A lack of skills and resources mean threats are going undetected.
Government networks host cyber threats for an average of 16 days before they are detected.
This is in part because of large amounts of data with a survey showing IT pros stating that their organisation is overwhelmed by the volume of security data.
Over half cited a lack of resources and specifically skilled personnel and nearly 8 in 10 stating that at least some of their security data goes unanalysed due to a lack of time and or skill.
So says the "Go Big Security" report from Splunk and MeriTalk.
However, despite the majority agreeing that big data analysis will improve cyber security, only 28% are currently using it.
Kevin Davis, VP, Public Sector, Splunk, said: "Government organizations have access to a wealth of cyber threat information."
"The challenge is managing that data and connecting the dots in real time. That’s how we get immediate insight into threats. Agencies need to detect threats faster and start to predict when and how they will occur."
According to government cyber security professionals, Big Data will help to make cyber security risk management more effective and proactive. With 76% stating they their security team often currently operates reactively rather than proactively.
Big Data would help to better detect a breach in process (61%), monitor streams of data in real time (51%) and also help conduct a root-cause analysis following a breach (49%).
Steve O’Keeffe, founder MeriTalk, said: "Moving from compliance to risk management is a mindset shift."
"Agencies need to think about ‘big security’ alongside big data. CDOs need to be on the court. Data is the MVP."