One of the world’s leading analytics solutions provider and a data warehousing powerhouse, Teradata is known as the gold standard in data warehousing. The Ohio headquartered company made several key acquisition in the last few years, especially Aster Data Systems that strengthened its solution portfolio. Another key driver is the company’s focus on doubling down on addressing the analytics and data management requirements of clients. In an interview with Analytics India Magazine, Rajesh Shewani, Head – Technology and Solution Architecture, Teradata India talks about the long-term strategy in India and how the solution portfolio has evolved over the years.
1. Can you tell us about your role in Teradata India?
I lead the technology and solution architecture for Teradata in India. Essentially, we are deciding the solution strategy for the country in terms of what makes sense to take the market beyond, i.e. if a customer will be able to relate to it.
Two important aspects to take note of here are:
a) Solution strategy
b) Furnishing those solutions and services to clients, prospects and partners
My team and I work with our customers to give them an idea of what capabilities we offer and try to map it through their business requirements. Before getting into business requirements, we highlight to them what are the possibilities across their industry. For an instance, take a bank, they might be aware what analytics is, but they will be able to relate better with real-life use cases. Everything we do is to make our prospective clients understand what Teradata could do in their industry, and specifically in their organization.
2. Can you give us an overview of your clients in India and across Asia Pacific? The two chief aspects of our solution portfolio are:
- Data management
Our portfolio has also shifted as the market evolves. Now, Teradata is close to a 40-year old company. It started as a data warehousing firm, combining hardware and software. However, the market changed quite a bit due to open-source and cloud coming into play. There’s a lot of analytics solution requirement that customers have asked.
What we lead with is business outcome, rather than simply focusing on the technology part. We created something we refer to as BVF (Business Value Framework). BVF further gets BIOs (Business Improvement Opportunities). If you integrate analytics it provides answers like, where will our business get impacted, where will it improve, cost side, process side, etc. We created BIOs for the same reason, and created it to be applicable across 8-9 industries. BIOs are a part of a larger Business Value Framework (BVF). For instance, let’s consider banking. There will be over 300-400 BIOs. Those BIOs are set into functions. For eg, there is a BIOs that specifically performs customer analytics, and there’s a BIOs that’s specifically designed for retail distribution. There’s a set of BIOs related to before operation and each has a usage.
3) Can you provide uses cases against different domains such as finance, healthcare?
Use case: The task is to increase the wallet share for a customer. For this, our team had to do a lot of cross-sell, upsell, marketing campaign, personalized messages, capturing the feedback, all of that, and these shall constitute my outcome. But what is the type of data needed for this outcome? Is there customer demographic information? Do we have all the transaction data and behavioral data? Next question would be to ask what kind of analytics would make sense. Should we perform graph analytics, or is it good to get a 360 degree dashboard. How can one finally determine the success of the outcome?
So, for each of our customers across the industry, we have created these BIOs. For eg, first step would involve getting a 360 degree view of my customer. Once I know the client, doing personalized communication for targeted campaign is easy.
There’s a lot of shift from capturing data as transactions. We are trying to capture both transaction and interaction for two of our clients. We are trying to capture every click on the portal. We can’t wait just for transaction, we have to capture interaction data and combine it with transaction data. Moreover, it’s not only limited to customer information, but also includes machine-related information.
Case in point, a person visits a website to check for the eligibility for a loan. At this point, he hasn’t taken a loan, rather just thought about it and hence, wanted to find out whether he will be eligible to take the loan later. So the system starts capturing these interactions in order to better project analytics based on customer interaction for events which are yet to happen. Basically, interactions may include a simple call to any call center for an inquiry.
4) Robotics process and machine learning -based automation have taken over the path. Companies are strengthening their portfolio through acquisitions and partnerships? Has Teradata made any key acquisition in automation space?
From a strategy point of view, there have been many acquisitions. One of the key acquisitions was of Aster Data Systems, a discovery platform. Then the acquisition of Hadapt and ThinkBig Analytics allowed Teradataa to bring in consultants from the data sourcing and Hadoop background.
5) Can you give us an overview of different types of data captured and how?
• Different types of data click stream, various kinds of log data
• Different types of algorithm – data mining, machine learning, deep learning
• Core technology – is basic computers and servers and the technology available to process a huge load of data in an efficient way in a short time. Also there is the application of Cloud, Big data, and Hadoop as well
6) How is AI and machine learning playing a key role at Teradata (Aster platform)?
Aster as a platform is a discovery platform. The Teradata Aster Discovery Platform reduces difficulties associated with big data analytics. In a way it is a multi genre advanced statistical platform, capable of doing path analysis, churn data available before a particular event occurred, in order to better analyze the event flow. In that manner, we get an idea regarding the customer path, meaning how and which actions are going to follow based on the user’s past activities.
The second key area is Graph analysis – that represents how different events are inter-related. For instance, LinkedIn uses graph analysis shows people you may know. It actually analyses and tallies different key points from users’ profile. Aster was constantly used in the LinkedIn’s base engine to determine different data collected across the user base to find out which profiles may be related and hence the suggestion comes in. In machine learning and AI point of view these techniques are very important and have very high application in creating statistical and calculated approach with high resolution.
7) Any uses cases where cloud based AI solutions are used in India?
In digital marketing, high volumes of open texts require analysis to get customer feedback on brands and products. It can help in determining future business plans in product sale enhancement based on customer interaction, feedback and demand.
8) Could you list down some of your clients in India and give us a headcount of data analysts in India?
Globally Teradata serves the world’s biggest companies and biggest brands such as Barnes and Noble, P&G, American Red Cross and many other. In India, LIC has been a client for a long time. Their entire policy related analysis happens through Teradata. Then there is Aditya Birla Financial Services group and we partner with Indian Government on some initiatives as well. We also have retail clients and in telecom we have Aircel and Telenor.
As for the staff, we have 2000 data analysts in India.
Try deep learning using MATLAB