The Art of Asking: A Data Scientist’s Guide to Stakeholder Interactions

The Dashboard Dilemma

Picture this: You’re in a meeting, armed with your laptop and a strong coffee, ready to dive into some serious data analysis. Suddenly, a stakeholder interjects:

“Hey, data wizard! Quick question—what’s our 3-year Average Revenue Per User (ARPU)?”

You resist the urge to sigh audibly. Instead, you take a deep breath and prepare to explain why this question, while seemingly straightforward, might not be as useful as they think.

The Art of Asking (and Why It Matters)

Here’s the deal. When you ask a data scientist for raw metrics like “How much traffic do we get from each marketing channel?” or “What’s our 3-year ARPU?”, you’re essentially saying one of two things:

1. You believe you can interpret the data better than a trained professional (spoiler alert: that’s unlikely).
2. You’re not entirely clear on the business question you’re trying to answer.

Both scenarios are about as productive as trying to paint a masterpiece with a sledgehammer. So let’s talk about the art of asking the right questions and why it matters for your bottom line. After all, that bottom line is probably why you brought a data scientist on board in the first place.

The “What” vs. The “How”

Here’s a paradigm shift that could revolutionize your business: Do your job, and I’ll do mine. As a stakeholder, your job is to handle the “what.” What are the business objectives? What are the key challenges we’re facing? What insights do we need to drive growth?

For example, instead of asking “What’s our customer churn rate?”, try framing it as “What factors are driving customer churn?”

My job, as your friendly neighborhood data scientist, is to handle the “how.” How can we analyze the data to answer these questions? How can we leverage advanced statistical techniques to uncover hidden patterns? How can we translate raw numbers into actionable insights?

Together, we’ll tackle the “why.” Why are certain strategies working (or not)? Why are customers behaving in specific ways? Why should we invest in one area over another?

From Metrics to Insights

So, instead of asking for raw metrics, try these on for size:

• “How can we boost customer retention?”
• “Which product features drive the most value?”
• “What factors are most influential in predicting customer lifetime value?”

Now we’re cooking! These questions provide the context I need to work effectively. I’m not just pulling numbers out of a database; I’m using my expertise to dive deep into the data, uncover meaningful patterns, and provide insights that can drive the business forward.

The Stakeholder’s Guide to Data Science Interaction

To help you navigate the complex waters of data science interaction, here’s a handy guide:

1. Define the business problem: Before you even think about asking for data, clearly articulate the business challenge you’re trying to solve.
2. Ask open-ended questions: Instead of requesting specific metrics, ask questions that allow the data scientist to explore and analyze.
3. Provide context: Share relevant background information and any hypotheses you might have.
4. Be open to surprises: The data might tell a different story than you expected. Be prepared to challenge your assumptions.
5. Collaborate, don’t dictate: Work with your data scientist to refine the questions and approach as you go along.

I can almost hear some of you protesting: “That’s all well and good, but the data scientists I’ve worked with didn’t have the domain knowledge to contribute this way. They couldn’t see the forest for the trees!”

First off, ouch. But also, fair point. Let’s unpack this a bit.

The Domain Knowledge Dilemma

If your data scientists seem to be operating in a vacuum (and not the kind that cleans your floors), unable to connect the dots between data and business impact, you’re facing one of two problems:

1. The Hiring Hiccup: You might not have hired the right data scientists for your needs. A good data scientist isn’t just a number cruncher or a code monkey— they’re more like a Swiss Army knife with a penchant for statistics. Let’s break this down further:
   • Technical skills aren’t enough: While proficiency in Python, R, or machine learning algorithms is crucial, it’s not sufficient. It’s like hiring a chef based solely on their ability to chop onions really, really fast.
   • The curiosity factor: Look for data scientists who ask probing questions about your business model, industry trends, and specific challenges. If they’re not as curious as a cat in a room full of laser pointers, keep looking.
   • Adaptability is key: The business world changes rapidly. Your ideal data scientist should be able to pivot quicker than a politician during election season.
2. The Silo Syndrome: Your organizational structure might be keeping your data scientists too far removed from the business. If they’re tucked away in a corner, only interacting with stakeholders through JIRA tickets and the occasional email, they might as well be on a deserted island, sending SOS messages in bottles. Here’s why this is problematic:
   • Context is king: Data doesn’t exist in a vacuum. Without understanding the business context, data scientists might focus on statistically significant results that have little practical impact—like discovering a strong correlation between ice cream sales and sunburn cases.
   • Missed opportunities: When data scientists are isolated, they miss out on casual conversations and impromptu meetings where valuable business insights are often shared. It’s like being the only person who missed the memo about casual Friday.
   • Lack of buy-in: If data scientists aren’t integrated into the business, their recommendations might be viewed with skepticism or misunderstood, leading to poor implementation of potentially valuable insights. It’s the corporate equivalent of bringing a chess set to a football game.

Consider this scenario: Your data scientist discovers a strong correlation between customer churn and the frequency of product updates. Without proper business context, they might recommend increasing update frequency across the board—about as helpful as suggesting everyone should eat more kale to solve world hunger. However, if they were more integrated into the product team, they’d understand that frequent updates are costly and might annoy certain customer segments.

The goal isn’t to turn your data scientists into business executives, nor to make your business leaders into statisticians. We need a shared language and understanding that allows the business to deploy both groups’ unique strengths effectively. Think of it as a corporate version of the Avengers—but with more spreadsheets and fewer Chitauri invaders.

Breaking Down the Barriers

What’s the solution? It’s time to break down those barriers and bridge the gap between data science and domain knowledge. Here’s how:

1. Hire Holistically: When bringing data scientists on board, look beyond technical skills. Seek out candidates who demonstrate curiosity about your industry, ask insightful questions during interviews, and show a willingness to learn.
2. Integrate, Don’t Isolate: Embed your data scientists within business teams. Let them attend strategy meetings, customer calls, and product demos. The more they understand about your business, the more valuable their insights will be.
3. Encourage Cross-Pollination: Set up regular knowledge-sharing sessions where business experts can educate data scientists about industry trends, and data scientists can explain the potential of different analytical techniques.
4. Invest in Onboarding: Develop a comprehensive onboarding program for new data scientists that includes deep dives into your business model, key challenges, and industry dynamics.
5. Foster a Learning Culture: Encourage continuous learning and development. Support your data scientists in attending industry conferences, taking relevant courses, or even shadowing colleagues in different departments.

Remember, a data scientist who understands your business is worth his or her weight in gold (or bitcoin, if that’s more your style). They’re not just there to answer your questions, but to ask the right ones, challenge assumptions, and uncover insights you might never have considered.

The Bottom Line

Remember, data scientists are not human calculators or walking dashboards. We’re problem solvers, pattern finders, and insight generators. When you treat us as such, that’s when the real magic happens.

So the next time you’re tempted to ask your data scientist for a quick metric, pause. Take a moment to think about the real business question you’re trying to answer. But also, reflect on whether you’ve set your data scientist up for success. Have you given them the context they need? Have you invited them into the heart of your business, or left them on the periphery?

By bridging the gap between data science and domain knowledge, you’re not just improving the quality of your analytics. You’re creating a powerful synergy that can drive real, tangible business value.

Together, we can turn data into decisions, metrics into meaningful insights, and questions into quantum leaps for your business. Now, isn’t that a lot more exciting than just asking for the 3-year ARPU or complaining that your data scientist doesn’t “get it”?

The right question is worth a thousand metrics. It’s time to upgrade our stakeholder-scientist operating system for greater interoperability.

Ready for a reboot?

For more columns from Michael Bagalman’s Data Science for Decision Makers series, click here.