Data governance
Modern digital businesses have a high degree of data literacy. It’s essential that everybody in the organization understands the fundamental importance of data. Every project is also a data project, as it produces and consumes data and affects the overall data flow in the organization. Enterprise-wide data governance defines the rights and duties of data providers and data consumers. Every decision in the organization is based on trustworthy data.
Data discoverability
Information needs to be at everyone's fingertips, where it’s very easy to ingest by everybody in the organization (following data accessibility and legal data privacy rules, of course). Information can be accessed by simple and standardized APIs, self-services and freely combined to gain new information and insights.
Data trustworthiness
People in any organization will only allow intelligent and data-driven decisions to be executed if they have trust in their data. Pre-digital organizations introducing data-driven decisions often find that people in the organisation don’t trust the data and the decisions based on them, and overrule the automatic processes by manual decisions. As in human relations, trust comes with understanding and positive experiences. In modern digital businesses, people have great trust in their data and the decisions based on them. They understand the provenance of the data and the way decisions are made. They’re also aware of possible biases in the data. A high data trustworthiness doesn’t necessarily mean that data always has a high quality. Modern mathematical and statistical algorithms can (like humans) deal with incomplete, inconsistent and often noisy data.
Data insight ability
Modern digital businesses have a strong capability to generate information and insights out of data. This is done by data scientists and analysts. They model the data with advanced analytics and machine learning frameworks. Data time series and streams are modeled to predict future events. The derived insights are at the core of every decision and action in the organization.
Data actionability
The best data insights are useless if they don’t lead to actions. Therefore, bringing information and insights into actions in a fast and efficient way is another fundamental proficiency of a modern digital business. Insights are visualized to support business decisions (data storytelling) or included in the software development and operations process by CD4ML.
Data security and protection
As data is becoming the fundamental asset in a modern digital business, it’s important to secure and protect it. The data has to be protected against unauthorized access from inside and outside the organization. It also has to be protected against loss and corruption. More importantly for person-related data digital privacy and the obligations of legal requirements like GDPR (General Data Protection Regulation in Europe) must be secured.
Data is at the core of every Modern Digital Business
Becoming a truly Modern Digital Business isn’t solely an IT project. It’s a cross-organizational effort that affects every aspect of your organization, with technology and business working hand in hand. We’ve identified five building blocks that you have to develop in your organization. We’ve shown that data plays an essential role in each of these building blocks. In order to successfully develop these building blocks, the organization will need six data proficiencies in their IT environment and workforce skills.
Here are some examples of how you can start building these proficiencies in your organization right away:
- Understand the importance of data. Everybody in your organization — starting from the very top to the grassroots — should understand the importance and criticality of data for the business. Data is a raw material for your value creation, it should be available for everybody involved and not owned by single business units or “silos” in your organization.
- Make your data discoverable. Data in ERP systems or relational databases aren’t considered discoverable. They’re fixed in their tabular data schema and hard to extract and wrangle with. Data scientists love data lakes with simply structured or unstructured data, but be aware: setting up huge organization-wide data lakes are mega projects, too big to swallow for most organizations and usually doomed to fail. Multiple smaller, domain-specific data lakes like data meshes, with clear ownership in the domain, offering data per self-services to other domains to build data services and products is the way to go.
- People must have trust in the data. Set up continuous processes and high standards to clean-up and quality-check incoming and generated data. Decisions based on data should be made as transparent as possible and, if employees have doubts, treat them seriously. Find out where these doubts are coming from, look into the data and the decision process (including the speed of the decision process), compare with human decisions and convince people. You can’t simply mandate that automated decisions are trusted. And be aware that bias in your data can have a devastating impact on the decisions made based on these data and therefore on your business and public image.
- Request every decision to be data-driven or at least data-proofed. Don’t accept decisions by gut feeling, opinions or “we did it always like this” any more. Continue on this route and you’ll fail, as those that use data-driven decisions adjust quicker to the changing environment.
- Try to automate decision making as much as possible. Ask your data scientists to model decision making or use RPA (robotic process automation), at least for mundane processes in your organization. Upskill the employees whose daily job was to make these decisions to train and maintain the data models and RPA software or to work on more strategic topics.
- Analyzing and modeling the data and creating information and insights out of it should be a core competency in your organization. Don’t outsource this. It’s a major pillar for creating value out of your data and becoming a data-driven organization with a leading edge. You’ll need data analysts, data scientists and a proper data infrastructure and processes. Make sure their work is transparent, openly available, repeatable and not bound to single heads.
- Pave the way from data insights to operation in your organization. Great data science models and insights are useless if they only reside in Jupyter notebooks. They only create value when they reach the end user. Don’t let your data analytics and data science team work in isolation: set up mixed teams together with data engineers, developers and IT operations. Done means that data is available for the end user. CD4ML helps you to set up processes and an infrastructure to move data science models smoothly, securely and continuously from development to production.
- Last but not least make sure your data is protected: against theft and misuse from inside and outside your organization, and that the proper protection of sensitive and personal data — obeying applicable regulations like GDPR and against catastrophic loss — are in place. Data will become a main ingredient of your business value creation and “the blood”, flowing through your organization and keeping it healthy!
Unfortunately, there is no Holy Grail plan to becoming a modern digital business, nor should there be. It’s a journey that each organisation needs to define and learn as they go. Not every company needs to become the Amazon, Netflix or Google of its industry. Instead, every organization has to find their own authentic digital self.
ThoughtWorks has developed the Digital Fluency Online Tool to help organizations to understand where they should be versus where they are. It helps you to identify what are the core capabilities you need to work on to bridge the gap to where you aspire to be.
But remember, this is just a starting point of a complex never ending journey, there’s no such thing as a plan: the ambiguity is constantly challenging us to keep changing to remain digitally fit!