Financial institutions retain much of the data they use that relates to their customers. Some is retained for legal and compliance reasons; other data is kept because it could be useful in marketing additional products to customers. And it costs money to store data. Although data storage costs are going down on a unit cost basis, one bank I worked with saw its total costs increase because it was storing more and more data. When I started working with the company, it already had several data centers, and the amount of data being kept grew 6 percent in just the four months I was there. None of the data had to do with my project–this was simply the natural growth rate of data storage in the company.
Some of the increase was related to company growth, but that didn’t come close to explaining all of it. Data storage costs were spiraling fast enough to cause concern and appeared to be out of control.
The manager of the data centers didn’t have detailed information on the data being stored. He ordered some research into the type of information being kept and whether it was ever used after storage. When the research revealed that email traffic was a big component of the stored data, the manager asked himself whether all the emails that the bank sent or received had to be stored. The problem lay in determining which emails might be needed later. You don’t know what you need until someone asks for it, so it is almost impossible to say which emails would have to be kept. Still, he figured that there had to be some way to reduce the amount of email being stored.
The next day, as he was using his BlackBerry and scrolling through a seemingly endless email chain to find what he needed, he had a brilliant idea.
“What if we delete the disclaimers?” he wondered. Did the bank really need to retain that long paragraph that describes what you should do if you receive the email in error?
Every reply to an email contained all the previous emails in the chain, and the disclaimer was repeated on each and every email. He didn’t know how much of the stored data disclaimers accounted for, but some quick research by his team showed that the bank could save at least the equivalent of an entire data center by deleting them.
Once he had this information, he took it to a friend in the legal department. “No way can we cut the disclaimers,” his friend told him. Wise in the ways of our change process, the data manager persisted. “Are you sure?” he asked.
In this case, the company had Incorrect Information to make the decision (how much storage space the disclaimer was using), and the lawyer initially made a Bad Assumption (that the disclaimer had to be kept for legal reasons). There was a substantial cost to storing unnecessary information. When the change process broke the Incorrect Information and Bad Assumptions barrier, the savings in the data storage area were huge. Information overload makes it difficult for people to sort through data to find the right material to make a decision, and information overload can also prove very expensive in its own right.