With the mass of available data and the variety of options, the choice is often difficult. Identifying the right use case is therefore one of the most important and biggest tasks right at the start of a data science project. What is even more serious is that although the right use cases are often identified, the results still do not meet expectations. Actually deliver 80% of data science projects Despite the original assumption, not the desired operational added value in everyday business. The main problem is not the analysis method, but the lack of user-orientation of the developed solution. So how is it possible to identify user-oriented use cases with high business potential? A challenge that can be met with the help of the data thinking method. In this article, you can find out exactly what this is all about and why we at pacemaker rely on data thinking in our workshops.
What is data thinking?
Definition: Data thinking is a holistic approach that helps companies to systematically identify user-oriented use cases that create actual added value.
Data Thinking combines data science with the proven innovation method Design Thinking.
Data science: An interdisciplinary science that deals with extracting knowledge from data.
Design thinking: A creative process for finding ideas and solutions that is user-oriented.
Data thinking: The creative development of user-oriented data science use cases with high business potential.
By using design thinking methods within data science projects, the focus is on the user and the development of a data-driven solution that meets their needs right from the start of the project.
Why data thinking?
Traditional consulting approaches, which are often also used in the data science environment, traditionally start from the perspective of existing data. Does mean what data is available and what can be done with it? This approach usually results in solutions being developed which, in the end, are of little use to the end user. In contrast, the data thinking approach starts directly with the problem to be solved and the needs of the customer. The underlying question is therefore a completely different one: What is the problem and who has the problem?
The data thinking method therefore focuses on relevance for the company and in particular on the user's perspective. Only in the second step is it examined which methodology and data can be used to solve the problem and to what extent they already exist or need to be collected first.
The principles of the data thinking approach provide companies with the following benefits, among others:
Problem instead of technology: Business and user benefits are at the center of all considerations, not the technology behind them.
Added value before data: The use cases are identified based on the business challenge and not on the basis of existing data.
At pacemaker, we therefore use the data thinking approach in our initial workshops to ensure that the right questions are asked right from the start when using the right use case. In this way, important technical and process-related questions are clarified in advance. Finally, those data-driven use cases are tackled that can be implemented with reasonable effort and that are really worthwhile for you, from an economic perspective and user perspective!
What does the data thinking process look like?
The aim of a data thinking workshop is to gain an understanding of the problem, the relevant business processes, the areas of responsibility, the existing data and IT infrastructure, and project expectations. To achieve this, we at pacemaker have developed a three-stage process that is based on the Design Thinking process but reduces it to the steps relevant for data science projects.
The graphic below provides an overview of the process and the goals of each phase.
The process is of course not set in stone, but is flexibly and individually tailored to the needs of each customer.
In order to approach customer challenges and a possible solution to them, we use various methods from the Design Thinking and Six Sigma environment. The addition of Six Sigma methods is particularly helpful for data science projects due to their focus on data-based process analysis. In this way, our data scientists individually assemble the right toolkit for each customer in order to dive into the customer's working environment and tease out the necessary information.
Before we take a closer look at the individual phases, the question remains as to who should ideally be involved in a data thinking workshop? Different people are needed to obtain a complete picture and to meet the various information requirements. It is important that, in addition to technical experts, i.e. data scientists and data owners (people responsible for data management in the company), and management, the end users of the potential solution are also represented. After all, it is important to develop a solution that should make everyday work easier for them. In these interdisciplinary teams, tangible and concrete results are then worked together in every data thinking phase.
Data Thinking Phase 1: Define
The first phase is all about the problem to be solved. What is the challenge facing the company? Which specific problem should be solved? Since customers often approach us with lots of ideas, it is important to get an overview and to work together on a specific problem or question. The following applies to every data science project: The question and the data basis must go together! The participating data scientists therefore already have an eye on possible data sources and potentials and can thus help with a realistic assessment of use case ideas. In this way, we minimize the risk of setting unrealistic goals as a result of an incorrect assessment of the data situation. In general, it has proven useful to focus on smaller use cases at the beginning in order to build up experience and achieve quick wins.
The aim of the Define phase is then also to define the customer's expectations of the project and to document these for the rest of the project.
In short, the first phase includes the following steps:
- the development of a specific problem
- the formulation of a measurable target
- the formulation of measurable acceptance criteria
Data Thinking Phase 2: Empathize
The second phase is about getting to know the customer better based on the defined problem. The aim is to obtain a profound understanding of the day-to-day processes and the environment in which the solution is ultimately to be used.
The empathize phase therefore includes the following steps:
- the appointment of relevant contacts in the process
- the presentation of relevant processes and procedures
- Defining relevant data and data owners
- the presentation of relevant infrastructures
While traditionally the focus is primarily on understanding data, the data thinking approach also places great attention on process understanding. The early focus on operational processes lays the foundation for subsequent successful operationalization right from the start of a data science project.
Data Thinking Phase 3: Ideas
In phase three, it's time to get creative: It's important to define what users expect from the solution so that it creates real added value. Does it mean, for example, which features must it have? It can help here to simply step on the whiteboard to outline an initial solution.
Starting with the problem to be solved, after completing the three phases and thus at the end of the Data Thinking Workshop, there is a roadmap for a user-oriented and data-driven solution to the problem.
Looking at the typical project flow of a data science project, the described workshop successfully implements the first step of business understanding.
Conclusion
A successful data science project is much more than just about data. With the help of the data thinking method, you can unlock the full potential of your data by approaching use cases in a user-oriented manner right from the start. This ensures that, after successful implementation, the developed solutions also create actual added value for your company and end users. This increases the trust and acceptance of the resulting solutions by your employees. In addition, your data science initiatives can be measured through a clear definition of objectives and acceptance criteria. All in all, even skeptics become fans and paves the way for further use cases.