As your organization grows, its data becomes increasingly messier until the point it becomes basically unusable without large cleanup efforts.
Master data management (MDM) is the process and tools that are used to reduce redundancy and conflicting information. A company trying to do data analytics or data engineering is likely too fail if they are not routinely cleaning and cataloging data that is coming from different sources. Master data management is the process and tools that a company use to have one consistent and accurate source of truth.
MDM is essential for large, enterprise scale companies, that are growing organically or acquiring other companies quickly and want to make data actionable and able to be used for decision making. If a company distributes or sells data, MDM is essential. When I worked in market research we made heavy use of MDM to give accurate data to customers.
For a lot of companies, customer information is a good example of master data. It changes infrequently, but needs to be managed as the customer changes their name, moves or their situations change. The data needs to be mastered in one system.