What happens when you make a purchase at a large store like Lifestyle and put that purchase on your loyalty card?
An entire pre-set list of activity begins to unfold: The store system checks to see if you made the purchase on a particular credit card, based on which the credit card company may send you a discount offer for related purchases; the store loyalty system may check to see your loyalty points tally and send you an SMS encouraging you to utilise those loyalty points; your name and address may be shared with a mutual fund company who can send you an appropriate investment offer, depending on your shopping pattern and spending capability; and a call centre may get in touch to sell you membership to a spa or a holiday home.
Some years ago, life was simple. You went to a store and made a purchase. That's it. The transaction ended there. Today, a purchase is only the start of several transactions, making business more vigorous.
Behind the flurry of activity initiated by that single purchase is a set of rules and policies that examine your profile, filter the possibilities that may add value to you, alert the correct officers, agents and delivery systems. The rules and policies are called algorithms.
Amazon, the online book store, uses algorithms to generate real time data to assist shoppers in making the right choice: Computers at Amazon examine your purchases at the store in the past, combine it with the decisions made by others with similar purchases, and show you a vast variety of choices.
The maths behind this process is complex and it gets even more so when it needs to be done in real time. But, mining data for usable nuggets of information using mathematical techniques is what businesses today need in order to trigger sales.
Creating these algorithms used by airlines, banks, retail chains, credit card companies, medical facilities and just about every business is Bangalore-based Pattern Recognition Research Institute.
"The stuff we create helps decision makers delve deep into the data their businesses are generating to show views that were previously impossible, to forecast scenarios that can save lives and discover drugs that were previously unimaginable," says Sudarsh Kailas, founder of the Pattern Recognition Research Institute.
The algorithms written by Kailas' outfit are helping medical researchers rapidly count molecules that help predict cancer in patients and analyse speech patterns and inflections in a phone conversation to predict if the caller is a terrorist.
The cutting-edge algorithms can be used to predict storms, internet outages, assist in shelf stock management, car number plate verification, signature verification and stock market prediction.
The beauty is the same algorithm can help count people entering a hotel and analyse how many of them spend at the hotel or isolate product defects on a conveyor belt. In essence, large parts of the algorithms are re-usable and can be deployed in a variety of scenarios with minimal tweaking. And thanks to companies like Google and Amazon, we know that the market for such algorithms is rapidly expanding.
Kailas says the Pattern Recognition Research Institute deals in a new form of content technology. He says that businesses that are spewing mountains of data are unable to see logic and patterns quickly enough to act upon them. Processing this content produces knowledge which typically tends to last longer and be more meaningful than the data points.
Kailas' contention is that content technology will now begin to create a layer of value over information technology. "Content technology will make India [ Images ] a unique place in the world," predicts Kailas whose company has customers and partners like GE and Titan.
A company that used the algorithms created by Kailas' team is Cumulus Systems, which focuses on high-end R&D for Fortune 50 companies.
"Using pattern recognition algorithms we developed unique ways of extracting relevant information from pictures. For example a picture of the back side of a CISCO router resulted in automatically extracting the serial number, model and make of the router. The extracted data could then be fed into a database and thus typical questions such as model and make of the hardware need not be entered manually reducing errors," says Arun Ramachandran.
He adds that although he did scour around for companies that could deliver the work they need, he found none in India other than Kailas' company. "They are unique and I hope there are more companies with their mindset in the future in India. They are small, but truly beautiful."
Kailas believes that Indians are naturally positioned to create such algorithms because of their mathematical bent of mind. Historically, Indians have excelled at maths and this is the basis of his contention. Kailas believes that with the world moving from agricultural economies to industrial and information technology, it is natural that it migrates towards extracting meaningful content using the technology.