Interview with Mike Scharf, CoFounder of Capspire

Two guys recording a podcast interview

Interview with Mike Scharf, CoFounder of Capspire

I had the pleasure recently to interview Mike Scharf, CoFounder of Capspire, about the role of digital innovation in transforming the trading function. Here’s a transcript of the conversation.

Geoffrey
My name is Geoffrey Cann and you’re listening to ‘Digital Oil and Gas’ and another episode about the world of digital innovation in oil and gas. Today I’m joined by Mike Scharf, C0-Founder of Capspire. Mike, welcome to the show.

Mike
Thank you for having me!

Geoffrey
Not at all, this’ll be great fun. So tell me a little bit about your personal background. Where did you start out? What’s your story?

Mike
I studied at Oklahoma State University—Industrial Engineering and Management with a significant focus on operations research—and then spent a couple decades helping energy companies across the world implement commodity trading risk management solutions. Throughout that time, you were able to really get a feel for the challenges and the diversity of the supply chain.

Geoffrey
I remember this (commodity trading risk management ) really blew up in North America around 1999 or 2000. I recall California as one example going through a massive upheaval in its power sector. Is that part of your background?

Mike
That is. I was telling some folks that about that time I was on the Williams trading floors. And there were some things that were blowing up at that time.

Geoffrey
Yes, that was the Enron era for some of us—those of us who are old enough to remember what Enron actually is—a milestone signature event that blew up, created and drove the Sarbanes Oxley legislation in the US, caused the separation of the consulting industry from its audit and tax firms, recast this entire energy trading and risk management area into a major thing for almost 20 years. And so today you’re in Capspire. Tell us about Capspire, was does Capspire do?

Mike
Capspire was founded about 10 years ago, and we do two things. One is we are the world’s best commodity trading risk management solution firm. We help organizations select, implement, and support commodity trading risk management solutions throughout the world. We have offices in North America, Europe, and we just opened our first office in Australia. Second, we help energy companies leverage digital. We have a number of digital offerings that help energy organizations advance their digital strategies.

Geoffrey
Is that digital work associated with commodity trading and risk or in other areas?

Mike
Most of it is. We’ll help energy companies establish digital roadmaps, or how, within their organizations, they can leverage a digital strategy, how can they get value from a digital strategy. Most of that work is is an extension of the work we do within company trading and risk management. We’ve identified really a number of gaps in the marketplace and we’ve developed products that that address those gaps. So we’re working hard to kind of push those platforms and agendas forward.

Geoffrey
Yes, I recall the commodity trading area, a few years back, was very much in the news, as rogue traders were blowing up banks and financial institutions because of a lack of controls and exposure over trade positions. I clearly see the market there. But why why would digital be a topic in that trading world, because it would seem to me that the industry understands its risk profile, related to trading and risk management in the commodity trading area. Why is digital suddenly a topic there?

Mike
Our strategy and digital agenda is helping organizations that have complex supply chains, such as an integrated refining organization, or a large NGL midstream organization, or even a commodity gathering organization—organizations that have enough flexibility and complexity to their supply chains or value chains—with technology to drive to optimal decisions. Our philosophy is that we can leverage modeling tools/technologies to model complex supply chains, and by doing that we can help energy organizations drive to what we would call general interest decision making. So how can we help organizations break down department silos, geographic silos, to drive towards the best decisions for an organization.

Geoffrey
To be a devil’s advocate, I can hear my trading friends and contacts in the trading industry saying “You know, we’ve been doing this for years. What’s changed?” Is it the fact that the volume of data is now outstripping the Excel models that people have used in the past or is it the sheer complexity of the infrastructure just creates more does optionality now? What’s driving or is it all of the above?

Mike
I would say the opportunities have always been there. if you look at how most mid-sized and large companies are organized, you’ll have traders that are focused on a trading book. You’ll have marketers that are focused on local metrics, marketing metrics, and then you’ll have transportation functions, inventory functions (with their separate focus areas). It’s really really difficult in today’s world for those individuals to make consistent decisions in the best interest of the organization. A trader can make a decision in the best interest of his (trading) book. A marketer can make the best decision in the best interest of his customers, his terminals. Barge schedulers are going to try to keep barges busy.

Geoffrey
But the challenge is that you can’t see the opportunity to optimize at that more macro level, given the structure that you’re actually in because all the all the participants in the structure today—traders and barge operators will tell you they’re doing the very best they can. How do the market leaders, the ones who get it, see that the others are unable to visualize that motivates them to do this differently?

