Leveraging benefits from Big Data- Interview of Mr. Jayesh Shah, MD & CEO, Prism Cybersoft
What is big data?
The term big data is used for data sets that are so large that traditional tools of database management don’t work on them. Managing them requires special approach and technology. Data sets now are bigger than what it used to be because data is now being captured by multiple devices like mobile phones, cameras, RFID devices, sensors etc and also because of prolific use of social media where the culture of ‘likes’ and ‘shares’ reveal a lot about likes and dislikes of people.
How big is big data today?
The volume of data generated today is enormous. It is said that 90% of the data generated in this world was generated in the last two years and this period is only going to get shorter. In terms of business investments, the size now is about USD 25-30 Billion and will cross USD 50 Billion soon. In terms of data size, we create 2.5 quintillions bytes of data each day. That is 25 followed by 17 zeroes in terms of US measurements.
Where is it being applied?
Big data is now being applied everywhere. Apart from financial services where we are discussing, big data is being applied to spot business trends, predict whether, predict and combat diseases, in science, space exploration and research, media and telecom. In short it is being applied everywhere.
What are the challenges in Big Data?
While the benefits of big data are enormous, having the data itself is not enough. There are several challenges that need to be overcome. Some of them are –
- Having a good analytics team to convert big data into actionable insight
- Most companies are unaware of what policies to put in place and what data to capture. Even now Indian insurance companies quote car insurance rates even without knowing whether the driver owner is a male or a female. Unfortunately it is the entire industry that works in the same way and even if one single company wants to improve the quality of data capture, it is tough for them to do.
- Many institutions are worried about the potential cost implications of putting in place more hardware, analytics software and data analysts
Financial institutions have however realized that this tsunami of data cannot be avoided and infact can be leveraged to make business intelligence out of it that can be precious.
What are the various dimensions of big data?
There are three popular V’s associated with big data. They are – volume, variety, and velocity. Volume is how much data is collected. Gone are the days when brokerages used to work on only live prices. Now they want to store tick by tick data. Imagine the kind of data storage that is needed. Velocity of data is how fast data is being processed. Variety of data is the different kind of data structures in which data is represented. With social media, it can be videos, pictures, likes, tweets, posts, usual fields and so on.
As you can figure out, the complexity of data analysis grows multifold because one now has to analyze from these different kinds of data and sources.
What is its role in the context of financial services industry?
In a survey conducted, about 62% of the companies felt that big data has significant potential to create advantages for them. This really tells something. The business case for such implementation is very strong as companies can draw insights in their business that was previously not available. Gut feel can be replaced by actual facts.
But before companies can really start taking advantage of big data, they must evolve themselves as data centric organization that is comfortable with working with data.
Who is responsible for data within a company?
Generally it is the CIO who is responsible for big data within an organization. However, it is the business and analytics team that must own it and benefit from it. Financial institutions are seeing big data as a technology challenge. That must change. It is a business challenge and opportunity.
What are issues around analytics in big data?
Financial Services Businesses have all the data they need for better analytics. They just need to define what they want out of it. It is important to have clarity on the objectives of big data implementation. This has significant downstream impact. The kind of analysis needed will determine what data to capture, what to ignore, the data format etc. Organizations also need to have the right kind of people to make sense from the analytics generated.
What are the differentiators of big data management Vs traditional data management techniques?
Traditional data management techniques involved querying a database and working with small samples of data. Big data on the other hand involves working with very large amounts of data in different formats and in different storage devices. Generally, software for big data access allows parallel access of data and processing. Traditional methods involved storing data on hard disks but cloud has pushed big data into a different realm by storing data in cloud and making localized hardware storage redundant and useless. It has also reduced the cost of storage drastically. Big data normally works across different data sets and formats simultaneously and has several sophisticated algorithms that analyzes quickly through both structured and unstructured data and updates results on a real time basis. The business insight gained is also unparalled.
How can financial services companies leverage big data for business benefits?
