In Episode 3 of our Inside Story with FarmTiger, we asked founder Justus why he decided not to build, but to buy a Business Intelligence (BI) solution, what are the advantages and disadvantages, and what specific requirements such a solution should absolutely meet.
minubo: Hello Justus. It was clear to you very early on that you wanted to build your data-driven culture with the help of a business intelligence solution. You also knew from the start that you would not build it yourself, but buy it. Why was that the only option for you?
Justus: To answer this question, I would like to take a step back to explain why we wanted to use a BI solution. In our case, there were two reasons for this: On the one hand, we wanted to create more transparency within the company to better understand our own business. On the other hand, we wanted to be able to derive better decisions from figures and analyses based on this transparency.
If that is my motivation and my goal, then I would like to implement it relatively promptly, because for me better decisions mean more sales and even more profit. It was therefore clear to us that there was only the Buy option, as I can simply get transparency and better decisions more quickly when I buy it. I also believe that it won't be cheaper if you develop it yourself. So the two decisive factors were speed and cost.
minubo: In addition to the setup, what were other important factors for you that influenced your decision?
Justus: Let's think about what I would have to develop myself for a good BI solution. Very important, an intuitive front-end with which all employees can easily understand the data relevant to them. If I don't have that, then I need someone in the company whose job it is to make central database queries and then distribute these results to the departments. But my goal is to get the employees to do their own analyses and to evaluate their key figures independently so that they can gain their own understanding.
Of course, I also need a data warehouse. Developing just this independently could be relatively difficult and a lengthy process. It also has to lead to a correct data model.
The data model is extremely important to us because it enables us to create a common understanding of key figures in the company. So, I have a very uniform definition and it leaves no room for interpretation. So, there is no longer a large number of different key figure definitions, but there are a few that can be agreed on. This is very important for consistency, to avoid miscommunication and to establish trust in the data.
From our point of view, it made no sense to develop a comprehensive data model including the predefined key figures. If there are companies that have worked on this over the years, and have come up with a very intelligent data model, and have optimized it again and again, then I see no reason to have to reinvent the wheel at this point.
From these three points alone, only ‘Buy’ was worth considering for us: First, I would have to create a graphical visualization, that is, an intuitive usability. Second, I would have to program a database myself and third, I would have to develop a data model.
I could imagine that employees would be involved in such an in-house project for several years. If I buy a BI solution and thus the know-how, then ideally I can do the first analyses myself after a few months.
minubo: In case of doubt, a smaller company does not have the expertise to set it up independently from the outset ...
Justus: Yes exactly. I would first have to find the data experts with the appropriate knowledge. I also need a wide variety of areas of expertise ...
minubo: Where do you see any disadvantages in the 'buy' approach?
Justus: In fact, I honestly see very few disadvantages or almost none at all. Presumably, you are set up a little more flexibly in some areas if you set up everything independently. You can make all the decisions yourself: what the setup should look like, which programming language should be used, what the dashboards should look like, how the database should be structured - I then have maximum control. I could imagine that this could be relevant for a large corporation. For us, it wasn't a crucial point.
minubo. Finally, we would be interested in what specific requirements a BI solution must meet for you?
Two points in particular are important here. I need a reasonable form of user interface. By this I mean that every user without IT knowledge should be able to make various analyses or to be able to call up their data. In my opinion, you will not be able to build an organization that is really data-driven if there is only one person in the company who has to prepare and distribute the data centrally and on request. Unfortunately, this is still the reality in many companies.
My goal is to give employees the access and the competence and to explain how the individual KPIs are calculated, what they mean and how they can be influenced. This is a cornerstone of any data-driven work in an organization. Therefore, from our perspective, the most important requirements are a comprehensive data model and intuitive usability.
Even more insights into building vs. buying? minubo CEO Lennard Stoever has also given his thoughts. He illuminates these in his blog article.
Overview of Episodes:
- Episode 1: A Startup as Pioneer for Data-Driven work
- Episode 2: Data-Driven from Zero Hour – it pays to start early
- Episode 3: The initial consideration: BI – Build or Buy?
- Episode 4: Get Started in eCommerce with these Top KPIs