From Analyst to Experimentation Leader: The Skills That Got Me Promoted

From Analyst to Experimentation Leader: The Skills That Got Me Promoted

There's a moment in most analyst careers where you realize that being good at your job isn't enough to advance. You can build the most elegant dashboard, write the cleanest SQL, produce the most insightful report -- and still find yourself stuck.

I know because I was there. I was a solid analyst at SVB, doing good work, getting good reviews. But I was reporting -- describing what already happened. The people getting promoted were the ones shaping what happened next.

The shift from analyst to experimentation leader wasn't a single promotion. It was a deliberate, multi-year process of building what I think of as T-shaped expertise: going deep in experimentation while going broad across analytics, behavioral economics, UX, and product strategy. Here's how that happened.

The T-Shaped Skills Approach

The concept of T-shaped skills isn't new, but it's underappreciated in analytics careers. Most analysts go deep in one area -- they become the SQL expert, or the GA4 specialist, or the Tableau wizard. That's the vertical bar of the T. It's necessary, but it's not sufficient.

The horizontal bar -- breadth across adjacent disciplines -- is what creates disproportionate value. When you understand behavioral economics, UX research, product strategy, AND analytics, you see connections that specialists miss.

My vertical bar is experimentation. I can design, execute, analyze, and communicate A/B tests with statistical rigor. That's my core competency.

My horizontal bar spans:

• Behavioral economics -- understanding why people make irrational decisions
• UX and information architecture -- how design shapes behavior
• Product strategy -- how features connect to business models
• Data engineering -- enough to build my own data pipelines when needed
• Team leadership -- managing cross-functional teams toward shared goals

I always believed that cross-discipline is where new knowledge is harvested. The most interesting insights I've had in my career came not from going deeper into one field, but from connecting ideas across fields that don't usually talk to each other.

The Arc: From Reporting to Leading

SVB: The Reporting Phase

At Silicon Valley Bank, I was an analyst. I built reports, maintained dashboards, answered questions from stakeholders with data. I was good at it. And I was slowly becoming irrelevant.

Not because the work didn't matter -- it did. But because reporting is inherently backward-looking. You're describing what already happened. The strategic value is limited because by the time you've reported on a trend, the organization has often already moved on.

The pivot came when I started asking "so what?" about my own reports. Revenue dropped 3% in Q3 -- so what? What should we do about it? That question -- "what should we do?" -- is the bridge between analyst and strategist.

I started appending recommendations to my reports. Not just "here's what happened" but "here's what I think we should test based on what happened." Some of those recommendations were naive. But the act of making them changed how leadership perceived me -- from someone who reflected data to someone who had opinions about it.

AdMixt: The Experimentation Phase

At AdMixt — an adtech company that ran paid marketing for clients across Meta, Facebook, Instagram, Pinterest, and TikTok — I learned experimentation through sheer volume. The team was constantly testing audiences, creative variations, and bidding strategies, even if we didn't call them experiments.

Different headlines, different audiences, different creative, different landing pages -- all tested, all measured, all optimized. The pace was relentless, and it taught me something that no certification could: intuition for what's likely to work.

That intuition isn't magic. It's pattern recognition built from hundreds of experiments. After you've seen enough tests, you start to develop a sense for which changes will move metrics and which are cosmetic. That sense doesn't replace rigorous testing, but it makes you dramatically better at prioritizing what to test.

Agency work also taught me client management, which turned out to be directly transferable to stakeholder management in enterprise settings. Explaining to a client why their pet idea lost an A/B test is basically the same skill as explaining to a VP why the data doesn't support their preferred strategy.

NRG: The Leadership Phase

At NRG, I wasn't just running experiments -- I was building the experimentation program. Setting the roadmap, establishing the methodology, training teams, building the culture.

This required a completely different skill set than individual contribution. Suddenly, I needed to:

• Sell experimentation to leadership. Not everyone believes in testing. Some leaders trust their intuition over data. Making the case for experimentation requires understanding their language and concerns.
• Build a testing roadmap that balances quick wins and strategic bets. Too many quick wins and you're optimizing buttons. Too many big bets and you won't show results fast enough to keep support.
• Establish statistical standards. When is a test done? What confidence level do we require? How do we handle tests that are directionally positive but not significant? These decisions shape the entire program.
• Train non-technical teams. Product managers, designers, marketers -- they all need to understand experimentation basics to participate meaningfully.

The transition from "person who runs tests" to "person who builds and leads a testing program" was the biggest career jump I've made. And it required almost entirely different skills than the ones that got me my analyst roles.

The Certifications and Learning That Mattered

I've invested significantly in continuous learning. Not all of it was equally valuable. Here's what actually moved my career:

Mindworx Behavioral Economics (Ogilvy Group UK)

This was transformative. Behavioral economics gave me a framework for understanding why people behave the way they do, not just what they do. Concepts like loss aversion, anchoring, default bias, and social proof became practical tools in my experimentation work.

