First Look – Opera Solutions

I got a chance to catch up with Opera Solutions recently. A company that aims to improve “machine intelligence” and couple it with human insight to help companies with sustained profit growth, Opera has 400 staff worldwide with over 125 analytic scientists. This makes it a very large analytic solutions provider in a market dominated by small specialty firms.

I got a chance to catch up with Opera Solutions recently. A company that aims to improve “machine intelligence” and couple it with human insight to help companies with sustained profit growth, Opera has 400 staff worldwide with over 125 analytic scientists. This makes it a very large analytic solutions provider in a market dominated by small specialty firms. Like me, Opera sees a compelling case for applying “machine intelligence” – to develop smart (enough) systems I guess. And like other companies who have studied results, they see analytics improving results significantly. For instance, an assessment of technology companies found that those using these analytic techniques had better sales and EPS growth by a significant margin.

Opera sees a growing opportunity for analytics. With “big data” and with continually changing or flowing data, they feel the traditional approach to mining data has to change. They argue that it must be possible to extract more meaning or signal from an environment with more noise, it must be possible to bring human insight and analytic insight into the same environment and all of this must be delivered through a widespread environment that permeates all channels. They really focus on combining human intelligence with machine intelligence and take a lesson from chess where machines have beaten human experts in specific instances but the combination of an OK player with a machine creates an unbeatable combination.

Opera has built a global team of 120+ scientists, libraries of techniques, user interface technology, simulation/simulation tools, technology for finding signal in noise and various specific platforms. Companies work with Opera by buying Insight Bureau Services (recurring insights delivered as a service) or through various consulting, partnership or licensing deals. Opera has six areas of focus:

• Marketing services
• Credit and risk services
• Supply chain services
• Global markets analytics
• Data transformation services
• Custom applications.

In these different areas they often develop platforms to solve specific challenges- how to combine human intelligence with machine intelligence to solve a particular problem. For instance, Mobiuss is a platform for portfolio risk and valuation initially focused on Residential Mortgage Backed Securities. The platform takes a bond which may have 4,000 to 5,000 individual mortgages.  It then applies advanced likelihood/willingness to pay models determine the default and prepayment risk of  each individual borrower.  It allows a user to vary scenario variables like employment and interest rates, then model the cash flow for different scenarios.

This allows pricing and analysis so that buy / keep / sell / hedge decisions can be made. As I like to say a platform like this lets you use an aggregate of granular simulations rather than trying to simulate against an aggregate. The platform actually runs various decisions (will this person pay off, default etc) using very granular data then aggregates these to provide an overall risk.

Companies in this space are only making hundreds of decisions a day – not normally enough to justify a decision management platform like this – but these decisions are being made based on millions of elements so this volume is more than enough to require a degree of automation like this.

Another example was a used car auction pricing optimization solution. This company had 250,000 cars come back off lease each year and wanted to optimize the floor price for auctions. The system uses auctions of other similar cars from other vendors and previous sales to predict the best floor prices for the cars coming up tomorrow (this is location specific). It also makes recommendations as to where to send cars to be auctioned to maximize the value of the auction.

I liked the way this solution exposed the analytic models to the user and made it clear what was making a difference so that the person can change elements that drive the price and then get a new floor price. Rather than allowing a simple override, the people using the system change the quality of the inputs using their visual and business skills and this better data drives a new (analytic) decision. A nice example of how to bring human decision makers into the loop without losing the analytic edge.

Opera is a company with big ambitions and some interesting solutions. It will be interesting to see where it goes from here.