Modern enterprises have to maintain a number of systems and applications to manage data on a variety of products and hundreds of thousands of customers. This makes data management inherently complex and fragmented.
As critical business data can change from time to time, there can be instances of variance across all of these systems. For instance, if a sales rep adds an incorrect customer address for a new order, the incorrect information will go through multiple systems of the order fulfillment process until some layer, such as the accounting department, introduces a correction. This would leave two data versions for the same order across systems – one correct and the other incorrect.
With thousands of transactions happening every day, acquisitions being made and customers updating their information, the issue is exacerbated multifold, leaving a large chunk of data out of sync and no way of determining which version is correct, incorrect or outdated. Using this data for analysis can lead to incorrect decision-making. This is where master data management (MDM) steps in.
What is master data management (MDM)?
Also known as a golden record, master data is a key organizational data asset that contains the most up-to-date and accurate information for day-to-day business operations. It supports transactional, non-transactional or analytical data and is usually shared across departments to help personnel conduct analytics and make decisions around service, sales, marketing and other areas. Master data serves as the source of common data and often includes application-specific metadata and corporate dimensional data (that directly participates in transactions like a customer ID or department ID), although the data type can vary depending on the organization and its needs.
Broadly, master data has to have three key qualities – less volatility, more complexity and mission-critical. Plus, it should be used and reused over and over again. A good example is contact numbers collected from customers. This kind of data is less volatile, mission-critical and can be very complex, meaning a slight error in numbers can result in a missed opportunity for the business.
The ongoing practice of creating and managing the master dataset for all critical information is called master data management (MDM). It encompasses all of the technologies, processes and people that help organizations consolidate, cleanse, de-duplicate, organize, categorize, localize and augment master data as a single source of truth and then sync it with all business processes, applications and analytical platforms in use across the organization. When MDM is executed perfectly, the master data being disseminated is in sync, highly accurate and trustworthy.
Various domains of master data management include the following:
- B2B and B2C customer master data management
- Product-specific master data management
- Supplier-specific master data management
- Location master data management
- Asset master data management
- Employee master data management
In the early stages of MDM, each type of master data had to have its unique data store and business logic. Companies looking to optimize the performance of their employees, for instance, centered their MDM strategy only on the employee master data domain. However, these “single-domain” approaches are not as effective today because the data has become more complex and intertwined than ever. In order to extract full customer information at present, enterprises not only need customer master data (demographic, etc.) but also product-specific master data, which may signal their buying preferences. This is exactly where “multi-domain” MDM, which manages all different types of master data in one repository, comes in.
According to a study from Aberdeen, companies leveraging multi-domain MDM have seen better results than those using single-domain MDM in terms of completeness and accuracy of data.
Importance and benefits of master data management
Efficient master data management gives an organization a single place to have an authoritative view of information, which in turn eliminates costly redundancies that occur when organizations rely upon multiple versions of data distributed across siloed systems. For instance, if a customer’s information has changed, the organization will update the master data with the new information and not turn towards driving sales and marketing efforts using the old data point present in other systems.
Better data quality
With MDM, organizations also get access to better data quality that is more current and suitable for analytics. The discipline makes the data format more consistent and uniform, which makes it more usable for various business processes and simplifies answering basic questions, such as “What services did the customers use during the last quarter?” or “Who were the top buyers during the same period?”
Revenue growth, improved customer experience
Since MDM consolidates complete and trustworthy data from customers in one place, organizations can use it to better understand their target audience as well as their preferred connection channels. Then, using those insights, they can make personalized sell-up or cross-sell offers to the right person at the right time. This would not only drive up revenues, but also help cut down spending on unnecessary customer acquisition methods. Issues like sending emails to incorrect customers or making calls about resolved matters can also be avoided.
Easy to back up
MDM’s centralized copy of business-critical data also makes it much easier for organizations to create a backup. Backing up silos is a costly affair, but MDM simplifies the process, giving organizations an easy way to recover their data in the unanticipated case of a disaster or loss of information
Accelerated product launches
Enterprises can accelerate their time-to-market with MDM. In particular, product and supply chain master data management can help organizations cleanse and enrich almost every aspect of information required to set and meet product deadlines. A study from Stibo Systems reported that 58% of organizations using their MDM solution reduced their TTM from months to weeks and 36% brought it down from weeks to days.
