Advancing Your Career in Data Analytics: Lessons from 20 Years in the Field

After spending two decades in data analysis and leading data teams, I've learned some lessons that most courses and certifications never teach you. Let me share what actually makes the difference.

The Real Challenge Isn't What You Think

One of the most complex parts of analytics isn't necessarily the messy data. Everybody focuses on the messy data: "If only we move to this new database, it's all going to be solved." Or "this new product will fix everything."

I don't think that's true. Messy data will never be solved entirely. That's why many organizations have the medallion architecture, the gold layer, the silver layer, and the bronze layer, where you always have some messy data. It's just a fact of life.

What many of us don't focus on, and maybe don't want to focus on, but definitely should, is the business. Understanding the business needs. Understanding your stakeholders in the organization.

What does the marketing team care about? What do the different sub-teams in marketing care about? What do the salespeople care about? Product support? What do all these organizations care about? What are their interests? What acronyms do they use?

Document this information. Make it part of your knowledge base. Don't just document the data model. Don't just document "this is how we tackle data." Also, document "this is how we tackle the organization."

When you bring in an AI tool, having that context will be tremendously helpful. It can take any shape or form—a Google Doc, whatever works for you. But having this context will make these AI tools incredibly powerful.

I strongly recommend investing the time in this.

The Shift That Changed Everything for Me

The significant shift for me, after managing data teams and working as an analyst for many years, came when a guest speaker visited our company. He was the CEO of a successful data company. He had inherited it and realized that the core product wasn't what they thought it was. The company's core product was data.

He challenged us to consider: Is data as important as the core product you're selling?

Take Amazon, for example. The core product they're selling is a range of goods shipped to you. But their second product is the pattern of what people are buying and what they care about: what they want to see, what they want to buy, and what they might want next.

Data for many organizations is the second product. And it might be as necessary, or even more important than the main product they're selling, because it provides competitive differentiation.

For me, as a leader of the data organization, this was transformative. This mindset change fundamentally altered how I behaved in the organization. From that day forward, I took a step back and thought: Maybe I should have a strategy for how we handle data and how we make the organization mature with it. How should we set up the team?

Seeing stakeholders in the company as my own customers was a significant shift.

For anyone looking to advance their career, I recommend a fundamental mindset shift: see yourself not as a passive listener or recipient of requests, but as a product owner who manages data as a product within your organization.

This means you need to:

• Understand your customers (stakeholders) - What are their pain points? What keeps them up at night?
• Build a roadmap - What capabilities will you deliver and when?
• Prioritize ruthlessly - Not every request deserves immediate attention
• Measure impact - How is your "product" driving business value?
• Iterate based on feedback - Continuously improve based on what's working

This shift from order-taker to strategic product owner changes everything about how you operate and how you're perceived in the organization.

Using the Analytics Maturity Framework

One tool that really helped me be seen as strategic was the analytical maturity framework. You can Google it—it's widely available.

Imagine two axes: value to the insights and level of maturity.

What is happening: What was revenue last week? How many new customers did we get? What was our churn? These are "what" questions.

Why is it happening: Why was revenue up? Why was it down? You use cohort analysis and anomaly detection. Asking why moves you up on the analytical maturity curve. Getting the "what" is the foundation; getting to the "why" advances you.

What if scenarios: What if we lower our price on this product? What if we don't offer as steep a discount in Southern California? Now you get to real ROI situations. You need the foundation of what's happening and why it's happening to get to the "what if" optimization.

I've used this framework for C-level conversations because it's very helpful to explain: "We're currently at the 'what' level. If we want to reach the level that Uber or Amazon operates at, we need to put pieces in place to get to that level of maturity."

You can break it down by department. How mature is each department on that what, why, what-if curve? And what needs to be true for them to advance?

By breaking it out by department, you become very strategic. You have a plan. You have your six different departments. You have a roadmap for each one of how you will mature them.

Now you're a leader in the organization.

Where to Start: The Squeaky Wheel

One question I get frequently is: "Who do I start with?"

My personal recommendation—though some people disagree—is to start with the squeaky wheel. The person who complains a lot and is quite vocal about it.

Because that's someone who talks extensively about their needs, and even if you can help them a little bit, they will talk about that as well. They will talk about how this has helped them. And that is what you need internally. You need support from people who talk about your contributions and give you additional momentum.

There's some risk, of course, in going with the squeaky wheel. I would first check: do you even have the data available to help this person? Is this accessible to you? Is this super complicated? There are several criteria to consider.

But I would focus on the person who is the most vocal about their needs. Having someone who talks extensively about your success and contributions will greatly benefit you.

Start with that person and then move to the second—perhaps not the squeaky wheel at that time, but maybe someone who's truly impactful in the organization and widely respected.

I typically start with someone in sales or marketing, then move to someone in customer support or product.

And one thing I highly recommend is putting an ROI number together.

Communicating Success and ROI

One thing that did not come naturally to me at all, and I still need to work on it, is managing upwards and communicating success.

I always thought you did good work and that it would be recognized. But the truth is, that's just not reality. I still believe good work should be recognized on its own merits, but the reality is that the person who communicates their success effectively will benefit far more.

