How to build a data team from scratch

With data analysis becoming key to staying competitive in almost every industry, it’s time to think about building your own in-house data team, we understand that not all businesses are the same and there is no one size fits all approach, but where do you start? Read on for a step-by-step guide of how we have seen teams put together successfully.

1. Who to hire first?

Should you start by hiring a data engineer to build your technical infrastructure, or a data analyst to determine what data you need and therefore what your infrastructure should look like?

Before you decide, look at who you already have on board. If you have someone with a bit of engineering experience, get them pulling data together and hire an analyst to start making sense of it. If you have someone who can handle data analysis in SQL or Excel – think former consultants, financial analysts and project managers – hire an engineer to start finding them data. 

If you don’t have anyone with either of these skills, you probably want to start by hiring an analyst with some basic data engineering skills. Advances in data tools mean that anyone who can use push-button data tools can build a modern data stack.

2. How to structure your team?

Most data teams are structured on one of three models:

Centralised: A team where everyone reports to a single leader, acting like in-house consultants for the rest of the business. Works best in small and medium-sized organisations. 

Decentralised: Business units hire their own data professionals who specialise in their area of the business. Works best in large organisations.

Hub-and-spoke: A combination approach where a centralised team liaises with data professionals in other teams. Works best in organisations that need more specialisation from data team members.

3. How to start working with data?

Data analysis is only as good as the quality of the data: garbage in, garbage out. New data teams tend to have to deal with a lot of garbage. After making your first hire or hires, a good first step is to start centralising raw data from different sources into one place, then cleaning it up and modelling it using tools like Airflow or dbt. 

This can be a multi-step process, so it’s best to approach data management in an agile, iterative way. Don’t worry too much about getting your whole strategy planned out in advance–keep revisiting it, evaluating it, making improvements, and adding new tools as you go. With the industry evolving so fast, there are always things you can add to make your data management system better.

If you are looking to build your data team and would like to discuss how we can help you please contact us.