BR Research recently sat down with Stephen Brobst, the Chief Technological Officer for Teradata to discuss the science of big data and other tech developments. He is widely regarded as a leading expert in data warehousing and joined the Teradata team in October, 1999. He has served as an advisor to the National Academy of Sciences in the area of IT workforce development and also served in Barack Obama's first term as a member of the President's Innovation and Technology Advisory Committee (PITAC), a working group of the President and Council of Advisors on Science and Technology (PCAST).

<B>BR Research: Please tell us about your role at Teradata and what the company is all about.</B>

<B>Stephen Brobst:</B> As the Chief Technology Officer (CTO) at Teradata, I have responsibility for the technology strategy and the deployment of the technology solutions in the field globally for the company. In Pakistan, we have our local customers and our local teams, but we also have a global development center. Actually we have two of them, one in Lahore and one in Islamabad, which we use to create talent pools that serve globally and not just Pakistani customers. Of course, Pakistani customers benefit from that also because a lot of expertise is brought here from international experience, but for us it's a business model that works very well because we have talented enthusiastic people that are happy to travel when they are required to and deliver effective solutions that we need.

The Lahore facility for the global development center (GDC) is relatively new. We have the one in Islamabad for quite some time now, and recently we opened in Lahore to expand the capability; and we continue to hire in Lahore. Currently, the GDC at Lahore is roughly half the size of Islamabad, but it's growing faster. It is a good place to get and retain talent as there are a lot of universities here.

<B>BRR: How long have you been operating in Pakistan and what sectors are working in?</B>

<B>SB:</B> In 1999, the Teradata division was formed under NCR. Teradata came in to a Pakistan around 1999, and the first customer we had was NADRA. Then later we expanded in to mainly telecommunication companies like Ufone. Today all of the telecommunication companies in Pakistan are doing business with Teradata. We then expanded from telecommunications into financial services, which is strong in all the three major cities. Karachi does not have much telecommunications, but there is a lot of banking in Karachi and we also have some banks in Islamabad and Lahore as well. The third area we focus on is government with initially the federal government, but we have actually been working with the provincial government here in Punjab quite significantly as well in a number different efforts related to crime prevention, health care, tax compliance and other areas give a lot of value. We are also working with PIA as well as in the manufacturing sector.

<B>BRR: What sort of services do you provide to the telecom and banking sector? Is it data warehousing or big data analytics or a combination of those?</B>

<B>SB:</B> It is a combination of things; so we obviously provide data warehouse platform, which is the technology that they use to deploy the analytics. We provide them human resource and talent to develop their analytics and do their implementation. We have analytical customer relationship management (CRM) and software analytical applications that are used by many of the telcos for doing event-based marketing, targeting and segmentation and those kinds of things. Then we have the data model, which has been implemented widely in Pakistan for telecommunication companies, both for fixed and mobile business.

<B>BRR: What kind of projects are you doing with the Government?</B>

<B>SB:</B> There are number of projects related to law enforcement and safety. We have done some work in health care especially related to dengue fever tracking and prevention. Although some projects were commercial based, many of the projects are being done as part of our CSR initiatives.

<B>BRR:</B> Tell us about big data. What prospects do you see for big data in Pakistan and do you see any progress being made in this sector in Pakistan?

<B>SB:</B> Big data is bit of a hyped term. People are confused by this term. It's really not the bigness of the data that matters but the diversity of the data that really matters. We should call it diverse data not big data, but the term has become prevalent and I predict very soon we will stop using it. It was important initially in differentiating a new kind of data, but now the distinction is so apparent the term does not actually need to be used. For example, nowadays smart phones have become so common that people have stopped calling them smartphones. They just call it a mobile phone.

So I think the same thing will happen with big data, it will just be data. You don't care if its big data or small data or triangular or square data, it's just data. The main idea is how we exploit the content and diverse data sets regardless of what we call it.

If we look at big data exploitation there are three phases that are easily identifiable. The first phase, which is already history by now, is web logged data; it's not just what people buy online but every click and every search that lead up to the buy.

Pakistan is not a very strong player in online retailing so the companies who serve Pakistan are collecting and analysing that data, but it's not happening within Pakistan that much. On the other hand, online banking is something that they are definitely interested in analysing. Also, for airline companies, it is useful to analyse online booking and airline flights that people are looking at along with their purchasing patterns and finding out who's the looker and who's the booker is very interesting. So the web logged data that gives you the interaction below the level of the transaction is the first phase of big data exploitation. I think Pakistan is starting to look at that level of detail for certain industries.

The next stage of big data exploitation is social media: What are people saying about my brand, my products and my competitors and so on. In Pakistan that is probably most interesting to the telcos and the banks which are our main customers. The government is also interested in this to some extent as well because they want to understand which programs are working well and which aren't and what is on the minds of citizens. So in this scenario, there is a lot of interest, but the exploitation is relatively in its earlier stages.

The third phase really has not been widely adopted in the market place yet. I would say the third phase is sensor data. Being able to track every vehicle, every part in the supply chain, every item on the retail shelf; this is not really happening in Pakistan yet. The sensor isn't here yet but it will be eventually. One of our customers for example, Volvo cars whose brand is all about safety, uses sensors to increase the safety of the vehicle. They have made a commitment to the market place that by 2020 no one will die in a Volvo car, which is a pretty big deal. Sensors will eventually be everywhere from manufacturing to home applications. You can use sensors in your home for automatically monitoring the temperature and the lightning which is the concept of a smart home. This may seem like science fiction to you here, but for example in California building smart homes are emerging common practices.

