Firms will need to reassess the way they manage their data archiving and storage systems in the coming years to help reduce costs, according to IBM.
Steve Wojtowecz, vice president of storage software development, said the need to understand the data being stored and set policies for deleting old information is a key trend the company sees for the year ahead.
As such, Wojtowecz told V3 that firms need to undertake a "data intervention" in which, rather than constantly acquiring new hardware, executives discuss with IT and legal staff policies for when and how archived data can be deleted.
"At some point we are going to get to a tipping point where it is simply unaffordable," he explained.
"There is going to be this notion where there are business processes and procedures to outline what we keep and how long we keep it."
Improved management is one of several trends IBM is predicting for the coming year.
Wojtowecz also foresees a growing need for lifecycle management and data transfer tools as companies look to migrate data from old storage mediums to a combination of on-premise and cloud computing platforms.
In doing so, IBM believes firms will see a need to purge outdated information such as archived email messages.
By removing the "low-hanging fruit" of stored data, firms could slow the costs of storage in the short term, he explained.
Before that can take place, however, Wojtowecz said companies will need leadership to step up and take charge of the data management process.
"Companies are trying to build policies to simply understand what data we keep and using technology to implement the policies," he explained.
"There is no one team or department that is willing to step up and say 'I own the policies for our data retention'."
Last week, IBM said it expects 4G technologies to enhance the spread of social data, which could lead to yet more data for firms to store and archive.
Finger pointed at "advanced" nation state attacker in Norwegian health records cyber attack
Kaspersky claims the ban is based on subjective, non-technical public sources - and unconstitutional
Google unleashes Cloud AutoML tool to enable ordinary Joes to train AI systems without having to write code
Next step: machine learning systems that can generate their own machine learning software
Pixel devices could be hijacked by confusing the engine and escaping the Chrome sandbox