The importance of data storytel­ling in large-scale projects: More than just data quality

When discus­sing data quality in large-scale projects, many focus on obvious issues such as data outliers or incor­rectly recorded manual entries. While these challenges are important, they are only a small part of the problem. Much more important to the success of a project is the ability to correctly tell the story contained in the data – a concept known as data storytel­ling.

Data as a narra­tive of the business process

Over time, all opera­tional business processes in a large organiza­tion generate a tremen­dous amount of granular data that is stored in enter­prise resource planning (ERP) systems. This data and its associated tables not only reflect the current state, but also tell the story of the under­lying process. Hidden within are patterns, depen­den­cies, and trends that map the entire flow of a business, and the real chall­enge in lever­aging this data is to correctly identify and inter­pret these stories. When this is done, not only can existing processes be better unders­tood, but new, value-adding or even disrup­tive algorithms can be developed using these insights. However, this requires that the data be viewed in a consis­tent and coherent frame­work.

Fragmen­ta­tion as a chall­enge for data storytel­ling

In large enter­prises, diffe­rent business units such as sales, purcha­sing, produc­tion, and delivery often operate their processes in their own, often isolated, silos. These silos consist of separate process centers that operate independently of each other. Other large enter­prises also have decen­tra­lized struc­tures that include not only diffe­rent organiza­tional units, but also diffe­rent system landscapes. This fragmen­ta­tion is a natural conse­quence of humans breaking down large and complex tasks into smaller, more manageable units, and for data science teams and AI projects, this fragmen­ta­tion is one of the biggest challenges at the start of a project. The question is whether it is possible to ensure consis­tent and end-to-end data storytel­ling. Only when data is seen as part of a coherent story can it unfold its full impact and unlock the poten­tial for new insights and innova­tion.

Local solutions and the big picture

It is not only possible, it often makes sense to develop solutions locally, within indivi­dual business units. These local solutions can address specific problems and provide valuable insights. However, it is critical that the under­lying data be placed in the context of the entire enter­prise. The data must be seen as part of the same story that is told across the entire business process.

Conclu­sion

While data quality remains an important aspect of data proces­sing, the bigger chall­enge in large-scale projects often lies in the area of data storytel­ling. The ability to develop a coherent and consis­tent narra­tive from the granular data in an ERP system is critical to the successful imple­men­ta­tion of data science and AI projects. This is the only way to unlock hidden poten­tial and develop truly trans­for­ma­tive solutions. The fragmen­ta­tion of processes and systems within an organiza­tion makes this task parti­cu­larly complex, but maste­ring it is the key to sustainable success.