Getting a Seat at the Table as a Data Scientist

By Becoming an Asset, Not An Expense

Hello, it’s Ethan & Jason. Welcome to Level Up: Your guide to grow fast, avoid mistakes, and make optimal career moves.

EXCITING NEWS: We’ve been named a Top Leadership & Career course and Maven is offering 20% off for a limited-time:

• Stuck at Senior Manager - How to Break Through to Executive - For leaders who want to grow to Principal, Staff, Director, or VP. (save $299)

• Get Promoted Faster - For ICs & new managers (<10 years experience). Get tactical “how-to” insights for rapid career growth. (save $99)

This is the biggest discount that will be offered. Both come with Maven's strong money-back guarantee, so you can sign up with confidence.

Code MAVEN100 applied automatically in the links above.

---

We are thrilled to bring you a guest post by Yuzheng Sun (staff data scientist at Statsig.com). He was previously a director of data science at Tencent, a data scientist at Meta, and an economist at Amazon. He holds a PhD in economics from Cornell. Yuzheng is also a top AI instructor on Maven.com and a content creator who shares insights on data and growth on LinkedIn and YouTube with over 300K followers across the Internet.

Throughout his career, Yuzheng has seen three types of data teams: 1) SQL Monkeys, 2) Scientists, and 3) Decision Makers.

In this post, he breaks down what a seat at the table looks like, why you want that seat, how AI will replace those not at the table, six principles in becoming a decision maker, and three weapons for success.

(these lessons are applicable to ALL domains, read on)

---

What does a seat at the table look like?

Most data scientists and analysts fall into the first two categories, either SQL Monkeys or Scientists, making it hard to envision working as a Decision Maker. At Facebook Marketplace, the influential data scientists were definitely in the third category. At Tencent, I transitioned my team from a mix of the first two to a mix of the latter two.

Here are some characteristics of Decision Makers that I have observed:

• Involvement in Decision-Making: Data science leaders participate in all key decision meetings. No important decision is made without them. This hierarchy means a product-lead VP needs a VP/Senior Director DS's sign off, a product-lead director needs a Director/Senior Manager DS’s sign off, and so on.

• Influence on Decisions: DS leaders don't just sign off the decisions—they shape and sometimes initiate them. For example, our DS leader drove a major shift in Facebook Marketplace priorities from B2C to C2C.

• Ownership: Data scientists are seen as business owners, not just data handlers.

• Strategic Involvement: From leaders to individual contributors, DS are involved in setting roadmaps, priorities, goals, and evaluating team performance.

• Valued Collaborators: Engineers and PMs value discussions with DS and seek more interaction. DS are viewed as teammates, not obstacles.

• Responsibility: DS owns significant scopes and is sometimes directly responsible for KPIs, such as DAU, retention, or incremental revenue.

Being a decision maker has lots of benefits. Most importantly, it gives us a sense of purpose and ownership, thereby making our work meaningful.

Why do you want a seat at the table?

Many DS and DA remain "SQL monkeys" because it's comfortable. You meet requirements, feel needed, and avoid deep thinking. When things go wrong, you can blame others, and things go wrong all the time. So, why take the risk of growing your role?

Some prefer the "Scientist" role. We were often rewarded for being smart students, and we want to continue that pattern at work. We solve problems others can't, demonstrating our intellectual superiority while boosting our technical skills, salaries, and career growth. That’s the dream, right?

Not exactly.

Businesses hire us to help them earn more money. But, unlike engineers or sales, our impact is indirect. We don’t ship code into production, we have to rely on engineers, leaders, and other roles to take actions based on our advice to make tangible improvements. So, how do we contribute?

We might justify our work by saying that our segmentation model increased email opens or that our sentiment analysis provided valuable insights. But the truth is simpler— effective Data Scientists help businesses make better decisions.

Erik Gregory summarizes this well.

Ultimately, our role as Data Scientists (excluding machine learning engineers) is to ensure the business chooses X over Y, where X is better.

The danger for “Scientists” lies in pursuing complex problems instead of identifying the right problem and a practical solution that drives a business outcome. In my interview with Ethan, he described this tendency as "building monuments to their technical expertise," which is quite accurate.