Mike
The world’s leading energy companies have seen the margins that they’re leaving on the table by not making general interest integrated decisions. And most of those organizations are either in the process of creating, or have created, a value chain organization, that sits between, for example, refining and marketing that’s really responsible for driving and coordinating the best decisions for the organization. Leading organizations see this benefit of having buyers and traders making decisions that are consistent with refined product opportunities. Or refined products traders making decisions that are consistent with what marketers want to sell at the terminal.

Geoffrey
It’s easy to see a barge operator motivated and measured to keep the barge highly utilized, and not be as concerned about the margin for a particular barge move, and a trader who’s motivated to maximize the margin on a trade and take advantage of a market opportunity. And if those two parties don’t talk to one another, the barge operator is going to make a sub optimal choice, by keeping the barge busy but missing out on the best margin opportunities because the trader can’t see what the what the barge operator’s up to. And similarly the barge manager can’t visualize the trades coming at them. Am I summarizing that in a fair way?

Mike
Exactly.

Geoffrey
Imagine the complexity when you get into say, five refineries, two continents, multiple pipeline systems. The whole integrated value chain creates a tremendously different operating landscape.
So tell me a little bit about how the solution works. I’m now visualizing the supply function as a complex organization, a lot of moving parts. What does the technology actually do that’s different from, say, your basic trading and risk management front, mid, and back office product?

Mike
We’ve spent the last five years building a solution that, and I’ll use one of your words, can create a digital twin.

Geoffrey
[Laughs] No fair using my words against me, Mike!

Mike
We create a digital twin of your physical supply chain. So, yes, let’s take for example a North American refined products market and a refiner that’s trying to optimize, let’s say, their downstream refined products market. We would be modeling all the output of their refineries. We would be modeling all of the different batch open stock pipeline systems. We would be modeling all of their storage terminals, we would be modeling any ethanol routes, truck routes, barge routes, to really being able to mathematically create a digital twin of that network. It’s not easy, but once you’re able to do that, it really helps an organization drive to better decisions, whether it’s a strategic decision of increasing batch sizes on the Colonial pipeline, or if it’s entry into a new market, or if it’s evaluating the number of barges you need in your barge fleet. It really becomes an amazing tool.

Geoffrey
As we add more complexity of real movements, complex pipelines, tank farms, barges, and so forth, all of that modeling is a real challenge. The tool of choice today is probably Excel, I think. And so different tools, yielding a better and different outcome.
Through Capspire you model a lot of data. Are you using or applying novel digital technologies to process that data? And here I’m thinking about machine learning or artificial intelligence or some other kind of sophisticated modern era analytic capability?

Mike
It’s a great question. I would say we’re leveraging artificial intelligence, in somewhat of an old school fashion using new age technology. We leverage cloud data technologies for their ability to process and store large amounts of data. We’re leveraging native cloud computing technology like Kubernetes, so that we can scale. But then we’re using what most data scientists would probably say are old school approaches. We’re using mixed integer linear programming. Our base model is a network flow model. And then we leverage Monte Carlo simulation, to be able to apply stochastics and variability to it. We haven’t found, in the specific problems that we’re solving, tremendous application for machine learning.

Geoffrey
It’s a very data intense problem. And how dynamic is the model? A dynamic model would allow me to take in today’s orders or today’s production in a field as the orders come flowing in? Does it allow me to dynamically manage and make decisions? Versus a static model, in which you provide a bunch of inputs, get a bunch of outputs, and you try and make make decisions best as you can based on this fixed data. How do you that playing out at the moment? Is it more dynamic or more fixed?

Mike
The model and the technology were built to be dynamic. We have a range of clients, based on their maturity level, that will use it at different time intervals. Some organizations may use it only for annual planning. Others may use it to do a monthly run, where others will be driving daily decisions. They will be integrating rail car tracking systems, so they know how many rail cars are in transit. Where are they located? What’s the best view of their ETA, so that we can be making decisions that are consistent with latest information on our supply chain? So, and that really depends on the maturity and the need of the organization.

Geoffrey
Could I use this model and data for transaction planning? In other words, I could, say, based on my interpretation of the data where I’m seeing either bottlenecks or shortages, invest in new assets in that location (expand a tank farm for instance), or purchase someone’s existing operation. Is that what some companies are doing?

