So you want to work in data

I’m often asked how to break into data (weird flex, I know). While there’s a ton of content on this topic, providing commentary is still valuable for a few reasons.

For one, many follow a unique, circuitous route to a data role, and sharing these experiences may encourage others to follow a similar path. For example, I’ve had colleagues on operations and customer success teams leverage their responsibility as “the only person on their team who did data stuff” into a full-time data analyst role. More traditionally, data professionals have technical backgrounds in computer science, finance, or management consulting, as these fields provide the generalist, logic-driven skillset required in data roles. However, data-specific programs have been popping up at universities (shoutout to UVA’s School of Data Science), and boot camps are increasingly more popular - both lower the entry barrier for data roles for students and those pursuing mid-career transitions. In summary, there’s not a one-size-fits-all approach.

Secondly (and maybe more importantly), the way to “break in” is counterintuitive. Most junior-level data analysts and mid-level career jumpers looking to develop their data chops focus on becoming experts in [enter coding language, data tool, etc.]. Several others have commented on this,¹ but people often forget that data tooling is a means to an end, not the end itself, and focusing solely on technical skills will not get you far. Communication, stakeholder management, and product management are far more desirable skills and more difficult to hone. As a result, it’s important to guide your skill development and job search with this frame of reference in mind.

My professional journey

I started my career at a boutique management consulting firm focused on marketing analytics, customer segmentation, and price elasticity. Our bread-and-butter was marketing mix attribution (MMA) models, and these two to three-month long engagements sharpened my skills in stakeholder management, project management, and data modeling.

However, I quickly realized a few things:

1. I found the technical aspects of the role particularly enjoyable: building an end-to-end data product that delivers value. On the flip side, client management can be a terrible beast.

2. I valued building a highly scalable data product over the repetitiveness of ongoing, high-touch consulting work.

3. Navigating the internal bureaucracy of massive corporations is a skill in and of itself (and a highly valuable skill at that).

Fortunately, I leveraged the skills I’d learned (even Microsoft SQL Server; yikes) into an in-house marketing analytics role for a growing eComm startup. Further, after wedging my foot into a data role, I picked up responsibilities across business functions (Product, Finance) and along the data lifecycle (ingestion, warehousing). I gained broad exposure to people, processes, and problems across the business while refining both hard and soft skills.

This exposure led to my most recent position leading a data team at a growth-stage software startup. My role ranged from designing and building the initial data infrastructure and core models, product managing embedded data products, and most importantly, managing and developing a high-performing data team.

Guide to landing your first data job

If I were entering the workforce or looking to pivot into a data role today, there are a few areas in which I would invest my time and effort.²

Build and write in public

Building a public portfolio in GitHub, via a newsletter or blog, etc., demonstrates a strong interest in the field and the ability to complete projects from end to end. From hypothesis development to data sourcing, ingestion, modeling, and visualization, you will showcase your ability to replicate a project similar to one within an organization. This involves clearly and concisely communicating a problem statement, tradeoffs, and an analysis of results in a way that resonates with your audience. Much of my writing involves DIY tech, and The 2024 U.S. Open (source analytics) is one example that would fit particularly well in a data analyst portfolio.

While tutorial learning is important, nothing can replicate practice. Don’t get sucked into tutorial hell - go out and build!

Learn hard skills, but showcase soft skills

Most entry-level roles will require a rudimentary understanding of SQL and the ability to navigate a BI tool. There are plenty of free resources available to build a basic technical skillset, such as:

1. Practice problem sets on W3Schools or HackerRank

2. Hands-on learning tracks on CodeSignal

3. LLMs to ask clarifying questions

However, like most technical positions, strong technical chops will only get you so far. As you develop to senior levels, attributes like stakeholder management, ownership, and communication are more relevant. Therefore, when building out a public portfolio, double down on:

1. Problem Identification: why does this project or analysis matter? Why are you interested in this topic?

2. Communication: do you explain technical concepts clearly and concisely, prioritizing storytelling?

3. Tradeoff Evaluation: did you sacrifice quality for speed, and if so, how did you come to that decision?

rather than building highly complex SQL models.

Connect with other data professionals

The Federal Reserve Bank of NY found that referred candidates are more likely to be hired and referred workers experience an initial wage advantage.³ The takeaway here: touch grass every once in a while (read: talk to people). I highly recommend two data communities, Locally Optimistic and dbt, as venues for connecting with others in the industry. Not only are jobs posted in these communities daily, but they also provide a space to request coffee chats with individuals whose backgrounds you’re most interested in and stay up-to-date on the latest data trends. I’ve used this template to connect and learn from folks with different backgrounds:

Decomposing the template:

1. Personalized and up front “why” - demonstrating your interest in their work with specific examples and a straightforward ask.

2. "Friday at 3pm ET” - providing a time eliminates scheduling friction (shoutout Atomic Habits). Also, people are less likely to be busy on Friday afternoons.

3. “Venmo you for coffee” - people appreciate the thought of offering to pay for their time, even if most will not accept this gift.

While your hit rate might be small, the more outreach you send, the likelihood of connecting increases.

Closing thoughts

Breaking into data isn't about following a predetermined path - it's about leveraging your unique background while showcasing a strong mix of technical and interpersonal skills. The most successful data professionals I know didn't necessarily start in data, but they all share a few key traits: curiosity about solving problems, ability to communicate complex ideas simply, and willingness to learn continuously.

Start small but start now:

1. Pick a topic you're passionate about and build something with data, even if it's simple.

2. Share your work and thought process publicly, whether through a blog, GitHub, or LinkedIn.

3. Reach out to one data professional today (yes, today).

Every experienced data professional started as a beginner. A non-traditional background isn't a limitation - it's an advantage that brings a unique perspective to approaching data problems.

1  Ethan Aaron (Portable), Benjamin Rogojan (SeattleDataGuy), and Solomon Kahn (Delivery Layer) hit this theme often.

2  I’m mainly referring to a data analyst or science role, but these concepts also apply to data engineering and analytics engineering roles.

3  “Do Informal Referrals Lead to Better Matches? Evidence from a Firm’s Employee Referral System” (link)