A well-executed MDM strategy, simplifying the handling of customer information, can also help organizations comply with regulators as well as privacy laws. Centralized management of customer master data would be especially useful in the case of General Data Protection Regulation (GDPR), which gives customers an option to get their data deleted or ported to another company providing the same service if needed.
Master data management framework: Key components
- Discovery: This aspect of MDM practice revolves around identifying where master data resides. Organizations have to check their entire data landscape to see which systems on-premises or on the cloud have fragments of business mission-critical data, the type of that data, and its quality.
- Acquiring: After discovering data sources, organizations have to acquire all the relevant information. This is done by connecting the data distributed across all applications – from ERP to CRM systems – to a central repository.
- Cleaning: Post-acquisition, all the information gathered from source systems is cleansed, with its overall quality being improved. To do this, organizations have to deal with data inaccuracies and make the data format more consistent for downstream use. Effective cleaning is critical to make the master data trustworthy.
- Enriching: In this step, the master data profile that’s created is enriched by connecting to third-party sources of trustworthy data such as Dun and Bradstreet and Axiom. Enriching can address the incompleteness of source system data.
- Matching: After cleansing and enriching, the master data asset is matched to check for duplication. Here, certain tools are used to leverage established business rules and identify different records for the same thing/person.
- Merging: The duplicates identified through matching are subsequently merged into a single version or golden record. In essence, the results of the previous step are run through a resolved process, where rules of survivorship are used to select the most accurate and relevant piece of information for the person or object concerned. If there’s uncertainty around any data point, this process creates a set of tasks for data stewards who can then conduct manual merging.
- Relating: The master data is ready by this stage, but simply using it won’t be enough. This is why it has to be related to data points from other domains, such as information from the supply chain, production or employee systems. This stage makes up multidomain MDM, giving organizations a 360-degree view of data to work with.
- Security: The next step revolves around implementing security measures to make sure that the master data, which contains business-critical information on customers, products and other areas, remains protected from unauthorized access. To do this, organizations often mask their sensitive data and limit who or what systems get access to the master data.
- Delivery: Finally, MDM practitioners look at the delivery of trusted, relevant and secured master data from the central repository to the right applications. This ultimately helps people with analytics and other data-driven decisions.
6 top master data management strategies for 2022
Groundwork for discovery
In order to ensure that the discovery of master data is done to perfection, you need to do the groundwork to make the entire enterprise data landscape accessible. This would include giving access permissions to the MDM team and flagging any obscure sources of data that might not be directly available.
Beyond this, a significant portion of time should also be spent on data profiling and usage patterns so that only relevant pieces of information are taken into the master data repository. Consultation with subject matter experts within the business can also come in handy here.
Look at the multi-domain approach
Instead of creating and managing master data for a single domain, try to go for the multi-domain MDM approach. This will connect hidden data points across functions – from product to supply chain – and give a more holistic view of information to drive better results.
Implement with transparency
After deciding to switch to MDM, it is important to make sure that all people who will be asked to use the master data are made aware of the incoming changes and why they are being made. Implementation should not come as a sudden change. In addition, the users of master data should be given sufficient time to adjust to the changes and an option to share their feedback, ask questions and identify gaps (if any).
Training is important
Beyond transparency, all personnel and departments getting access to the master data repository should be trained and retrained on various aspects of formatting and using the data. There should be workshops to educate the personnel on how they could leverage the master data to meet the set business goals.
Observation and measurement are critical
Organizations should implement MDM in phases instead of going all-in at one go. Once the process is completed, project managers must continually work with management, employees and other stakeholders to discuss feedback and suggestions aimed at improving the system for better business outcomes. Simultaneously, they should measure the ROI of the entire process, starting from business case formulation to post-implementation, and regularly audit crucial aspects such as installation, configuration, data models and alert queues to avoid systemic delays when new business requirements and challenges arise.
Triage data issues and prioritize
When conducting audits, there is a good chance of running into one or more data issues. Project managers need to expect this and be prepared to handle whatever comes up. This could be done by implementing a suitable process for triage, where the issue could be assessed and addressed depending on how urgent or critical it is.
If the issue is small, it could be addressed immediately. However, if it requires long manual work, then it could be pushed off until the launch. In a nutshell, organizations should have a set structure in place to address data quality and governance issues.