential to provide the context and the “why” behind projects so they can understand the significance of the work being done. It also gets teams in the same mind space — with similar context and expectations — preventing data teams from wasting effort on misaligned tasks.
Frequent communication can increase velocity. When information flows freely, teams can make decisions faster, and everyone can work toward achieving the best outcomes.
Perhaps most importantly, communication builds trust and empathy. Hiding information from stakeholders can lead to distrust and skepticism, ultimately throwing sand in the gears of projects. By having more informal conversations, teams build stronger relationships with stakeholders and can better anticipate possible objections and questions about technical data science work.
Data science is highly technical. When it comes to updating non-technical stakeholders on your team’s latest efforts, how do you translate the complexities of your work into digestible language?
As data scientists, communicating complex technical work to non-technical stakeholders is crucial to our job. To make this task easier, I approach it like a story. I start by focusing on the “why” before diving into the details. This helps motivate people to follow along with the technical material.
Communication is essentially teaching, and empathy is key when engaging with non-technical stakeholders. I take the time to understand where they’re coming from and meet them at their level of understanding. From there, I carry them toward a deeper understanding of the technical work.
Providing examples is also crucial in making technical work digestible. Without real-world examples, general statements can feel abstract and unrelatable. Concrete examples can illustrate how technical work impacts the business and our stakeholders.
Lastly, I always test my understanding of the technical work by asking myself whether I can explain it to my four-year-old son. If I can explain it in simple terms, it means I truly understand it. Therefore, when communicating with non-technical stakeholders, I simplify complex concepts and avoid jargon as much as possible.
Empathy is key when engaging with non-technical stakeholders. I take the time to understand where they’re coming from and meet them at their level of understanding.”
Can you recall a time when you needed to get buy-in for a data science initiative from a non-technical stakeholder? How did you secure that buy-in and what learning did you take away from it?
One example was when I needed to replace a production machine-learning model with a more modern algorithm. The data team had a strong point of view, but my product stakeholders didn’t see why it was worth replacing.
I approached the process of getting buy-in like a campaign. I collected facts on what was wrong with the current state, how it impacted the business and how the new algorithm would unlock new value. I summarized this in a written memo — not a slide deck — to help drive the conversation.
Long-form memos let stakeholders absorb the details asynchronously, share with colleagues and provide feedback. This adds better information to proposals and gets stakeholders invested in the initiative. I treat written memos like a sponge, absorbing new ideas from feedback and adjusting my proposals to work around criticism and defuse it early.
Another helpful strategy I used was pointing to other industry leaders who had successfully implemented similar algorithms, sharing papers and blog posts from leading tech companies. Doing so helped de-risk the proposal for stakeholders by showing that other companies had successfully taken similar approaches.
Publica is a connected TV ad platform that maximizes revenue and efficiency for publishers and advertisers.
How does frequent communication with cross-functional stakeholders — especially non-technical ones — benefit your team?
Frequent communication helps increase the awareness of irregularities for the non-technical audience. Understanding the cause of such irregularities helps transform our company and allows non-technical stakeholders to showcase examples and case studies of how we’re going above and beyond to service our clients.
It’s important to foster a healthy dialogue with stakeholders in different parts of the company, as it improves the end goal: the client experience. For instance, receiving a frequent feedback loop enables us to key in on insights that are most pertinent to our client’s business needs. At Publica, we focus on building the best ad breaks for streaming TV service providers. So if, for example, a client alerts us of a change in their tech stack, we can mine the current data to see if that change might negatively impact our ability to serve ads and make recommendations accordingly.
It’s important to foster a healthy dialogue with stakeholders in different parts of the company, as it improves the end goal: the client experience.”
Data science is highly technical. When it comes to updating non-technical stakeholders on your team’s latest efforts, how do you translate the complexities of your work into digestible language?
We mainly use data science to mine data, identify correlations and patterns and draw conclusions. Regardless of the complexity, it’s always helpful to start at the end and then elaborate on the findings that brought you to that conclusion.
Can you recall a time when you needed to get buy-in for a data science initiative from a non-technical stakeholder? How did you secure that buy-in, and what learning did you take away from it?
We helped encourage a client to take a more scientific approach to maximize their yield per ad break. We got their buy-in by showing them a trend in the data over a long period of time from year to date that illustrated if we were to switch to “x,” then we would likely get “y” results. After we implemented the change, we gradually saw a lift in revenue.
