TheVentureCity Data Team
TheVentureCity data team, June 2023. All of these people are curious and persistent.

In the beginning, it was just me, a solo data scientist dude working with the rest of TheVentureCity team to figure out how to use data to our advantage (please see my previous post for a full discussion of my approach). I was lucky enough to have some strong engineering colleagues (h/t Roberto Carlos Navas and Juan Ramiro) to help me achieve some early impact, and soon we decided that I could use some help. So we hired some people for me to lead. That’s how I ended up founding a data team.

Founding, building, and leading the data team at TheVentureCity involved more than just an understanding of how to apply data analytics to startup investing. It also required building a culture of trust, openness, and continuous improvement. My journey as the leader of a data team taught me valuable lessons about what truly drives a team towards excellence. Here’s my best attempt to break down those lessons.

Hire Curious People

The job description we posted contained some gobbledy-gook about the company, the responsibilities for the position, and the skills required to be successful. But all of that was wrapped around what I considered to be the core of what I was looking for, a list of seven personal characteristics that define a good data analyst or engineer:

  • Consultative, collaborative, problem-solving approach to working with teammates and the companies we support
  • Curiosity
  • Perseverance
  • Love of learning
  • Capacity to work autonomously and be accountable for deliverables
  • Ability to handle multiple concurrent work streams
  • Affinity for working on a geographically distributed team

During job interviews, I most heavily emphasized and looked for the trifecta of curiosity, perseverance, and a genuine love of learning in candidates. Nobody knows all there is to know about data analytics or data science and engineering. It’s simply too big a field, and it changes frequently. To build a great team over the long haul, we needed people who were “down for whatever:” ready to help founders and build infrastructure without thinking they already knew it all. A good data professional can always get better.

Give Them Space for Deep Work

Our team needed to deliver results from cognitively demanding tasks. This means we needed to quickly master hard things and produce at a high level, both in terms of quality and speed. We all needed uninterrupted time to get into a problem and stay there, without any context switching. In his book Deep Work, Georgetown computer science professor Cal Newport says that deep work is not just a routine habit but a skill that needs to practiced and strengthened. If you get good at it, he says, “cultivating a deep work ethic will produce massive benefits.”

Deep Work, by Cal Newport
Deep Work

I saw it as my role to help our team members manage their calendars to find four-hour blocks of uninterrupted time as often as possible. The team managed work through a Kanban board from which we could pull work, rather than have it pushed to us, giving us more control of the inflow. We all familiarized ourselves with our phone settings to minimize or eliminate alerts during deep work time. In our case it helped to have half the team in Europe and the other half in the Americas. We could meet during the times when our schedules overlapped and go deep when the other side of the Atlantic wasn’t working.

Embrace Self-Management and Trust

My goal in managing our team was to help it reach “self management,” where the team members coordinate their own activities and make decisions without a central authority. As the manager of self-managed team, my role would be that of a servant leader, removing roadblocks and setting ambitious goals. We understood that for our team to manage itself effectively, every member needed to feel empowered and trusted. This trust extended to handling mistakes and failures positively. We fostered a culture where mistakes were seen as opportunities for learning and growth, believing that each error helped us refine our methods and deepen our ownership of projects.

A team’s degree of self-management exists on a continuum. I believe we moved far down that continuum on the way towards “full” self-management without getting completely there. But we became a much better team in the process of continuous improvement required to get as far as we did.

Give and Receive Feedback

Many moons ago, I was lucky enough to work with a truly great colleague, the late Adrian Wible, who taught me about the importance of feedback when working as a team. A healthy feedback culture is the best way to build trust. Adrian later consolidated his thoughts in a Feedback Manifesto that is well worth anyone’s time to read. Our team culture emphasized both giving *and *receiving feedback as skills to be developed at every opportunity. We conducted regular retrospectives where everyone was encouraged to provide candid feedback. This process was always about focusing on behaviors, not individuals. These sessions were crucial for us to adapt and fine-tune our processes continuously.

(For more about the importance of receiving feedback in particular, I recommend this from organizational psychologist David Burkus, PhD and this from former Facebook executive Sheryl Sandberg.)

Deepen Personal Connections Through 1:1’s

Another golden opportunity to give and receive bidirectional feedback were the regular one-on-one meetings I held with each team member. I wanted to understand them as people: their motivations, hopes, and dreams. These discussions were instrumental in crafting roles that not only mapped to each person’s skills and aspirations but also offered them challenges and growth opportunities. We didn’t limit these exchanges to just the managerial level; team members were encouraged to engage with each other, fostering a supportive network.

