There have been a lot of articles on managing data teams [1]. John Hopkins University even has a course on building a data science team [2]! I will share my own personal experience in building a data team in a startup. This article will share what is a data team, when and why startups need this team, how I formed a data team, and finally I will close with recommendations on forming a data team. This article aims to give insight for new startups leaders that have started to think about how should they form a data team.
A Brief Introduction To Airy Indonesia
Airy, founded in 2015, is a virtual hotel operator in Indonesia [5] (in Bahasa Indonesia). We have invested in technology that may change the accommodation industry. Unfortunately, we have to close down due to the current pandemic situation [6]. However, feel free to browse what we have written in the past about our technologies! [7]
What is a Data Team
A data team is a team that has a function to bring values from large and overwhelming data sources. Each team has its own function and by function, a first definition that comes to my mind is given an input transform that input into output [3].
Large and overwhelming data sources are inputs. As I have mentioned, there are 3 characteristics that need to be fulfilled: large, overwhelming, and various (hence I said data sources), and this characteristic translates directly from the definition of Big Data [4]. The volume of data would typically be large, thus it will need special software architecture to bring down the cost of storing that data. Overwhelming here means that the velocity of the data is so large that it should be handled with care to bring down the latency of processing. Finally, various means of the variety of data formats will add complexity to process the data itself. These factors will require a significant amount of effort, thus forming a team does make sense.Bringing value is the output. The data team needs to bring values to the business in the form of descriptive, prescriptive, or predictive analysis. Most of the efforts and value of this team will be spent on descriptive analysis, either in the form of dashboards, reports, or EDA. Prescriptive analysis will need more collaboration from other divisions like marketing, product, design. Predictive analysis will be conducted when both business and team skills have matured.
When Does A Company Need a Data Team?
I was recently asked, by a former colleague, when should we form a team if we are in the early stage of a startup? It is an interesting question, in my opinion, because nobody even bothers to ask me that simple but challenging question. I came at the relatively early stage of startup straight to the data team, at that time we had a different data team lead so my CTO has envisioned that this company needs a data team.
#management #startup-lessons #data #startup #data analysis #data analysis
