With quadrillions of dollars on the line, banks and financial institutions pay close attention to the emerging exaflood of available data about their customers and the world around them. Here at the O’Reilly Strata conference on big data, the panel on big data in the banking world was fascinating. It’s likely an indication of the way the rest of the world is likely to move in the near future – at least if you believe the predictions of the people on the panel.
Huge opaque markets are about to become transparent because of new regulations and that means a whole lot of new data available for analysis. Scalable processing of that data will require outsourcing, giving birth to new industries. Millions of people will need to be trained to deal with all this. Below, my notes from this fascinating panel discussion.
Can big data help banks avert the next financial crisis? Could regulation resulting from the last crisis yield newly available data that could become new mega-resources for innovation themselves? Those were among the topics discussed.
Abhishek Mehta (Tresata)
Roy E. Lowrance (New York University), Richie Prager (BlackRock), Allen Weinberg (McKinsey)
11:20am Wednesday, 09/21/2011
Richie Prager: Dodd Frank Consumer Protection Act changes the OTC derivatives market. It was the domain of large financial institutions, measured in hundreds of billions of notional dollars.
Everything in that market that was opaque will now have to execute contracts with transparency, all the executions will get turned into data and all the risk is reported to regulators. What was very opaque will now be a transparent market, reams of new data available.
Allen Weinberg (McKinsey): All kinds of businesses will need to learn to work with this data, new providers will need to emerge to serve them. They have to get outside of silos, it will be a dramatic opening up. Banks won’t have enough capital to build it all themselves, it will have to be an open source model to build the infrastructure.
“We manage over $3 trillion and we believe in analytics so much that we’ve created a dedicated research staff. We actively use tools to understand everything from trading cost analytics to capacity of a certain trade, performance metrics for our traders. Now it’s about how you create Alpha opportunities to outperform based on that data.” – Richie Prager, BlackRock
Roy E. Lowrance (New York University):
Data mining projects can take all day to think of an idea, then days or months to run. We need to figure out how to process it efficiently, with a consistent load. The financial data that is becoming newly available will need some smoothing and story extraction – that’s something that will be outsources. Lots of outsourcing will happen.
Abhishek Mehta (Tresata): Do you think our industry will embrace mining of data outside our own walls?
Lawrence: As new data sets become available, everyone will need access to them, cost will become an issue – outsourced providers can scale and you can’t.
Prager: Data providers are already here and big, third-party data is already key. It’s a growing space.
Weinberg: New interesting companies include data cleansing, data management – and how do we get a single view of the customer? New names you haven’t heard of will become household names very quickly just because of the serious need for their services.
Weinberg: Most of the companies in banking grew through silos and thus haven’t had a good single view of customer. Even if you had the data in one place – how do we manage then to execute? If we had visibility and transparency into the exact credit on every distinct loan, if you can see the whole customer then it’s no longer a mystery and we can avoid the mistaken assumptions of the past.
Abhishek Mehta (Tresata) is moderating this conversation very well and is a model I should follow for moderating panels like this myself.
Prager: For institutional investors, we manage over $3 trillion and we believe in analytics so much that we’ve created a dedicated research staff. We actively use tools to understand everything from trading cost analytics to capacity of a certain trade, performance metrics for our traders. Now it’s about how you create Alpha opportunities to outperform based on that data.
Lowrance: You have to crunch the MBA stuff, figure out how to use advanced number crunchers, make use of them. Capital One had the very best analytics and they believed them. They executed based on that. That’s the hard part.
Weinberg: It’s well developed to see what the best data is, run it through intelligent analysis, though there is some concern about getting lost in the model. On the retail side, though, people are stuck to see where is the value? You still need to sell a single thing to a single customer. We need the ability to test quickly; small offers, tested quickly, then ramp up. We need to know what the total size of the pie is – then we’ll see the money start pouring in. You get sophisticated, expensive tools and a lot of people sign up but don’t use it a lot. First people will make a lot of money, then it will become commoditized, but we’re at the beginning of that curve.
Lowrance: Key training opportunity is with middle managers. You have to get them trained and aware of the possibilities.