Another risk is becoming isolated. It's fine to be an expert and tackle challenges others can't, but spending too much time on this can detach you from reality, especially if you are purely focused on technical problems because others handle the real-world issues.

That's why we need a seat at the decision-making table. Otherwise, we're justifying our existence through others' perspectives, following arbitrary rules, and will lose agency and meaning in our work.

In my first year at Meta, I was mostly churning out analysis handed down by others. Some of it mattered, but a lot of it felt like “drawing a target around where the arrow landed.” I was miserable, and I couldn’t push back — my voice got lost in too many layers.

In my second year, things changed. I started influencing decisions and built direct relationships not just with data science leaders but also with engineering and PM leads. My insights started to matter. Directors and VPs regularly noticed my notes. That’s when the work started to feel meaningful, and I could mostly shape my own agenda.

How AI will replace those not at the table

I didn't need to bring AI into this argument, but it's more relevant than ever, and I happen to have something to say as one of Maven’s top AI instructors. So let's make it explicit.

When I first joined Facebook, I had some bad habits from being a "Scientist." I wrote a detailed 36-page analysis of our business using random forests and robust panel data analysis. An engineer kindly read through it all and then asked me, "YZ, it's an impressive study. Could you tell me what I should do with it?"

I didn't have an answer.

Despite the fancy analysis, I couldn't tell her how to do things differently. I've since changed my approach, but I still see similar analyses from other scientists—technical, comprehensive, and thorough. These sound like the qualities of a good paper, but most of them fail to guide better business decisions.

When GPT-4 came out, I quickly saw the problem. GPT-4 can easily produce lengthy analyses like these, but it can't reliably advise on improving business strategies. That means that any scientist who only provides analysis is at a greater risk of being replaced.

For ideas to be truly valuable, they must be original and contrarian, backed by solid reasoning. Lengthy analyses often reorganize existing information in complex ways without producing anything new or useful, which is exactly what AI does as well.

AI is trained on consensus and can transform information better than we can. So, AI will render these types of analyses obsolete. I look forward to that day, because I always felt like such type of analysis were red herrings that distract us from what truly mattered. I hope by then, data scientists will stop producing lengthy and toothless analyses.

Six principles in becoming a decision maker

Understanding how data scientists influence decision-making is a nuanced topic that has intrigued me for a long time.

At Meta, whenever I met with senior data science leaders, I asked how they established a truly data-informed culture. I also spoke with product and engineering leaders. At Tencent, I applied these insights successfully. I’ve summarized my learning into six principles.

1. [Leaders] Are right, a lot; have backbone, disagree and commit

These two Amazon leadership principles encapsulate these insights well:

The first time Data Scientists are at the table, people won’t take us seriously. This is because, unlike other roles, we lack immediate leverage. Things will get moving with or without data. So, we have two choices:

• Compromise: If a leader wants data to support their conclusions, we might skew data to fit their narrative. This erodes trust and respect over time.

• Disagree and commit: The harder choice is to stand by what you believe is right. Make your dissent known, but do it wisely. The goal is not to prove others wrong but to move things forward. Sincerely disagree and commit without gloating about being right later.

2. Produce content

Since we can't create products directly, content amplifies our influence. It serves three purposes:

• Build reputation and relationships at scale: Writing is a force-multiplier for trust. A well-crafted analysis travels far beyond your immediate network, silently introducing you to hundreds of colleagues who now see you as “the person who really gets X.” At Meta and Tencent, publishing just three deep-dive pieces in a niche domain won me more credibility than months of one-on-one meetings ever could. Long after I left these companies, The takeaway: invest the time once, let the document keep “socializing” for you.

• Make sure your insights are actually heard: In live conversations, you cede control to the room’s dynamics; attention drifts, side-tracks erupt, and — worst of all — counter-intuitive findings get waved away before you can lay out the logic. Writing flips that script. A document lets you present the full chain of reasoning, complete with evidence and nuance, so even surprising conclusions feel inevitable by the last paragraph. It isn’t just broadcasting; it’s preserving the integrity of the idea long enough for people to absorb it.

