Data fusing technologies developed in New Zealand are beginning to play a big part in marketing banking and insurance services in Asia.
Specialist data analytics company Datamine is part-way through its first job in China, providing services for a joint venture formed by the subsidiaries of an Australian bank and an Australian insurance company wanting to cross-sell the insurance products to the bank’s eight million Chinese customers.
Datamine has also done work for similar ventures in Hong Kong, Korea, Taiwan and Japan.
In New Zealand, its client list has several government departments, including Inland Revenue and ACC, banks, insurance and financial services companies, energy clients and supermarkets, and the telecommunications sector.
It’s about targeted marketing, taking the data from a business problem, adding other data, say, from the Statistics Department, and finding the key drivers to customers’ buying patterns, according to company founder Paul O'Connor.
O’Connor started the company in 1995 based on research work he had done at Victoria University.
“I’m essentially a statistician,” he says. “For years, statisticians worked on small samples, but with businesses collecting much more information now, plus disk and processing becoming cheaper, statistics had to evolve.”
Datamine now has 14 staff in Wellington, four in Auckland and one in Manila.
“New Zealand is ahead of the game because our Privacy Act came in early and we have had to develop data fusing techniques [to avoid breaching the Act],” he says.
“In Asia, the banks are now facing much more competition, and they’re having to be smarter about targeting customers.”
He says the Asian work has come out of earlier work done in Australia.
The data is fowarded to New Zealand, either by DVD or broadband, and the analytics are done here.
Datamine uses high-powered SGI computers running Linux, which, O’Connor says, have the advantage of a strong floating point function.
Psychology comes into the analytics, O’Connor says. “In pure statistics, segmentation is about clustering. We had to determine techniques to get the data to match the way the business was functioning.”
New Zealand has, on a per-population basis, three times more debit card transactions than anywhere in the world. That data is one example of a wealth of information about usage that can be translated into targeted marketing, O'Connor says.