Data should be the foundation of all enterprise decisions and the driving force behind digital transformation. For this foundation to be strong, data should have integrity. However, data integrity is threatened by the persistence of data found in schema-less repositories, enterprises’ reactive approach to data management, lack of processes and systems, and solutions that deliver outcomes the business cannot understand.
According to BBA.org, an estimated 2.5 quintillion bytes of data are created on a daily basis. Enterprises – small and medium businesses included — should consider this a wakeup call for taking data management seriously. We at Liaison believe in the following best practices which should serve as the foundation for more advanced approaches to enterprise data management.
1. Plan for data management.
Enterprises must plan their data management strategy. When planning, enterprises should consider how data will be compiled, managed, and made accessible for its users. Which data is for whom? How will the data be collected and analyzed? What repository will be used? How will data be organized and who will be managing it? During this stage, enterprises should also plan how they will be producing metadata records using specified metadata standards and tools.
2. Collect usable data.
Data collection is one thing, but ensuring the data’s usability is another. Enterprises should carefully consider their data collection methods and documentation prior to collecting. They should also consider using templates for data collection to ensure that only relevant, usable data will be collected. It is during this stage that parameters should be set and descriptive filenames assigned.
3. Ensure data quality.
A study revealed that 40 percent of strategic processes fall short because of poor data quality. From as early as data collection to data analysis, enterprises should perform quality assurance and control on their data. Part of this quality assurance initiative may include double-checking manually entered data using quality level flags for indicating potential data problems, checking format consistency, and incorporating data cleansing methods.
4. Document data.
To effectively understand and leverage data for future use, enterprises should comprehensively document their data. This includes describing the data’s digital context, information and parameters, as well as identifying stakeholders who can best use the data. Documenting data also involves assigning comprehensive metadata, which enables users to discover and use the data.
5. Archive data.
When it comes to archiving data, enterprises should leverage a data center or repository that will support data discovery, access, and dissemination. Data with long-term value should also be identified—that is, data most useful for future stakeholders and data that is hard to reproduce. Remember, however, that when archiving data and regulating access, there are legal policies and regulations that must be taken in consideration.
6. Enable search-based data discovery, integration, and analysis.
Enabling data discovery may entail implementing search-based data discovery tools, which enable the development and refinement of views and analyses of structured and unstructured data. According to Gartner, search-based data discovery tools should have three attributes: a proprietary data structure for storing and modeling data from disparate sources, a built-in performance layer using RAM or indexing, and an intuitive interface.
Most data discovery tools also support integration, analysis, and visualization. Technological advances can support creation and management of complex data while enabling discovery, integration, analysis, and even visualization.
7. Implement schema-on-read technology.
Schema-on-read technology enables enterprises to allow their users to write their data first, and then figure out how to use it later. While this approach has a few drawbacks, Tom Deutsch of IBM Big Data and Analytics Hub enumerates a number of reasons why organizations should consider a schema-on-read technology, including: massive flexibility on how the data can be consumed; storing of raw data for reference and consumption years into the future; enabling experimentation since the cost of getting it wrong is low; faster time from data generation to availability; and more flexibility for storing unstructured, semi-structured, and unorganized data.
The Liaison Advantage
Liaison’s agile, cloud-based data management solutions supports schema-on-read data handling and can turn your enterprise data into insightful and actionable information. Liaison’s data management software solutions not only empower our customers to solve classic challenges, such as master data management (MDM), but also to solve unique and custom issues.
Liaison solutions are delivered via the cloud from the Liaison ALLOY™ Platform. ALLOY’s modular architecture is built upon microservices to incorporate any combination of solutions and elegantly address varying degrees of complexity, customization, and power.