Banks need big data to trump Google, Amazon: analyst

Until now, banks have had a stranglehold on the best source of behavioural data to tailor products to their customers — credit card records — but they are about to be "left in the dust" unless they start using big data techniques to analyse other data also, according to one analyst.

Speaking at Bank Tech 2012 yesterday, Ovum research director Denise Montgomery said that the latest generation of data, such as geolocation data, is much more powerful than banks' credit card and transactional information, when used in combination with other data. And combining data sources is something organisations like Google and Amazon already have a lot of experience in.

"Putting data together from different domains, such as retail location and spending, will, I think, lead to innovation and products and ... the concern there is that the Googles and Amazons — they've got the drop on the banks."

While Montgomery said that it was too soon to tell whether Google and Amazon would cut the banks' grass by offering financial services to customers, leaving the banks as nothing more than deposit takers, she did recall a conversation she had with one Australian bank, which demonstrated the degree of complacency she felt existed in the finance industry. The bank in question had stated that it was waiting to see which way big data was heading, with the aim of copying or buying up companies that had a proven model.

"I don't think that's taking into account that banks aren't the gorillas in the room any more. You don't just wade in and take Apple or Google over."

Google in particular has a first mover advantage in that, during its journey to build a web search tool, it invented the underlying technology behind Hadoop, a commonly used software framework that is often employed to help analyse large amounts of variably or lightly structured data.

"[Google] could not get the results back from their queries and searches on their data [efficiently], so they developed a new kind of data processing framework, known as MapReduce, and that is the core of big data," Montgomery said.

Amazon, on the other hand, has already implemented a three-tiered architecture that goes some way towards putting big data into a usable form, while overcoming speed limitations associated with analysing giant sets of data.

In its architecture, it has a bottom tier Hadoop "staging area" to rapidly bring in unstructured, relatively raw data. Its second tier is a traditional "warehouse" environment that is for the more structured and controlled data that finance and risk business functions are able to use. The top tier contains customer behavioural and relationship data.

"We're no longer thinking that big data and analytics are something that can wait. People have realised that they're part of the ticket to the game and we actually need to be working on big data now," Montgomery said.

The banks do seem to understand to a certain extent. She said that, while the finance sector generally said they had "bigger problems than big data" last year, all the major banks have launched some form of big data pilot this year.

Indeed, ANZ is one such bank that is already using big data. In the past, it would have set up several meetings with a business to ascertain the business's needs before offering them a product, but by analysing the business using big data, it is able to tailor a product within a shorter time frame.