One thing you can do in the data space that others will struggle with is articulating the ROI. If you helped someone with something, there's often a dollar figure attached. It doesn't have to be exact math. It can be approximate. But everyone will appreciate that you thought through your contribution to the organization.

This will help you tremendously at the leadership level.

Communicate the approximate ROI. And also get into a cadence of perhaps once a month, announcing: these are the new features we have released, the new capabilities we now have. Maybe we can now help the marketing or customer support teams with analytics.

Highlighting progress—even if people don't necessarily read the newsletter—lets the CEO know something is happening consistently. Twelve times a year, your team sends updates on progress made. That's how you build substantial support.

I highly recommend this approach. And if you can include an ROI, you're really onto something.

Why This Is the Right Time to Embrace AI

Is now the right time to embrace AI? Well, we are three years past ChatGPT's launch. Some people say, "My job hasn't really changed significantly." For others, it has changed dramatically.

I used to hire people to take notes, identify action items from each meeting, and follow up. All of that is now done by AI.

The parallel I often think about is 2007, when I was an analyst—this was not part of my job—but I was asked, "Hey, we're thinking of maybe going into the cloud. It's this new thing, what are your thoughts?" I did some research on what it would take to migrate to the cloud.

I did my first cloud migration in 2007, then another in 2009, and another in 2011. By 2011, I had done three cloud migrations. I never thought of that as a skill set, just something I was asked to do.

But in 2011, that was an incredibly powerful skill to have. Redshift came out. BigQuery was still early. Snowflake didn't exist yet. Having done multiple cloud migrations in 2011 was a really interesting moment in time. We spun up our Redshift cluster, I think in 2012, and I was already quite familiar with the technologies behind it.

I think of this moment in the same way: the train has undoubtedly left the station. AI is moving much faster than cloud migration ever did.

AI won't take your job, but someone who embraces AI and uses it every day and builds that skill set might.

I use AI every day for everything from meetings to content, email responses, and data analysis. I barely open a spreadsheet myself anymore and do the number crunching. I use tools that think things through for me. They triple-check everything. And this comes from someone who has spent 20 years in data analysis.

I would say use AI wherever you can right now, because every year of experience you gain, every product you master will give you an advantage.

What's Wrong with Traditional BI Tools

What's wrong with old-school business intelligence tools? There are several issues.

First, they're often harder to use. There are all kinds of buttons and things to push. Having a tool you can talk to natively, or one that anticipates and suggests ideas for you, is fundamentally different.

The other issue is that I think BI tools have never truly understood analytics. The builders of BI tools are engineers. They built these things. But the nature of analytics, the nature of insights, is to drive change in an organization.

As soon as you create a report, within three months, that report should almost become obsolete or at least change significantly. When you create a dashboard, some core KPIs won't change, of course. Revenue will always be necessary. Customer count will always matter.

But underneath, the "why" and the "what if" should be changing and evolving with you.

Traditional dashboards and BI tools struggle with this. They were not built for that. They were built for static, locked-in content. The result is dashboard rot. You have all these dashboards that are no longer used.

Every time I came into an organization—and I ran the professional services organization for Looker for many years—you see hundreds of outdated, unused dashboards or explorers in the Looker ecosystem. It's because it's not evergreen content. It doesn't evolve. It wasn't built to evolve with a business.

The Data Literacy Challenge

One challenge we consistently faced is data literacy. You show someone a dashboard, and they see that revenue was up. Nobody will take action because revenue was up, implying there's nothing we need to do.

But maybe revenue was actually down for their part of the organization, their span of control. They wouldn't see that if they just looked at the top-line numbers.

One of the challenges I've seen repeatedly is a wide gap in data literacy. The way modern AI tools work is that you don't have to ask the questions. They will understand your span of control, then ask the why, peel back the onion, and go multiple layers down for you. Then they say, "Revenue was up, but actually, for you as a marketing associate, it was below what it was the previous week. And here's what you can do about it."

I once worked with a CEO who didn't tell me that they didn't quite understand what these fancy graphs and numbers meant. I kept an eye on usage and realized the CEO wasn't using the dashboard. I finally sat down with him and said, "Look, I know you're not using this. What's going on here?"

He finally told me, "I just don't quite understand it." I walked him through how to read the pivot table and how it relates to his job as a CEO.

With AI tooling, it's very different. AI sends the insight to stakeholders, and they can ask questions in the privacy of their own computer: "I actually don't quite understand this. Can you explain this to me?"

There's a tremendous opportunity here to close a massive gap in the adoption of analytics.

The Path Forward

Advancing your career in data isn't just about mastering the next tool or programming language. It's about:

Understanding the business: Document what your stakeholders care about, not just your data model

Thinking as a product owner: See data as your product and stakeholders as your customers

Using frameworks strategically: The analytics maturity framework gives you a language to speak with leadership

Building momentum: Start with the squeaky wheel, deliver wins, let advocates spread the word

Communicating value: Calculate approximate ROI and share monthly updates

Embracing AI now: This is your 2007 cloud migration moment

The technical skills are table stakes. What differentiates your career trajectory is how well you understand the organization, how effectively you communicate impact, and how you position yourself as a driver of change.

Data might be your organization's second product. But for advancing your career, these insights should be your priority.

Lucas Thelosen has over 20 years of experience in data analysis and leadership, including running professional services at Looker and building data organizations across multiple companies.