How to Build a Successful Data Team (Without Searching for a Unicorn)

You can’t go into battle by yourself; it takes an army. The same is true with a successful data project. The fact that we’ve focused so much on finding unicorns, or individuals who can do it all, is foreign to me.

Personally, I can do a lot of engineering tasks and scientific algorithms but am not as good at visualizing things.

There was a team I was on that had excellent engineering and data scientists on board. We spent 6 months working on beautiful solutions to problems that we thought the rest of the team needed.

After 6 months we were called into the CEO’s office to present our results.

Problem was we didn’t have any, we had a pile of useless beautiful code and models. Long story short, that project was killed that day. 6 months of work down the drain!

That’s why today I’m going to be talking about putting together a successful data team; who should get on the bus? While things might change depending on the specifics of your organization there are four important archetypes on any data team.

Before I explain what the four archetypes are, let’s talk a little bit about what archetypes are. Carl Jung is a psychologist who came up with the modern interpretation of archetypes. He defined them as highly developed elements of the collective unconscious.

Basically they’re traits that are universal. A common example would be “the hero” which has been chronicled for millennia. Nobody is 100% any of these archetypes.

So for instance nobody is 100% “hero” (even though they might want to be), that idea of the “hero” is only a theory.

But what does Jungian psychology have to do with putting together a successful data team?

I believe strongly that there are four major archetypes that can be used when filling a team. They are the scientist, the storyteller, the domain expert, and the engineer.

 

 The four data archetypes fit into four quadrants. These quadrants are inside the data vs outside the data and inside the team vs outside the team focuses.

The four data archetypes fit into four quadrants. These quadrants are inside the data vs outside the data and inside the team vs outside the team focuses.

If you were to think about each of these archetypes, they fit into a quadrant of internal vs. external, and group vs. individual.

The scientist is individually focused on what is going on on the inside of the data, what makes it do what it does...

while the engineer is focused on moving the data around.

The storyteller is about telling the story to the rest of the group by examining the parts that the scientist and engineer find.

The domain expert knows the data inside and out and also the rest of the group.

All of these archetypes work together in conjunction by taking the problem, breaking it down systematically, discovering insight, and then finally presenting that back to the rest of the company.

 Customer happiness is influenced by how the archetypes work together.

Customer happiness is influenced by how the archetypes work together.

While I will be talking about each of these archetypes in future blog posts individually here’s some overarching information about each.

The Domain Expert

The domain expert in many ways is probably looked over. This is the project manager, the business owner, the person who gets the domain.

Of course, you’re probably thinking that data science should be impartial. Well yes and no…

Machine learning and AI can become unconstrained goal optimization. This leads us to tricky situations where we are either breaking the law or doing something unethical.

The less extreme example is if we are not following guidelines for the rest of the company.

The Engineer

After the domain expert has explained the problem, it’s up to the engineer to break it down systematically.

You probably know already who is the engineer on your team. These get it done type personalities who are willing and able to move data, program up scripts, and deploy things to production.

These are the make it happen kind of folks.

The Scientist

From the engineer comes the scientist who is like frosting on the cake.

While what they do seems magical, the real point behind science in a data team is to determine insights. Whether that’s through building mathematical models, clever ways of rearranging data, or anecdotal models.

The Storyteller

Lastly and probably one of the most overlooked and important steps is the storyteller. This person’s job is to take what the domain expert, engineer, and scientist have uncovered and then present it to the rest of the company or customers.

Without this last step, the project would stagnate. The best part about this is that the storyteller can also take feedback and give it to the domain expert.

Note: These Are Not Four Individual People

Realize that all of these archetypes can be blended. You can, for instance, have someone who is both a domain expert and a storyteller. You can also have an engineer scientist...though it becomes harder to find someone who satisfies three.

If you are looking for someone who satisfies all of these archetypes, then I think you are looking for a unicorn.