Dealing with big data

Dealing with big data

Make your business more effective

In today's business environment, the topic of big data is ubiquitous. In 2011, the global output of data was about 1.8 zetabytes (each zetabyte equals 1 billion terabytes). Amazingly, 90 percent of the data in the world was created within the last two years.

Yet, the times are such that dealing with big data in the most intelligent fashion has become de rigueur for many businesses.

Take the case of the airlines industry. The airlines business operates on thin profit margins, mainly due to fluctuating fuel costs and multiple taxes. Currently, for instance, China and the European Union are at war over paying emissions tax-all airlines have to pay this tax for flying in and out of Europe. China has said no to it.

In such a scenario, imagine having an algorithm that could help airlines fly more efficiently and save on their fuel consumption. That will help them become more competitive.

Enter Mahindra Satyam. The IT services company is doing exactly that-using its software framework to help some airlines save on fuel consumption. For example, it is helping a major Asia-Pacific airline through advance algorithms crunching several hundred thousand parameters and permutations to see how it can reduce fuel costs and meet tough EU emission requirements.

"A typical engine in an airplane generates 1 terabyte of data during a transatlantic flight," Venkat Narayanan, head of strategic initiatives, Mahindra Satyam, and an expert on business intelligence (BI) told MIS Asia. He was speaking at Management World Asia Forum on 8 February in Singapore.

"So if we collect this data from not only that one engine but all the engines of not one flight but all the flights, besides some other correlating factors such as the weather, altitude, speed, and so many other parameters, what you will be able to do is, for example, come up with a fuel consumption model. The various parameters that determine how much fuel is getting consumed will be developed into a model which will simulate different conditions. For instance, what if I fly at a different altitude? Will that change my fuel consumption? The airlines people can play around with controllable parameters such as cabin lighting to cabin temperatures, speed and altitude and so on to determine the optimal setting which will help them to save on fuel consumption."

Narayanan was at the forum to present his perspective and insights on how BI can transform and grow corporations.

"Asian corporations can harness next-generation analytical tools to grow and transform their businesses amidst a global economic slowdown, increased competition and the rapid emergence of social media," said Narayanan.

"Many corporations face challenges of unclear business requirements, inconsistent data, poor time to market and failure to harvest business benefits from analytical tools," he said. "These challenges come at a time when Big Data is emerging as a major trend and differentiator in view of slower economic growth, increased competition and the rise of social media."

For example, Narayanan said that his company helped Viva Telecom of Kuwait grow its business using iDecisions framework. The iDecisions framework is a BI tool which helps enterprises rapidly adopt the best practices of major corporations by customising data input which can help leaders make effective decisions cutting across various business functions.

"The iDecisions framework has helped us roll out a BI program which analysed how and why mobile customers migrated (or "churned") from one telecom provider to another, the effectiveness of marketing and channel initiatives as well as sales intelligence which

helped increase sales at retail sales outlets," said Arnold Ali Cender, data warehousing and BI program manager of Viva Telecom, who also spoke at the forum. Using this tool, Cender said that his telco could bring the churn rate down by 20 percent.

Harvesting data smartly

"A good amount of data is outside the organisation's visibility," said Narayanan. A lot of that data can be found on social networking sites. He calls such data low hanging fruit. What is better to do is to look at the data about a brand on the networks and come out with a sentiment meter. "This whole thing can be done in a real time manner so that organisations can start responding in the same media," he said.

He gave an example. When the train between Paris and London got stuck, people started tweeting. British Railways did not even realise this was going on till it became too late. Organisations can't afford this kind of folly any more. It might affect their business. "Today, you need to engage the customer in real time and in the same media," he said.

For example, there is a lot of customer churn in telcos and the banking industry (credit cards). What BI services providers can do, Narayanan said, is to identify the influencers. They can categorise the customers into those who are influencers and those who are not and come out with a customer influence factor. "This can help in attrition prediction or churn prediction," he said. "The BI system can identify the influencers. If they are saying negative things about your organisation, the system can help ring fence them. If they are saying positive things about your company, the system can tell you how to leverage on those influencers."

New trends in data analytics

According to Narayanan, one of the new trends in BI is how you assimilate the big data. "How closer to the action you are using analytics?" he said. "We want to reduce the action distance. Not only data gathering has to be real time, your ability to take action should also be quick and in the same media."

Another trend to watch out for, he added, is to be able to differentiate between fake and real social media chatter. "You know that success brings unsavoury attention," he said. "Today, there are social media sweat shops that can be used to talk good about you or talk bad about your competitor. How do you know those who are saying things about you genuine or fake? Good BI tools should help you differentiate between the two."

Verticals more likely to succeed using BI

"Businesses which are under a constraint or a threat will succeed well using BI tools," Narayanan said. "Take telcos, for example. Earlier, there was no pressure on them to use analytics. Today, they are under tremendous pressure. Their traditional voice revenues are vanishing. People are shifting to IM (instant messaging and VOIP) than voice. That's why today telcos are moving into media and other spaces. Their traditional business is getting yanked out. Right now, they are a provider of a dumb pipe. Companies like Apple are raking in the revenues. In this scenario, telcos are trying to make money by moving from dumb pipe to smart pipe."

Other verticals, according to Narayanan, that can use BI to succeed are banks and financial services sector, healthcare, and retail.

"Healthcare is a major adopter of analytics," he said. "Public healthcare's objective is to make sure people don't get sick. So, population-level healthcare data is very useful for the government."

Doctors are also veering towards evidence-based medicine. "It's an example of analytics-driven healthcare," said Narayanan.

"For example, Singapore going for electronic health record is an inflection point," he said, referring to the Singapore government's US$144 million National Electronic Health Record system (NEHR). The Ministry of Health and Accenture launched in June 2011 one of the world's first national electronic health record (NEHR) systems. This serves Singapore's "one patient, one record" vision. The NEHR enables a single patient health record for clinicians to access across the healthcare continuum. "Under this electronic system, you are getting a single point view of a patient's health records. There are many benefits of this system. Long term potential is high in reducing people falling ill, and there are cost efficiencies."

Defence and security is also one of the big users of analytics but they don't like to talk about it, Narayanan said.

Challenges for Asia

"Language is a challenge in Asia Pacific," Narayanan said. "It is a balkanised region, with many languages. One of the challenges of developing a good data analytics system for this region is the ability to handle the local languages. For example, people in Singapore use a lot of mother tongue words even in English. How would a system handle it?"

The other challenge cost of adoption. "Not all organistions have got the size to get into analytics by investing in it," he said. "Tier one retailers or banks can invest in BI but Tier 2 and 3 companies can't; They can't justify the ROI in investing in analytics."

"Responding to the cost challenge, Mahindra Satyam is trying to promote a cloud based consumption model," he said. They have a "pay as you use" model. "Also, what we [Mahindra Satyam] harvest from the Web, we use it for multiple customers. This allows us to pass on the benefit to individual customers."

The other model is result-space pricing-that is price based on business benefits a user is getting.

"Both the models are working well in Asia," he said. "For example, India has a lot of customers who prefer the result-space pricing service. But it is a risk reward model."

Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.

Join the CIO New Zealand newsletter!

Error: Please check your email address.

More about Accenture AustraliaAppleEUMIS

Show Comments