Cultivate a Fun and Productive Environment

Inspired by Ali Abdaal’s concept of “Feel-Good Productivity,” I aimed to create a work environment that was not just about getting tasks done but doing so in a way that was enjoyable and fulfilling. Abdaal says that fostering a sense of fun is the secret to unlocking our learning, creativity and productivity while reducing stress.

At one point we attempted to make “have more fun” one of our half-yearly objectives, as part of the Objectives and Key Results (OKR) goal setting framework. Ironically, that experiment didn’t last very long when we discovered that trying to measure and implement such a goal was the opposite of fun.

Instead of formalizing fun, we kept the atmosphere upbeat and incorporated light-hearted elements like nicknames. That’s where my “David the Data Dude” moniker comes from (everyone on the team got a similarly alliterative nickname). Our weekly meetings always included reminders about why we were there: to have fun and excel at our work. We started asking each other, “What would this <arduous task> look like if it were fun?”

TheVentureCity Data Team
What would this staged team picture look like if it were fun?

Link a Mission to the Work

From the onset of the data team, I articulated a clear mission: Become the best data team in the world of VC investing. Hey, why not? Why do anything if you don’t think you can eventually become the best in the world at it?

This bold vision didn’t just serve as a guiding star for our work and goal-setting; it instilled a sense of pride and purpose in everything we did. It helped us stay focused and motivated, pushing us to set high standards for ourselves and continually strive to meet them. And it meant continuously optimizing for delivering value to our founders and internal investment teams.

Recognize and Advocate

Recognizing achievements, especially for those in less visible roles, was key to maintaining morale and motivation. We celebrated not just major successes but also the small wins along the way, making particular use of the wonderful “shout-outs” portion of TheVentureCity’s weekly 360 meeting–15-20 minutes of team members praising and thanking each other! Behind the scenes, I advocated for raises and promotions, ensuring that team members felt valued and fairly compensated.

Automate and Reduce Fragility

Our technical strategy focused on automation, rigorous unit testing, and setting up alerts for anomalies. These practices not only saved time but also reduced errors, allowing us to focus on higher-value activities. We loathed repetitive tasks and sought to automate everything we could.

Pay Close Attention to How Long Tasks Take

We consistently found it difficult to estimate the time required to perform analytical tasks with data. Most of the time, you don’t know what you don’t know. Some days you might get a dataset with European dates mixed with American dates and it will drive you crazy for an hour or more. Other days, you might explore super-clean data that inspires lots of lines of inquiry that you didn’t anticipate. This uncertainty is difficult to manage. Here are some things we did to mitigate the problem:

  • Prior to starting, decide on a timebox: a fixed number of hours (or days or weeks, depending on how big it is) to allocate to the project. Once the time box expires, assess where you are. You may decide you need another timebox, or you may decide you’ve done enough. Projects (and blog posts, lol) have a tendency to take as long as you give them. Timeboxing helps use that tendency to your advantage.
  • Remember that 80-90% is often good enough, and that last 10-20% may not be worth the marginal time you have to spend on it. Data scientists are prone to overcomplicating things, but knowing when to say “when” can have a positive impact on the team’s overall impact.
  • Allow for the subconscious mind to work its magic. We all know the feeling of having breakthrough ideas in the shower or during a walk outside. Use this phenomenon to your advantage and strike a balance between driving towards deadlines and letting an analysis “breathe.” It’s okay to put something down and return to it later with fresh eyes.
  • Leverage the collective wisdom of the team. In team meetings we would often recognize some prickly problem that someone was stuck on and that needed more thought before diving in. We would agree to schedule a separate meeting with everyone who had the knowledge of and interest in the issue. A 15-45 minute meeting could often set the person primarily responsible on a clear path to delivering value rather than spinning her wheels.

People Over Pixels

Our team’s internalized and personified TheVentureCity’s mantra of “people over pixels”—a reminder that while deliverables and deadlines are critical, the team’s well-being and culture are what truly drive success. By fostering a supportive, engaging, and trust-filled work environment, we didn’t just meet our goals—we created a workplace where everyone was excited to come in every day. I am proud of the work we did and the way we did it together.

Special Thanks

  • Andriy Radich
  • Garoe Gonzalez
  • Jon Ardinast
  • Juan Pablo Treviño
  • Juan Ramiro Meyer
  • Katya Skorobogatova
  • Laura Gonzalez-Estéfani
  • María Dancausa
  • Mario Cantelar
  • Mercedes Plaza
  • Roberto Carlos Navas
  • Santiago Canalejo
  • Yannick Ruby

Our little group it’s always been

And always will until the end

– Nirvana, Smells Like Teen Spirit