Financial institutions can benefit from big data in a number of ways. It gives very deep insight into customer behavior and financial institutions can profit from that. It gets a 360 degrees view of the customer, preferences, likes and dislikes. Social media analytics can also tell an institution what is in favor and flavor and what is not. Big data analytics can also lend a lot of power to the marketing team to segment clients correctly and evaluate what offers to roll out and when. Many a times marketing teams roll out festival discounts without understanding whether the offer will be profitable because of increase in volumes or unprofitable because of reduction in margins for the company. This kind of analysis can be very easily done by understanding the customer behavior. Another area where big data can add enormous value is risk management. This is because now it is possible to look at different aspects of data and behavior simultaneously.
Can the issue of customer retention also be addressed?
Customer retention is a very big issue in India because India as a country is very value conscious. By knowing the thresholds of acceptability, institutions can price their offerings more intelligently and even roll out retention offers that are more likely to be accepted by customers rather than offering the same prices to all. Generally, several aspects of customer service can be addressed and improved using big data. All this will automatically increase customer retention.
What benefits does it bring in customer centricity?
The organization can know much more about its customers, dramatically increasing its focus towards them. It can customize and personalize all services by analyzing what the customer wants. As customers we see several airline offers for a place we have searched on search engines. However, do you know that most of these offers are actually customized just for you? It is not something which is offered to all customers. A lot of offers on pre-paid mobiles that customers get are customized for them after studying their usage pattern.
What are the general concerns about dig data management?
Just as there are several benefits, there are several hurdles around big data and its implementation. Some of them are –
- Organizations are not willing to bet a big change in their data management strategy
- They may not be in a position to assess exactly where big data projects will benefit them
- Poorly defined data definition and data policy leads to poor capture and retention of data
- Lack of tools and people who can leverage the analysis that big data can generate
Insurance companies, especially in developed nations have used data analytics for a long time. What is the situation in India?
Insurance companies in developed nations are known for their technology adoption and extensive use of analytics. They are used to predictive modeling to analyze their customer behavior, conduct segmentation, roll out offers and forecast claims. However in India insurance companies are just beginning to adopt behavioral modeling and analytics. Since most of the business here is controlled by brokers and agents, pricing power is limited. Things will only improve as we go along.
What benefits can big data analytics can bring to brokerage and asset management industry in particular?
Like other industries, brokerages and asset managers can benefit tremendously too from big data analytics.
Sentiment of clients can be analyzed and accordingly sales can be done and offers can be tailored. Big data analytics can tell organizations when is the right time to introduce and withdraw new products and services. By analyzing customer’s responses to offers made, pricing can be fine tuned. By analyzing commission structures, channel efficiency can be improved and their remuneration can be fine tuned. Better margining can be done and better risk management systems can be put in place. Trading sentiment analysis can be done and compliance programs like Anti Money Laundering and Insider trading can be put in place. These are some very strong outcomes of using big data intelligently.
How should an organization go about implementing big data projects?
First and foremost, the organization should be convinced that big data is beneficial for the organization. The business case must support the implementation. It should establish the business area where the big data project will be implemented. For example, an organization may decide to implement it in the area of commissions that it charges to its clients and it pays to its channel partners. Then it must go ahead and carefully define the structure of big data set up. This will include definition of what data to keep, in what format and what data to let go etc. This will also include the data governance structure and also roles and responsibilities of team members who are involved. Then, it must conduct a small pilot to see how the management, extraction and analytics are working. This will act as a dipstick to see if the project is really adding value to the organization. If the results meet expectations, then the organization must go ahead with full fledged implementation. Else, corrective measures must be taken. After the project is on stream, a very careful, periodic analysis must be conducted to ensure that such an implementation is actually meeting its objectives. A lot of time such projects and its analytics will need fine tuning.
Organizations must realize that this implementation is different from organization to organization and hence deep patience, planning and insight is needed in every stage of implementation.