With rising data volumes and the advent of more powerful software tools for handling this information set to be one of the big trends for 2015, the question many chief information officers (CIOs) will be asking at the moment is why they need to be making this a priority.
This was addressed by research director at Gartner Alexander Linden, who was responding to a Q&A ahead of his organisation's upcoming Business Intelligence and Analytics Summits, which are taking place in Sydney, London and Las Vegas in the first quarter of next year.
He said the overall amount of data CIOs are dealing with is growing in every industry, so these executives will need high-quality data science skills and solutions in place to extract valuable insight from this.
Mr Linden stated it is now essential that businesses are able to determine how to acquire new customers, improve their cross-selling and predict future demand and failures. Normal business intelligence and descriptive analytics, and even traditional software engineering, will not be able to handle these situations as data volumes grow.
"Advanced analytics can surpass human capability in coping with significant volumes of data and dealing with highly complex digital business settings," he explained. "Digital businesses have to adopt data science methods in more use cases, by driving the availability of sensor data, expanding bandwidth and reducing storage costs."
Mr Linden added that extracting valuable information from all this raw data is "not a trivial task", so one of the key elements when looking to make sense of these assets is having people with the right skills.
To make the most of this, organisations need to understand that data scientists are not the same as traditional business analysts. As well as being able to derive mathematical models from data in order to achieve clear and hard-hitting business benefits, they need the ability to network well across different business units and work at the intersection of business goals, constraints, processes, available data and analytical possibilities.
The expert also observed that many businesses make the mistake of assuming the new generation of analytics solutions will be broadly similar to what they have been dealing with in the past. It is a common misconception that once a firm has mastered traditional analytics, it is simply a case of some additional learning and the introduction of extra software tools to progress to the next level. But this is not the case.
He explained 'normal', or descriptive analytics, merely reports what has happened, but advanced solutions use prescriptive and predictive analytics to solve problems.
"Predictive analytics predicts future outcomes and behaviour, such as a customer's shopping behaviour or a machine's failure," Mr Linden said. "Prescriptive analytics goes further, suggesting actions to take based on the predictions."
For example, after predicting a machine is likely to break down after producing a certain number of parts, a prescriptive analytics approach might suggest that the company conducts maintenance before that maximum number of parts is reached, in order to prevent unscheduled and costly downtime.