Mike
Yes, spot on. One of the most significant use cases is attacking bottlenecks. Whether it’s a bottleneck on a pipeline storage location, a barge dock. You want to analyze where am I constrained. Where am I hitting a constraint? And how do I attack that? What’s the value? If I was to double my storage capacity at this location, or if I was to increase my pipeline capacity, or double the number of barges that I had, what is the cost of that? What’s the benefit?

Geoffrey
Mike, can you share some insight as to what customers get from using the technology? Do they see a capacity/throughput improvement? Do they realize better margins? What’s the benefit that people tell you about?

Mike
It’s game changing, because what we’re talking about is totally changing the way that decisions are made or assisting in the way that decisions are made. For those organizations that are successful, it’s game changing, and they are amazed by how it’s making decisions, versus how they made decisions before. What we find is most people get caught in historical bias, or historical rules of thumb, using a model that maybe I’m just running once a month. The digital twin constantly challenges that historical bias, to make people think differently. So it makes people think that there there may be another way of getting product to that market.

Geoffrey
One of my clients likes to remind me that many people in oil and gas benefit from having 30 years of experience, but the challenges is that it’s the same year repeated 30 times. And if certain rules of life get embedded deeply in that first year, because you were brought up through the organization to treat a tank farm’s behavior in a certain fashion, you’re going to apply the same rule for the rest of your professional career, unless you’ve got some data to tell you that there’s other way to do it.

Mike
Exactly. The market is constantly changing. One thing that I read recently was how chess players were re-learning how to play chess because of some of these AI chess algorithms that are playing the game very differently. So that’s teaching professional chess players to play the game differently. And we think of this technology in a similar way. It’s not always going to be driving the end decision for every organization, but it is going to be teaching those organizations different ways of making decisions.

Geoffrey
I wrote about digital twin technology in my book. You take the digital twin and build around it a virtual version of an operating business, which you could turn it into a game platform for your own employees. And you use it as a teaching tool—here’s our network, here’s the network of our assets, here’s how they play out. Let’s teach you how this business works and how we make decisions, because we can run the digital versions faster than you can a real business. It takes time to move 50,000 barrels in and out of storage. You can do that in a blink of an eye in the digital platform. You could rapidly accelerate decision learning and behavior change if you had the digital twin. Do you have clients who think like this now?

Mike
I don’t know that they think about the digital twin as a way of educating employees, it’s a great idea.

Geoffrey
Well, it’s like a flight simulator for network asset business versus a flight simulator for an aircraft.

Mike
They will use it to run hundreds of different scenarios that they can imagine for their business.

Geoffrey
And then make choices. Where do you see this going in the long run, Mike? If you have a digital version of the system and start to understand how it behaves, how much automation could you bring into the system? Is there still a layer for humans to do the creative problem solving?

Mike
I don’t think human intervention is going away. I do believe that there is a tremendous opportunity for automation. It really depends on the problem and the supply chain and the data that’s available. Take for example one problem that we’ve leveraged the technology to solve. Imagine I’m a large retail organization, I have 500 convenience stores, and I have a complex supply chain. I’ve got 200 trucks, and I have a number of different ways that I can supply products. I can supply products from my inventory, I can buy the daily rack posting, I can buy off an index agreement that may have a monthly, weekly, daily commitment. I can buy a prompt gallon. And I’ve got some variability on demand at the stores and I’ve got tank limits.

Today that type of problem is addressed by five supply specialists, 10 dispatchers, and a coordinator. Those are the types of problems where in the next couple of years we will see tremendous automation, because it doesn’t need a lot of human intervention. But this next half decade will really see this technology sitting side by side with schedulers and traders to help them drive better decisions.

Geoffrey
That’s the world of the digital twin and where it takes takes us. Mike, it has been fascinating to hear about these ideas and where and how they apply in the world of commodity trading and risk management. Thank you so much for joining me today.

Mike
You’re welcome. Thank you for having me.

Geoffrey
Mike, if people want to learn more about you and what you’re up to, where do they where do they find you?

Mike
Just go to www.capspire.com

Geoffrey
Fantastic. Thanks very much and join me for the next podcast. Bye for now.

 

*****

Check out my book, ‘Bits, Bytes, and Barrels: The Digital Transformation of Oil and Gas’, available on Amazon, iTunes, Audible, and other on-line bookshops.

Mobile: ☎️ +1(587)830-6900
email: 📧 geoff@geoffreycann.com
website: 🖥 geoffreycann.com
LinkedIn: 🔵 www.linkedin.com/in/digitalstrategyoilgas

No Comments

Post A Comment