With data oozing from practically every electronic pore on earth, organisations face a tougher task than ever in trying to wrangle value out of the deluge.
Thankfully, the means of collecting and analysing information has also improved exponentially, with new systems able to tap into vast fields of knowledge, across virtually limitless resources, and make sense of it all - at light speed.
According to IBM, sophisticated analytics systems now have the ability to understand and interpret human speech, identify complex patterns, and gain insight into layers of seemingly unrelated information. This has the potential to improve decision-making at all levels of business, government and public service, the company says.
One particularly powerful version of such a system, which made its debut on the US quiz show Jeopardy earlier this year, is IBM's Watson supercomputer. Made up of 90 IBM Power 750 servers, the system builds on the company's Deep Questioning technology, with the ability to answer questions posed in natural language. It can answer requests based on a number of inputs by interpreting large volumes of knowledge across an enormous range of topics, says IBM.
Shane Radford, business analytics and optimisation lead at IBM SA, says the approach to data in the past 15 years has been to learn from the past. The focus was on automating processes and business intelligence (BI), which saw companies extracting historical information from operational, finance, and transactional systems, and analysing that data.
“The concept was yesterday, today and tomorrow. But while BI focuses on what happened in the past, if you can gain some insight into patterns and see where things are moving to, you can plan for the future.
“If we look at where things are going in the next 10 years, data and information is going to transform the way businesses function. A business with a lot of automated data can take that information across its various channels and get insights it can apply at different levels,” explains Radford.
He uses travel and expenses as an example: “If you use analysis of yesterday's data to detect a pattern of where people traditionally take chances, you can use it to track patterns in real-time and predict the results, and stop the fraud before it occurs.
“It's about what's happening now and what's likely to happen in future, so you can move to predicting rather than simply reacting,” he notes.
These predictive capabilities enable companies to keep a finger on the pulse of customer numbers, how much it costs to service those customers, and the costs of delivering a product, says Radford. “It's all about making sure the business adapts in a more dynamic way.”
Increased competition and the glut of information have led to “a gradual evolution in the market for BI and analytics”, according to Gartner. In January, the research firm highlighted trends that are set to challenge traditional assumptions around these fields.
“By 2014, the metamorphosis of BI from IT-owned and report-centric will be virtually complete for a large number of organisations,” says Gartner. “These organisations will change what types of BI and analytics they use. They will change how they procure them and where they procure them from, and they will modify how information feeds decision-making.”
The firm predicts that by 2014, 30% of analytic applications will use in-memory functions to add scale and computational speed. In the same year, 30% of analytic applications will use proactive, predictive and forecasting capabilities.
Smarter society
In addition to its promise in the corporate world, advanced analytics could have powerful impacts at a societal level. “There are great opportunities for government to work with the private sector in using new analytics capabilities to improve service delivery,” says Radford.
IBM argues that advanced analytics systems like Watson could help better leverage government information on population, due process and policies, which is often extremely in-depth and complex.
When it comes to unstructured data, there's still a strong need for human intervention.
Corine Van Erkom Schurink, analytics team lead, PBT Group
Agencies will be able to look at existing government data and gain more comprehensive information about performance bottlenecks, operational status and interdepartmental dependencies, it adds.
According to Radford, the biggest challenge in government is that it works in silos, as organisational activities are divided into national departments, provinces, and districts. “They don't tend to work across silos very well in terms of sharing information and data,” says Radford.
Combining relevant information from various data stores will allow departments to better allocate resources, he adds, to improve planning for future initiatives and policy changes in areas such as crime prevention, economic development and infrastructure maintenance.
However, Corine van Erkom Schurink, analytics team lead at Prescient Business Technologies, says government often suffers from budget limitations and is slow to react to change and adopt new technology.
“It will be industries like telecoms that will be the prime adopters,” she says, adding that actual implementation will take a while, as the software needs to mature.
“Vendors need to do development work and need a whole cycle of proof of concept in many industries before we see full adoption.”
Radford concedes that deploying these kinds of solutions at government level requires building a solid information foundation. “We'll have to start building a foundation at all levels to drive strategic change at a societal level.”
Another public service area where Watson's technology could be a game-changer is healthcare, says Radford. In February, IBM embarked on a joint research programme with Nuance Communications in the US to determine in which areas Watson could be the most useful to doctors. The plan is to combine its deep questioning capabilities with Nuance's speech recognition and clinical language understanding software.
The ability to capture notes and diagnoses, and render them in a useful way is a powerful proposition, says Radford. “Doctors are essentially problem solvers. They spend seven or eight years studying so they can take the patient's history and symptoms, and use the knowledge they've gained to come up with a likely diagnosis.
“With Watson, you can put all the books and material doctors have studied and all the outcomes of various treatments, to help understand a particular patient's stage of disease and level of treatment required,” notes Radford.
He explains that while there aren't enough doctors in the system, there are lots of health workers and nurses, which are involved in about 80% of the initial diagnoses. Making the knowledge and actual case notes of doctors available in a collective, intelligent system could allow it to indicate solutions with a high level of certainty, helping to bring healthcare to underserved communities, says Radford.
Social intelligence
An area witnessing possibly the biggest explosion of unstructured data is the world of social media, which offers a mine of information about customer sentiments and competitor activity.
Van Erkom Schurink points out that companies are starting to adopt collaborative decision-making techniques.
“In the commercial private sector, the sheer amount of structured and unstructured information is becoming almost counter-productive. If organisations want to be competitive and grow with the industry, they need to take what is said on the Internet, on Facebook and Twitter, and derive insight from that into what competitors are doing and clients are saying.
“They then need to feed that back into their BI and modelling systems in order to make appropriate decisions.”
She says one of the major benefits for companies lies in becoming more proactive in areas such as customer service.
“Usually, when a customer leaves, it's because it was the last straw. But if you can model and predict customer relationships and interactions, then you institute preventative action to stop them from leaving.”
Take telecoms churn, for example. “With predictive modelling, you can pick up which people have the highest call rates and devise an anti-churn strategy aimed specifically at them,” says Van Erkom Schurink. “It allows you to target a very specific demographic with a specific action.”
She adds that in the near future, systems like Watson could unlock huge volumes of information not accessible before. “It will give businesses access to softer information such sentiment, which is valuable because it allows companies to take corrective action in the market.”
This includes things like managing the perception of a brand, by picking up what's said on Twitter, blogs and social networks.
However, Van Erkom Schurink notes there will always be a need for human interpretation. “Watson is wonderful because it's got powerful processing abilities and can access terabytes of information, but it still cannot simulate the human brain.
“When it comes to unstructured data, there's still a strong need for human intervention.”
Companies that can both tap into these stores of data and use it to react quickly and effectively will gain a huge competitive advantage, says Van Erkom Schurink.
The challenge now is bringing the two together. “The problem for the further development of advanced analytics is that people don't have the resources. IT has made data available, but that data is not accessible to the business person to leverage it.
She says businesses will need help to understand how to insert analytics into all their processes. But the outlook is positive, with these solutions on “an exponential curve”.
“We're looking at between three to five years in terms of BI getting to a really high level of maturity,” she says, after which more advanced systems will really pick up steam. “At the moment, the data is just sitting there, waiting to be leveraged.”
Share