* Sometimes it's because their boss tells them that they need to make data more secure. Developers at the Veteran's Administration are probably doing that as we speak and so are lots of other shops big and small dealing with government-mandated security constraints or the results of a hacked data store.
* Perhaps the existing data storage system has exceeded its capacity or the capacity of the system to support it. This could mean the database can't store the volume of data, can't support the number of users trying to access it, or the company has run out of places to store the 3x5 card boxes. This might also mean that the application or database is so poorly designed that no amount of upgrading or tuning can cure its ills.
* Sometimes they need to choose a way to share data among co-workers or to publish data in a way that's not only secure but easily implemented with the existing infrastructure. Increasingly, companies are finding new requirements to share data or simply expose it to their customers or the public. Once a useful database is created, it can spread uncontrollably like a virus or a video on YouTube. That's the downside of data sharing.
* Perhaps the data source has changed or a new source of data has been introduced that does not work with the existing system. Data import and export issues are a perennial problem often introduced as companies and departments merge. Many developers spend much of their bandwidth writing import/export/transform utilities simply to move data from place to place.
* Sometimes the choice of a data storage system is a function of long-term study, skill, experience, or simply ignorance. All too often, you hear about people (and companies) that choose their DBMS because that's the engine they know—or it's the one chosen by the CEO's son-in-law. When choosing any tool or DBMS engine, developers, consultants, managers, and relatives all need to make an effort to understand the strengths and weaknesses of all the alternatives.
Data Entry India