Before this certification, I was testing based on best practices and intuition. After it, I was testing based on psychological principles with decades of research behind them. My win rate on experiments went up noticeably -- not because the tools changed, but because the hypotheses got sharper.

CXL Institute CRO Certification

CXL's program is rigorous. The statistics module, combined with later sessions I had with Optimizely's statistics team, gave me a level of quantitative confidence that most CRO practitioners don't develop. I stopped relying on "the tool says it's significant" and started understanding why it's significant -- and when the tool might be wrong.

Optimizely Statistics Team Sessions

These were informal but invaluable. Understanding the math behind experimentation platforms -- sequential testing, false discovery rates, multiple comparison corrections -- made me a more rigorous practitioner. It also made me better at communicating with data science teams, because I could speak their language.

DataStory Workshop

Data storytelling is the bridge between analysis and action. You can have the most rigorous analysis in the world, but if you can't communicate it in a way that drives decisions, it dies in a slide deck. The DataStory workshop taught me narrative structure for data presentations -- how to build tension, create clarity, and drive toward a specific decision.

This skill has been disproportionately valuable. The analysts who get promoted aren't always the most technically skilled -- they're the ones who can make leadership understand and act on their findings.

How Data Analytics, Experimentation, and Behavioral Economics Connect

These three fields are often treated as separate disciplines. They shouldn't be.

Data analytics tells you what is happening. Traffic is down, conversions are dropping, users are churning at day 7.

Experimentation tells you what works. This new onboarding flow increases activation by 12%. This pricing page layout increases plan selection by 8%.

Behavioral economics tells you why it works. Users respond to the new onboarding flow because of the progress bar (goal gradient effect). The pricing page works because the middle tier is positioned as the default (default bias) with the most expensive tier creating an anchoring effect.

When you combine all three, you get a feedback loop: observe behavior, hypothesize about the psychology driving it, test your hypothesis, measure the result, and feed that learning back into your understanding. Each cycle makes the next one more effective.

This is the approach I brought to building Jobsolv. The entire product is built on behavioral design principles -- understanding that job search is as much a psychological challenge as a technical one. You can see more about this framework at atticusli.com/framework.

Learning From Every Role

One thing I want to emphasize: there's no wasted experience if you're paying attention.

My economics degree gave me systems thinking. Financial analysis taught me to connect metrics to money. AdMixt taught me speed and how to run performance marketing across every major social channel. SVB — where I was a Senior Growth Analyst on marketing data analytics — taught me enterprise data and stakeholder navigation. NRG taught me program building and leadership.

Even the roles that didn't feel relevant at the time contributed something. Facebook Ads experimentation at an agency isn't the same as running A/B tests on a product, but the underlying logic -- form hypothesis, test, measure, iterate -- is identical. Client management is stakeholder management. Financial modeling is business case building.

The common thread: I treated every role as an opportunity to add another skill to the horizontal bar of my T.

Advice for Analysts Who Want to Lead

If you're an analyst right now and you want to make the transition to experimentation leadership, here's what I'd prioritize:

Start recommending, not just reporting. Every analysis should end with "and here's what I think we should do about it." This changes how people perceive you.

Learn behavioral economics. It's the single highest-ROI investment for anyone in CRO or experimentation. Mindworx is where I started, but there are many good resources.

Run experiments, even small ones. You don't need a formal experimentation program to start testing. Find one thing to test, design it properly, run it, share the results. Then do it again.

Build cross-functional relationships. Experimentation leadership requires working with engineering, design, product, and marketing. Start building those relationships now, before you need them.

Learn to tell stories with data. Practice presenting your analyses as narratives, not slide decks full of charts. What's the insight? Why does it matter? What should we do?

Invest in continuous learning. The field moves fast. Certifications, workshops, conferences, reading -- keep learning. The moment you stop, you start falling behind.

The path from analyst to leader isn't about getting better at analysis. It's about becoming the person who shapes what gets analyzed and what gets done about it. That requires breadth, communication skills, and the courage to make recommendations that might be wrong.

You can learn more about my career trajectory and approach at atticusli.com/about, and explore the experimentation framework I've developed at atticusli.com/framework.

Atticus Li is the founder and CEO of Jobsolv, an AI-powered job search platform that has helped 30,000+ users land interviews. He's built small, effective startup marketing teams, led product development end-to-end, and is AI-native — building software himself with AI tools and adapting quickly to new ones. At SVB he was a Senior Growth Analyst focused on marketing data analytics; earlier in his career, he worked at AdMixt, an adtech company, running paid marketing experimentation for clients across every major social channel (Meta, Pinterest, TikTok). Reach out at [email protected].