Chapter 1 - HR Matters
IN THIS LESSON
Building Data Teams
A start-up in the Age of Big Data feels the pressure to be data driven. Aside from the many definitions of that term, such as whether you use data deterministically or probabilistically, the challenge is finding the right team. Realistically, you cannot outsource your data team so you must build it. The reason you cannot outsource is that the data team needs to understand and interact with your team at a high level. They need to have skin in the game and care about the business if they are to generate a useful and creative insight.
Since data has become so important to all aspects of everything, it might seem like an easy decision you need to leverage data early. However, think about it critically. The data team does not generate revenue. They can help generate data to optimize things, but you should be able to figure out things at the crudest level. Things like sales to phone calls made can be tracked easily. It is when you established a complex funnel or other complex systems where a data team can come in handy.
A start-up must be wary of building something that scales. You don’t scale a company by hiring a bunch of data scientists. You scale the company by adding people that make revenue and support those activities including leaders to manage them, and you hire data people alongside that, so you use that scale wisely.
The #1 advice for building a data team is to take it one step at a time. You need to know your needs and understand the roles. Data architects, data scientists, data analysts, and data engineers are all different though architects and scientists overlap. However, you might not need them all at once. Start with the analyst and crunch what you have in Excel if you want to start cheap. The best option is a utility player. This is someone who has skills as a data analyst, basic to intermediate skills as a data scientist, and basic to intermediate skills as a data engineer. Find someone passionate with a broad range of skills with the ability and desire to hone those skills further. You want someone with the basics who will figure it out and build out your initial models. Bonus points if this person has skills managing and building a team. That is the way to go for a first hire.
Next, you’d probably go with a dedicated engineer. The engineer will help on the technology side and bring the data game to the next level. You might get a few analysts, scientists, and engineers over time depending on your needs. When you have a good amount of capital to reinvest in the business, and your KPIs are at a steady state, you are probably ready to hire a data architect and possibly start from scratch. The architect will help you get a standardized, industrial enterprise-level data solution in place. This could be done using various services along with the help of the engineer.
When you let the architect lose to build it right, you probably must expand the team quicker. While it sounds like the architect should be the one in charge of the department, you’re probably better off going with the person with the best business acumen and strategic insight. This might be the data architect, but that utility scientist you hired early on probably understands your business best, understands the current data systems best, and if you chose right will learn what the architect builds.