A digital transformation is something that you do and keep doing and is highly customized to your company. It enables all important decisions — whether made by a person, a team or an algorithm — to be informed by data and context.
A digital transformation may encompass hundreds of discrete changes in process, behavior and technology. But as companies progress through the various stages of a digital transformation, their use of data and digital assets changes.
In my research I have tracked the use of data and how it changes as organizations create increasingly sophisticated digital processes and analytics.
This high-level digital transformation maturity model will help put these changes in context:
During the experimentation phase, the main drivers of the transition — CMOs, CIOs, CISOs, etc. — will wish to inventory all digital assets, tools, technologies and expertise available to them.
In terms of the data, that inventory is built from a process of discovery usually undertaken by central IT with participation from all relevant lines of business. The number and variety of data sources varies widely depending on company size and industry.
Early on, it may only be a handful of sources, including CRM systems, customer databases, clickstream data and data warehouses. Later in the process, a medium-sized enterprise might identify dozens, and a data-rich financial services company may have hundreds or thousands of sources.
Once an inventory of available data sources is complete, the next step is to extract that data and use it to populate a data lake. Modern data platforms such as Apache Hadoop or commercial platforms now provide basically limitless scale and capacity for ingesting data and creating just such a repository. So, at this point if you follow the data into the data lake, you will see large volumes of structured data (from databases and data warehouses) and unstructured data (from everywhere else) stored in their original formats.
Once your digital assets are discovered and extracted, they must be reviewed and rationalized by your data wonks (data scientists, data engineers, statisticians, BI experts and quants). This can be a highly revealing exercise that will significantly shape the process of creating digital processes.
During the implementation phase, the focus will mainly be on establishing and integrating new digital processes, and then proving their efficacy.
The characteristic shrinking of development and deployment cycles by agile businesses has their executives looking for quick ROI and a measurable proof of progress by the implementation teams. These teams will find that populating a data lake is really step zero in a digital transformation.