The unwritten rules of being a head of data

A friend of mine became head of data but they had a hard time stepping away from coding. They kept shipping code late at night after their normal work ended.

This was fine at first until they realized, months later, they hadn’t really led anything of note for the business. There was nothing they had really driven. 

They had been promoted, but still were working like a senior engineer. Delivering lots of well-designed work but not actually taking a stand on any specific project.

Once you become a data leader, the game changes. What the business needs from you isn’t just to write another Python script or notebook. But that’s not always made clear. Some companies or mentors will help guide you as you become a leader. Others will give you the promotion and then hope you figure it out.

In this article, I wanted to discuss some of the unspoken rules and lessons you need to pick up fast if you are a data leader.

You are no longer an IC

Maybe this sounds obvious, but actions speak louder than words. Far too many new data leaders I’ve worked with don’t act like they are no longer an IC. Your job as a data leader is no longer to build data pipelines and dashboards.

You’re shifting from execution to planning and making decisions. Now, for those who are reading this and are a team of one, well, then you are doing both. 

But in many cases, you’re likely leading a team. That means you need to shift from trying to fight every fire, as tempting as it might be. Especially if you spent the last few years or decades delivering that type of work, it’s a massive mindset shift to no longer consider writing SQL the most valuable work you can be doing.

You need to be willing to hand over the reins and execution to other people and realize they might not solve every problem the way you would. That’s part of it.

You are no longer the IC. Sure, you can set the tone, you can set standards, and own projects. But how it is actually delivered, how each line of code is written, won’t always be exactly what you expect. So you need to step back and let other people do their jobs.

Building support and buy-in isn’t always easy

Lots of businesses like to say they are data-driven. But when you actually look at how they use data, that comes into question. We, as data leaders, are the data experts and it’s up to us to both drive the data strategy as well as get buy-in for it.

The problem is, we can often focus on the wrong things that detract from getting buy-in. 

We talk about topics the business doesn’t understand the value of (and often doesn’t care about), like: should we use a data lakehouse or a data warehouse? Or should we use tool A or B? These types of conversations don’t make it clear what the business impact will be.

But we can make them care about the work the data team does and what it can do for the business. We just have to shift how we approach our conversations. 

For example, imagine your operations team is struggling with late deliveries and wants to understand where the bottlenecks are. As data leaders, we might start by debating how to store the operations data or maybe what chart to use, but that’s not what the business and operations teams want to hear. They need help improving their late deliveries and understanding their bottlenecks.

That’s why we need to start by thinking about the business.

• What do they care about?

• What outcomes do they care about?

• What’s holding the business back, data related or not?

Take those goals and turn them into your technical solutions. Sometimes that means fancy ML, other times you might be able to deliver a straightforward analysis that helps shift the way your stakeholders are approaching their department.

In the end, you have to make them care. You have to get them to see the value, because it won’t just happen because you’ve built a dashboard.

You’ve only got so many at-bats

There is no sugar coating it. You only have so many projects you can take on and so many times you can be wrong. This refers to more than just data quality. You also need to be wary of just handing off a dashboard that is perfect technically but doesn’t drive much in terms of business action. Eventually the lack of impact and lack of usage will eventually get noticed.

This means you need to prioritize ruthlessly.

Take on work that aligns with the business or has a clear end goal. Sometimes, that end goal is to get further buy-in for a larger project, other times it's clear business value. But be clear and celebrate your outcomes.

More than once!

Don’t just complete a project and hand it off. Where your work stops is often only where business value starts. 

So be intentional. Deliver less, but deliver things that matter.

As a data leader, your success isn’t measured by the volume of outputs; it’s measured by the impact those outputs have. When you’re selective, strategic, and clear about what you own, your “at-bats” turn into wins that actually move the business forward.

Be choosy about your stakeholders

Plenty of stakeholders will ask for your help on projects and ad hoc data requests. However, not every stakeholder will help you succeed. 

Not every stakeholder will support you. In fact, they might just see you as a cost center and not view data work as helpful to the business. My tip there is don’t waste cycles chasing people who consistently block or dilute impact. Specifically if they don’t have much influence on your work or career.

On the other hand, there will be stakeholders who clearly want to work with data. They need data to actually help them make better decisions and they want to be your champion. Those are the people you want on your team. And when you deliver, and other teams see it, they’ll also want to work with you. Invest deeply in these relationships: co-design projects, share credit, and let their wins showcase your team’s value.

Final thoughts

It’s easy to fall into the trap of thinking that the actions you took that made you a good IC will also make you a great data leader. After all, it got you to where you are. However, I think once you make the shift, it’s good to have a few pattern interrupts. That is, you need to actually pause and ask yourself, What do you want your data team to actually do?

Stick yourself in the business’s shoes, think about something other than SQL and pipelines for a moment. Will you still have to fix the occasional Airflow error or parse through a Hex notebook? Yeah probably.

However, you’re also going to find yourself busy managing relationships, budgets, and expectations. All of which take up time and drive value in totally new ways that require you to learn and strengthen very new skills. 

Whether you’re just starting out as a new data leader or have been one for a while, working with people and businesses is a skill that can always be further refined, and some of the biggest challenges are that what worked yesterday at company A might not work at company B.