Thought Leaders in Big Data: Interview with Lars Olrik, Group CEO of Verdande, and Jo Kinsella, CEO of Financial Services of Verdande (Part 1)

Lars Olrik is the group chief executive officer of Verdande, and Joanne Kinsella is the company’s chief executive officer of financial services. The company specializes in early problem detection and prevention, driving its insights from data applying big data technologies. In this interview, Joanne and Lars talk about the oil & gas, healthcare, and financial services sectors and how Verdande helps companies in these industries prevent all kinds of problems from happening by comparing current and historical data. Furthermore, they see the potential to work with entrepreneurs who are building on top of their platform.

Sramana Mitra: Let’s start by setting the context of your company. Tell us about the company, what customers you cater to, and what problems you solve.

Lars Olrik: As group CEO, I will deal with the company’s profile and what we do. And on the use cases, Jo will take care of the financial services side of the business. Basically, Verdande Technology is a Norwegian technology company that started in 2004. It is fundamentally focused on case-based reasoning, which is a clever algorithm to predict what has happened before based on historical experience to avoid the same problem occurring.

We have three businesses: one in oil & gas, which is dealing with how to avoid typical drilling problems and if they happen, having enough time to take corrective action. These are typical reoccurring issues, so it is about how to learn. We also have a healthcare business that is looking at a similar principal, but it is about how to avoid or predict the types of patients you are dealing with. Then you have the business that Jo is running, which is focusing on operational risk and risk management – again it is about avoiding problems before they happen. We are not a five-minute warning system. We are about taking a longer term view on situations that give management the opportunity to correct them way before they happen. We do that by sitting on top of middleware providers – big data – and looking at how to take big data and convert it into something useful that you can then take to do something clever with your business. That is fundamentally what we do.

SM: It sounds like you have domain-specific heuristics in each of those verticals – oil & gas, healthcare, and financial services.

LO: That is correct. The fundamental basis of each of those is the same platform that we use across the three businesses.

SM: I would like to discuss three use cases from those three verticals.

LO: Let’s take the example of drilling. The starting price for a driller well in the Gulf of Mexico, Brazil, or Norway is around $100 million. You can imagine that drilling is a very costly process, and at the same time there are lots of issues around health and safety. Traditionally, each rig has a whole raft of sensors. Data [from these sensors] is then streamed onto land, so that people on land can see and monitor what is going on out there at sea. They have huge amounts of data, lots of information to disseminate and lots of data to react to when something happens. At any given time these sensors go off – there are green, yellow and red. Those are alarms that have a short-term horizon – five to ten minutes. You need to decide which one to act on. The likelihood of your reacting to the incorrect one is very high.

We are using the same data but looking at the context between the real-time data and the historical data. So we say: “There are similarities now in real time to what has happened in the past.” Then we have a radar and the case will come up. This case is a learning case. That means that the operative can click on the case, go in and look at the current situation, compare it to what has happened in the past, and then take corrective action to avoid that problem.

Typically, an individual engineer has between three and 48 hours to activate the mitigation of a particular problem. We have a whole raft of agents that address different types of problems. Each of these problems can cost you from $5 million a day to $40 million over four or five days, and you have to stop your operation. Sometimes you even have to fly the crew off because it has a very significant health and safety impact. Now you have people who take action before [you get to that point], and you can mitigate or avoid problems.

This segment is part 1 in the series : Thought Leaders in Big Data: Interview with Lars Olrik, Group CEO of Verdande, and Jo Kinsella, CEO of Financial Services of Verdande
1 2 3 4 5 6