However, even here a lot interesting things are going on and we have had some very interesting conversations with the Karachi electric utility company (K-Electric) for a smart grid and smart meters, which is a form of sensor technology to manage capacity more effectively.

<B>BRR: What do you think is the time frame for sensor technology to become the new norm?</B>

<B>SB:</B> It depends on the industry so in industries like the electric utilities. I think that's going to be in the next 3 years, would be my guess. Utilities are interesting because they are typically government granted monopolies, so the incentive to intervene is typically not so high. But because of energy shortages and rising energy costs, there is a lot of motivation to be smarter about energy which is why you see the adoption even though there is lack of competition. I would say with in Europe and US it's largely already a norm but soon it will come to Pakistan as well.

<B>BRR: Where do you feel Pakistan stands in terms of technological competitiveness compared to its peers like India, Bangladesh and Sri Lanka?</B>

<B>SB:</B> That again depends on the industry. Telecommunication in Pakistan is extremely competitive so I would say that Pakistan is at a global best practices level. So I'm not talking about comparing just India and Sri Lanka, I mean really at a global level. I think Pakistan is above the level of your near neighbours that you just mentioned. For example, Telenor Pakistan has been awarded international prizes for best practices not just regional but international level. So you definitely see that innovation thrives where competition is greatest. There is a lot of competition in that particular market place in Pakistan. There has been early investment and very sophisticated use of technology.

When it comes to the masses, I would say there is definitely somewhat of a lag in adoption and its disproportional rite in Pakistan. There are people who have less opportunity, so obviously it creates a digital divide that is larger here compared to other developed countries. I think that there are some government programs to help provide access to technology, for example, internet access but it takes time and money, and both of those are not easy.

<B>BRR: You talked about E-commerce and online retailing. There is no comprehensive e-payment system in Pakistan currently. What do you think are the reasons why we don't have an e-payment system in Pakistan?</B>

<B>SB:</B> I think a lot of that has to do with political and regulatory constraints. For coming in to a country, Pakistan in particular, you are going to banking; it's a lot of regulatory and government hoops that you have to jump through, which takes time. Additionally, companies evaluate the market place to gauge if there is readiness for adoption of new technology. For there to be demand for e-payments you want to assume that people have credit cards, smart phones and at least access to a laptop or desktop and technological penetration in Pakistan is not as ready as it is in other countries. So investors and companies go where the money is.

<B>BRR: We have been hearing about the Internet of Things (IoT) as the next big thing. Can you please tell our readers in simple term how it can be beneficial for them?</B>

<B>SB:</B> IOT is just another marketing term for sensors everywhere. So IOT is really about sensor technology. So you have consumer IOT, which is smart homes and smart vehicles. Then we also have industrial IOT, which is more in manufacturing like when to do maintenance on machines and measuring quality at the products that you have produced. IOT will go through different phases of maturity in terms of relatively simple sensor data analytics to much more sophisticated interaction between things.

The initial IOT is not in fact really IOT. It is the intranet of things, not the internet of things because most people companies at least from the industrial internet of things point of view are initially using the sensor data only within their own company. I would call that the intranet of things not the internet of things. But then once they have produced a product and put it out in the field and then it is collecting data and delivering it back to the company, which allows them to do maintenance on the field of the product, that they have manufactured is what will be the internet of things.

<B>BRR: So big data is mostly for big firms. But in Pakistan 90 percent of the firms are small and medium enterprises (SMEs). Does Teradata have a strategy to focus on emerging market SMEs such as those in Pakistan?</B>

<B>SB:</B> So we have versions of Teradata that you can deploy in a public cloud so most of the small companies prefer a cloud solution not an on premises solution. So a big company typically has a lot of data gravity in their data centers allowing them to perform data analytics. SMEs are much more attracted to the cloud type of model. So we have released Teradata in Amazon cloud infrastructure earlier this year.

We will be releasing Microsoft Azure later this year, and Google cloud is planned for next year. So the answer is yes, we see that the small businesses are interested in a cloud type of model. An OPEX model instead of a CAPEX model in terms of funding. So we believe that Teradata will be adopted and we are already seeing that adoption take place in the market place for smaller sized businesses.

But you have to realise there's also a sales efficiency point of view. Small businesses can't afford a huge system, and we can't afford to have a sales force calling on people making small purchases. But if we can leverage the market penetration of say Microsoft Azure or Amazon web services, then of course that creates a different kind of sales model and efficiencies where it makes sense to go down to that market. The core of our revenue is still going to come from larger customers. More than 80 percent of IT related expenditure world-wide is spent by the top five hundred companies.

<B>BRR: Can you give us an insight into which field holds the most promise when it comes to data analysis in Pakistan?</B>

<B>SB:</B> So one of the areas which have great potential in Pakistan is in data science. Yes, we can do data warehousing, reporting and analysis, but increasingly there has been a focus on data science, which is different than business analysis. A business analyst is looking for an answer to a business question and Pakistani customers have been using our technology for a long time to do that.

Data science is doing more data R&D, and much more advanced exploration and visualisation type of work. In Pakistan, the talent here is being developed as part of a program we have invested in, which eventually will lead to more of adoption of data science techniques. We have support for open source technology, which is increasingly popular in Pakistan. We Aster data, which is a technology that allows you to do more advanced graph processing, social media analytics and more advanced type of data science work.

I think data science is particularly important to the Pakistani market place. It's still at early stages of adoption but growing quickly so that is an area you want to look at particularly in financial services, telcos and the government sector as well.