Analytics: A key strategy for enterprises

CIOL: How do you see the growth of analytics market globally and are Indian enterprises adopting analytics? Where does IBM see the adoption and benefits of analytics in India?

Deepak Advani: In analytics, it is important to have data that you can analyze. In the study we conducted with Oxford Universities, we also tried to find out what is happening in India and Latin America. What we saw was a larger percentage of people in India felt that analytics gives a competitive advantage.

Interestingly, when lot of people start with analytics, in certain industries, they tend to adopt analytics faster than other industries. And even within an industry, certain used-cases showed adoption of analytics faster than other use cases.

In some cases, India is actually ahead in their thinking than the others. For instance, telcos in India are absolutely all over analytics as it is a big deal to reduce a telco churn.

The interesting thing with Indian telcos is that they are pushing us even harder than some of the US and other European companies, when it comes to real time analytics. We have to have social media tools in India and Brazil as social media is huge here that generates high data volumes. They have to apply interesting Natural Language Processing to understand some of the sentiments like slang words, etc.

Indian telcos are saying that they want to figure out in real time how calls drops are happening to some very key customer who uses lifetime value. They want to make offers to them to hold for a couple of minutes instead of a week from now as they will tend to get irritated and unhappy with the damages done. So they demand us if we can build predictive models and than can run in real time so that they can make real time offers when calls are getting dropped.

So that is one instance where a lot of telcos are using 70 per cent data of the network data that they have and are not even analyzing other data. So analyzing network data is a great example.

In banking I see a lot of deployments happening here in India. There's a bank that we have, which is present in over 10,000 different locations and has 70 different databases from where we have to pull the data out. After which the data is cleansed, integrated and taken to a trusted information platform to do data analytic. So we have worked with them to build the right kinds of dashboards and key performance indicators, which help them to figure out what is happening every day.

We are also working with Excise Duty Controls, who wants to figure out on a daily basis about their collections and dive a little deeper into that. So used-cases have started emerging here in India and I want to put special emphasis on India. If you look at the power and energy sector, there are so much of value-creation opportunities. So that requires the right kind of infrastructure that is instrumented for the right type of data. It's only a matter of time.

CIOL: There's a lot of talk around big data and analytics. So from an enterprise perspective, how do you think that enterprises can strategize using big data tools?

Deepak Advani: A lot of people are talking about big data, but everyone has a point solution today and there is no company that comes close as the holistic depth and breadth that IBM has in big data analytics. We can process lots and lots of data.

We call it data in rest and data in motion. People talk about the 3Vs - volume, velocity and variety but perhaps there's the fourth V called as veracity. So we look at it holistically. We said that some of the data will be structured and some unstructured. You need the ability to manage structured data; some of it may be on the mainframe, some of it maybe on Netezza- like appliances and you got to analyze that.

Some may be unstructured and then you need to have a Hadoop to produce a scalar architecture. So we have got elements in our portfolio that do all of that from a data management point of view.

Then you have to layer on different type of analytics. And that's when you look at different types of portfolio. We have got predictive analytics, mathematical optimization libraries with iLOGs, business intelligence so you can visualize what's going on. And then you also have to get into things known as text analytics, Natural Language Processing, the ability to understand different tongue discipline.

If you look at Watson, the jeopardy machine, that's the ultimate case study of a cognitive system, that is a big data system. Among all this data, you need to understand the language and have this system as a learning system, which gets smarter as more and more data gets fed to it. We've got the largest Math department in a private enterprise with over 400 mathematicians and scientists for all the algorithms.

On top, when you say big data, you have got all the structured and unstructured data and even in social analytics, which tells what is happening outside your enterprise, you need to be able to manage that. Then you have got different types of analytics.

But then the other things that really sets IBM apart from others is when you start applying industry and domain expertise on top of analytics and big data. So for instance if you look at understanding people and predicting their behavior, the key to do these is that we have the Unica, Coremetrics, Tivoli, SPSS, a lot of these are pretty much effective.

If you look at risk, there's a huge opportunity for value creation. We acquired Algorithmics in open pages, it made IBM the leader in quantitative and qualitative risk management, we have got thousands different industry specialists, better services so that we can deploy analytics to get a lot of value.

We need to start putting all those pieces together and so the strategy of big data is more holistic. You have different types of data, apply different types of analytics and in some cases, a combination of different types of analytics to optimize the end decision.

What we really focus on when we talk about big data, it's easy to get enamored with the technology, the algorithms, the pettabytes and zettabytes, but at the end of the day, a lot of customers don't care. They want our assistance to get a better outcome. The fact is when we engage with a customer on big data, we may start a conversation like we are having just now. And because we got a broad portfolio, we can pick and chose that makes sense.