Scaling is a reflection of the system's ability to work with big data loads, as well as support for ways to increase performance. Reaching the limit of the data management system's capabilities is fraught with the search for new solutions, or the use of several systems at once, which leads to serious budget expenditures.

Here are some ways to increase the performance of MDM systems. Some systems may support several options at once:

• Using cloud services;
  
• Three-link architecture;
  
• Cluster solutions support;
• Database distribution;
• Support for connecting new data domains.

The license policy is one of the first criteria when choosing an MDM system since an unsuitable license can later cost much more. It is important to assess how the data management project will develop, what it will focus on, and what resources will be used for it.

There are several license types:

• Unified product license;
  
• By server numbers;
  
• By record number;
  
• By user number;
  
• Temporary subscription (no other restrictions present);
  
• Separate modules subscription.
  

Single-purpose. Systems that are developed for specific business subject areas, such as mechanical engineering. Such systems are ready for all the specific needs of the subject area, but cannot be used in other areas. However, not every subject area has a Master Data Management system. It is also important to keep in mind that specialized systems may have performance limitations, such as the number of records processed.

Universal. Systems that contain a wide range of tools that can be used to create an MDM solution for any business. The price of such versatility is the time and resources spent on product integration and implementation. The quality of integration and implementation directly depends on the qualifications of specialists.

The MDM system must have all the data management tools needed at the start of the project and during its development. In addition, you should pay attention to the functionality and convenience of the tools.

The set of basic tools for data management systems is as follows:

• Working with the data model, including classification and linking management;
  
• Working with data (for Data stewards);
  
• Data quality assurance (DQ);
  
• Business process management (BPM);
  
• Reference data management (RDM), including maintaining taxonomies;
  
• Interaction with other systems;
  
• System administration, including data security.
  

Besides the ones listed, an MDM system can have many different tools aimed at solving side business problems. For example, calculating data statistics and visual representation of indicators.

To solve individual business tasks, the MDM system can provide customization options for the following aspects:

• Adding, modifying, or integrating new logic in standard operations. For example, the implementation of the new checks when you save the record.
  
• Configuring the app's appearance.
  
• Using data from external services. For example, displaying addresses on a map.
  
• Implementation of custom business processes using BPMN 2.0 notation.
  

MDM systems can be customized in several ways at once. Each direction usually includes multiple expansion points at different parts of the system. However, there are no recommendations for the optimal number and quality of expansion points, since when choosing a system, it is important that the required functions, which are not available in the standard delivery, can be performed using expansion points. A high level of flexibility increases the chance of solving non-standard and individual business tasks.