Digital transformation is the “buzzword du jour” in every industry. There have been many initiatives that should have led to a digital transformation across many industries -- supply chain integration, global ERP systems, etc. These likely should have prepared us for the digital life. This fell far short in large part to one key element -- data. Data is key to any digital transformation journey, but the foundation of all data is master data. Information about materials and products, customers and vendors are the bedrock of a digital framework. However, companies big and small need a strategy to manage that master data before they begin building their digital transformation dreams upon it.
What is master data?
Master data is the core data that gives meaning or context to transactions and data analytics. It can certainly include data that’s defined inside the organization from outside sources -- suppliers/vendors, employees, customers, materials/products and organizational data (e.g., companies, business units, plants, consolidating entities). It’ll also include data defined outside an organization, either by industry organizations or other centralized entities (such as governments, ISO or The United Nations). This could include reference data such as country names and codes, state/provincial names and codes, currency codes, UN location codes and units of measure.
Some of this master data relates to other types of master data. Regarding materials and products from within a company, one attribute may be its classification as determined by the United Nations Standard Product and Services Code (UNSPSC). Master data such as this is essential for companies to exchange information between each other as customers and suppliers. Clearly, the geographical information that’s standardized by governments and international standards organizations is critical to determining the addresses and classifications of suppliers and customers (this also helps to determine duplicates.)
What are the key elements?
First and foremost, support (and enforcement) needs to have full management approval and buy-in at the enterprise level. Support from business units is also needed, but it’s secondary to support from the top of the organization. Enterprise support is also vital to the second element, the elimination of data silos, which also allows for a full data inventory. Oftentimes, master data and its processes are locked within business unit silos. These are often system-driven (e.g., global system for customer master data is SAP, but one or more business units have Salesforce CRM with its own customer master data that doesn’t tie to SAP).
By breaking down walls hiding pockets of data a full data inventory can be completed so that rules can be developed and applied. These rules may govern data field requirements, special coding or the definition of a duplicate record. In many cases, the enforcement of these rules can be handled by a centralized master data management or governance tool. Such a tool would capture all required master data and publish to the various systems that require it, giving all such systems a common master data record.
The next element of a master data strategy is data rule definition. This is usually mandated by a management or governance system, but it’s also key to process changes absent any system. Data rule definition generally includes naming conventions, common abbreviations and punctuation and rules for determining data duplication. In many cases, master data within the same silo structure will have significant inconsistencies (i.e., upper and lowercase used in some records, all uppercase in others).
A prime example most companies can point to is how the telecommunications company AT&T is set up. Generally, depending on the age of the system, you may have all the following: “AT&T”, “AT and T”, “A.T. & T.”, “American Telephone & Telegraph” and possibly others. The same holds true for companies that have merged, been acquired or simply changed names. These can often fall into the “duplicate” category, but these are harder to assess. Defining consistent data entry rules can resolve these issues.
Finally, there comes the evaluation and cleansing of existing master data. While this is a daunting task, it’s one that must be done. At the same time, any new data coming into the master data ecosystem would follow the same rules and data duplication assessment. This could be an instance where taking data sets from one master data category for evaluation and cleansing may be the most sensible alternative rather than attacking all data sets simultaneously.
How can formalizing the strategy build the foundation?
Enterprise means enterprise. All of it. You cannot have a master data strategy without involving the whole organization. Many organizations will try to experiment with their strategy by rolling it out in one region or in one business unit at a time. Doing so immediately breaks down the elements that we laid out in the fundamentals of an effective strategy. It also reinforces hazards of the data silos previously mentioned.
To be effective, a master data strategy should be a “Big Bang” across all business units and regions. If there’s a need for experimentation, select a single data entity (perhaps materials) and roll out the strategy globally. Doing so will allow the organization to adjust rules, processes and workflows and weigh the impact of building a foundation with a common master data strategy. This proof-of-concept could also identify potential issues with other data entities.
Is a master data strategy all you need for digital transformation?
Deploying an effective master data strategy across the enterprise is a good start, but it’s not the sole basis of digital transformation. While we hinted at it in the master data examples, we didn’t address the need for solid integration between systems and processes. Integration, along with sound data practices, is what makes digital transformation work. Without that integration, the elements of robotic process automation (RPA), machine learning and artificial intelligence (AI) cannot be effectively applied to any process or industry. Consider levels of data integration that are practical today and those to include in a strategy for tomorrow to complete your own digital transformation journey.
Rob Roberts is a Director in Opportune LLP’s Process & Technology practice. Roberts has over 20 years of experience focused on the delivery of mid-to-large-scale ERP implementations involving process optimization, system integration and application automation. His focus has been on the architecture, design and implementation of cross-functional solutions, including process integration, mobility and business analytics. He has been involved in multiple full life-cycle system implementations from pre-sales and system planning to implementation and support. Prior to joining Opportune, Roberts was responsible for ERP and technology services for multiple private consulting firms.