• Practice: Writing forces clarity and being engaging, preparing you for high-stakes discussions and presentations.

3. Go to market multiple times

Rarely do we find the perfect idea or audience on the first try. Repeated attempts help refine our approach and reach more people. Present to your friends, teammates, manager, stakeholders, and people at other organizations. You can constantly learn who cares about what and improve your angle.

4. Start with trustworthy data

As a Data Scientist,your reputation relies on trustworthiness. Admit when you don't know something, invest in data quality, and find the ground truth. Correct mistakes tactfully to maintain authority.

This video shows you the method of finding ground truth by going into the code.

5. Have opinions without data

Decisions often require opinions even without complete data. Lacking data never stops PMs or leaders from having strong opinions, and it shouldn’t stop you either. Make sure you are clear on what was suggested by data and analysis, and what is your personal judgment.

6. Stop being defensive

Avoid making presentations overly technical to dodge challenges. A useful presentation generates debate and seeks impact proactively. This video shows you why and this video shows you how.

Three weapons for success

The six lessons mentioned earlier are helpful, but not enough. To increase your chances of success, consider these three weapons.

1. Sponsor

Want more context and crucial information to make informed decisions? Learn from your sponsor.

Need to be right often without undermining others? Let your sponsor socialize your insights for you.

Looking to secure a voice at the table or get opportunities for high-stakes presentations? After crafting an insightful analysis, ensure it catches the attention of leaders through your sponsor.

Having a sponsor in your career acts as a cheat code. Finding a sponsor requires many traits from you — be coachable, show potential, align interests, be authentic, and invest in your relationship. This is a topic worth exploring in depth another time, but here are three actionable tips to help you get started right away:

1. Say good things about people who helped you behind their back.

2. Pay attention to when people go out of their way to help you, and call it out.

3. When you receive good advice from someone, implement it, then follow up with a thank-you note.

2. Your own contrarian product vision

The reason Steve Job’s “Here’s to the crazy ones” commercial is so timeless is because it’s true. Being a correct contrarian is a necessary condition for being valuable. A conformist has no role at the decision table.

A correct contrarian view is the product of a good mental model about your business. This takes time. Accelerate the process by learning the business inside and outside of your company, having strong opinions but weakly holding them, practicing your intellectual honesty, and most importantly, treating yourself as the leader of your business.

3. Scalable experimentation

How do you prove you're right? How do you make data the ultimate truth? How do you evaluate performance with quantitative evidence? Answering these questions strengthens the power of data.

Our competition includes:

• Convincing arguments from PMs who are trained to be persuasive

• Leadership opinions backed by authority

• Engineers who might dismiss you if unconvinced

Data and models alone can't win. With a high degree of freedom in choosing models and cleaning data, any conclusion can be supported. We all know how data can be manipulated. But, causal evidence from A/B testing is trustworthy.

Building the infrastructure, process, and culture for scalable experimentation isn't trivial, but modern tools make it easier than expected. At my company, Statsig.com, we’ve democratized Facebook's internal tools for clients like OpenAI and Notion, as well as startups and solopreneurs. I have a paper and a presentation to share technical insights.

Reach out to yz@statsig.com if you are interested.

If you enjoyed this article, give it a like so we know to write similar types in the future.

---

Thank you Yuzheng for sharing your clear outline and bold advice on how DS / DA (and other domains) can get their seat at the decision making table.

Get more of Yuzheng’s content on YouTube and connect with him on LinkedIn.

---

Connect With Ethan & Jason

• Follow Ethan on LinkedIn.

• Get Ethan’s career advice on YouTube.

• Connect with Jason (Ethan’s COO) on LinkedIn.

• Learn more about Ethan’s live online courses and on-demand courses.

• Contact us for corporate training, speaking, podcast appearances, and more.

• Maven’s Top Leadership & Career courses, enroll today to get 20% off:

  • Stuck at Senior Manager - How to Break Through to Executive - For leaders who want to grow to Principal, Staff, Director, or VP. (save $299)

  • Get Promoted Faster - For ICs & new managers (<10 years experience). Get tactical “how-to” insights for rapid career growth. (save $99)