15th November 2012
For the past seven years, the business and IT communities from across the country came together at a conclave called the IBM Software Universe, to discover the next technology trends and how businesses could leverage on these trends to stay ahead of the competition. Now in its eight edition, the theme for this year's conclave in Mumbai on November 6th was 'How software can transform the way we live, work and play' -- with the tag line 'Meet Possible'.
The overarching technical theme this year was Big Data and Analytics, but there was a lot of talk about how the focus has shifted from the back-end to the front-end, with customer facing technologies like Social and Mobility.
InformationWeek caught up with three IBM experts to discuss the application and use of Big Data and Analytics in business and government. IBM is also working with Indian universities and management institutions to provide Analytics solutions and courseware. Brian Pereira of InformationWeek spoke to Prashant Tewari, Country Manager - Business Analytics, IBM India, Deepak Advani, Vice President, Business Analytics Products and Solutions, IBM, and Katrina Troughton, Vice President Business Analytics, IBM Software Growth Markets:
Can you update us on last year's announcement about IBM embracing Hadoop and working with Cloudera?
Deepak: We've got a product called InfoSphere BigInsights that's based on Hadoop. It is increasingly being used for social media analytics. With 34,000 tweets and 600 blog posts being generated every minute, that's a lot of data to analyze and to look for patterns. That's what Hadoop can do. We also have a product called Cognos Consumer Insights that builds on Hadoop and BigInsights and it analyses social media data using the MapReduce architecture. And then there's InfoSphere Streams to analyze data in real-time. You don’t have to store all that data in order to analyze it.
We also have a product called Cognos Consumer Insight (CCI) that does customer sentiment analysis. It also does social network analysis to determine how influential certain people are on the network. It is based on Hadoop and BigInsights.
Can you give us some examples of how companies, perhaps your customers, are actually benefitting from Big Data and Analytics?
Deepak: XO Communications (a telco) used our predictive analytics solution and reduced customer churn by 35 percent. The way to go about doing that is to look at the historical data, going back a year or two. And you look at all the clients and what were all the touch points, and all the actions that customers took. You look for patterns and you also have customer demographic information. So you take all the different data points and build a model, based on mining your historical patterns. So when you start to see similar behavior for different clients the models can tell from the patterns that a person is likely to leave. The models can also suggest what type of offer you can make to reduce the chances of a customer leaving.
Another customer in Europe reduced customer churn from 20 percent to 2 percent. They integrated all the different data they had on their clients, from demographic to interaction to transactional and looked for patterns to run those models in real-time.
Lots of people are now running these models in a call centre. So when someone calls in you have all their details on the screen, and can identify their propensity to churn or their life-time value. Based on what the person is saying, that text is now getting analyzed in real-time, and the offers that you make can also change in real-time. It's important to know what type of offer is likely to resonate with this person, because not everyone wants the same offer.
Katrina: It is a shift that we see, about needing to know a client as a person or individual. So it is about how to respond to this person, and not to someone who sits in this generic group. In a recent CEO study that we did, Indian CEOs saw this approach as one of the most important drivers for business growth.
Which are the verticals where you seeing a lot of use cases for Big Data and analytics?
Deepak: It is across the board. Certain industries are more aggressive in deploying analytics than others. More companies are asking what the telcos have been doing to reduce churn, because they have been doing that for decades. We also see Market Basket Analysis used extensively in retail and now Manufacturing companies want to learn from Retail.
While the Telecom industry is using analytics to reduce churn, Banking is using it to acquire the right type of customers who can give the best life-time value. So they will offer certain services to high networth individuals. Manufacturing is focusing a lot on predictive maintenance -- they want to predict that something is going to fail and act on it before it fails. For instance BMW can alert the driver that based on all the data from the engine there's a 70 percent chance that in the next 40 kilometers the car is going to break down. And its advises the driver to take it in for service. So statistical algorithms are examining all the data and looking for patterns and predicting failure. This notion of predictive maintenance is becoming big, and now energy and utility companies, and oil companies are also asking for this. In retail, they want to predict what are the products that people want to buy, so that they can stock their shelves in advance.
Katrina: Energy and retail companies are also looking at predictive analytics. For instance, a clean energy company might want to know the best place to install a wind turbine. You need to analyze large quantities of big data to make that decision. Weather prediction, flood alerts, natural disasters are other application areas.
Can you touch upon some use cases that are specific to India?
Prashant: We are doing work around taxation and inclusion (data gathering and analysis); we work with financial inclusion authorities to provide them with analytical inputs. A large financial inclusion authority uses our BI solution for analyzing collections and disbursements on a weekly basis. Micropayments in India is not finely regulated. So our solution is helping in gathering data on a weekly basis and shows it on an executive dashboard, to enable management decisions.
We also work with many state departments to assist them in the collection of information. There is a Ministry of Statistics, a Department of Economics and Statistics, that collect data for doing the GDP calculation -- and to figure out the deficit in terms of the index of industrial production. So these ministries and departments are tasked with carrying out statistical and data mining analysis.
The third area is direct and indirect taxation - we work with both departments and provide them with tools and techniques for carrying out in-depth analysis.
Katrina: We have also worked in the Education industry and universities to help improve (technical) skills around analytics, by working closely with state governments, and providing courseware, and in designing the curriculum. It is also about providing products and solutions that address a broader range of problems, and which are simpler to use.
Can you elaborate more on what you are doing to address the demand-supply gap for data scientists? Which universities and institutions are you working with in India?
Prashant: There are two sides to this: the technical and the management education. The SPSS solution, which is for statistics and modeling, is taught in almost every engineering college in the country. Practically all colleges use the SPSS solution for doing statistical analysis, and it is part of the course content. In fact we have an academic team which provides courseware to colleges.
And for management education we have some marquee names, such as IIM-Ranchi. They have included Analytics as part of their course curriculum. Management institutions (and graduates) are beginning to realize that just having a Marketing or Finance degree does not make one unique or immediately employable. So they are looking at Analytics as an enabler to help students become employable.
And we at IBM have an ambitious target to certify more people on our technologies.