You know the phrase “too many cooks in the kitchen”? While that statement is true, maybe we should also be asking “but do we have the right cooks?” I always say "data is a lot like cooking" and if you are running or starting an analytics team for the first time, you may not know exactly what roles and stations you need in your data kitchen. Do you need just an analyst? What about a data scientist? How about a data engineer? In this guide, I will show you the different “data chefs” you need for your data kitchen so you can better plan how you want to operate it.

THE TYPES OF ROLES

These are the people who are in charge of sourcing your data ingredients. For restaurants, there are people who are in charge of finding the best ingredients. They may manage delivery of ingredients or may just have to go get the ingredients themselves.

For an analytics division, if data is not easily accessible, you will need a data engineer. In addition to sourcing and finding how to get data, they are in charge of making sure that your data is clean. For tools like Domo, there are literally +1000 connectors to easily hook into to get data where you mainly just need login information. However, if a connector does not exist, you may need someone who knows how to hook up APIs, collect data from sFTPs, etc.

Furthermore, their job is to ensure that the data is clean. This is where the prep cooks come into play. Just like finding ingredients that may be spoiled or not pristine, these data engineers / prep cooks should also be making sure the data coming in is accurate. So if you are interviewing potential data engineers, ask what experience they have with APIs, sFTPs, data sanitation and data quality control.

These are most likely to be your main chefs. Traditional kitchens work by having chefs in charge of different stations. For instance, there is usually a chef in charge of the grill, another who runs the certain appetizers, another who manages pasta, etc. Each of these chefs is in charge of perfecting their dish every time it is ordered.

When it comes to the data kitchen, you may have one analyst that does everything or several analysts that specialize in different departments such as an operations analyst, financial analyst, sales analyst and marketing analyst. So do you hire one analyst or several? If you only have one, it’s okay, but now you have a generalist who may not know the KPIs they need to look for. If you can develop and have several types of analyst, you allow for specialization which allows for better quality of delicious insights. If you are hiring for analyst roles, you really want to ask what experience they have with SQL, ETLs, algebra and I would go so far as to say UX.

Doing the math is not enough, you need someone who has a good eye for design who can also make insights look good. I always say “if a dish doesn’t look good, no one is going to want to eat it”. Similarly, if the insights are not visually appealing, they will be hard for your stakeholders to digest. You want stakeholders to be excited about consuming data insights because that’s the fuel of a data-driven culture.

Data scientists are highly skilled individuals who specialize in predictive analytics and machine learning. They are like your pastry chefs where you don’t need many, but you need the ones you have to be very good at what they do. When you go to culinary school, you usually have two main tracks: pastry or everything else. Pastry is so complex that it its its own specialty. Similarly, being a data scientist is a very complex job that takes lots of training and practice.

Moreover, your data scientists need to be doing data science work. I have seen many organizations think they need a data scientist when they really just need an analyst. As a result, they are paying a lot of money to have a data scientist do very basic math. It’s not fun for the data scientist either because they aren’t using their skills. Sure, a pastry chef can cut onions, but is that what you really want that person to do? Instead, your data scientists want to be doing predictive analytics and machine learning like they trained for. Their jobs is to make models to show what will or would happen based on lots and lots of data. So if you’re hiring for these positions, ask what experience they have with Python, regressions, modeling, Jupyter, etc.