By Geoffrey Cann
How can upstream oil and gas overcome its siloed processes and unconnected data to free up time for engineers, accelerate decision making, and improve decision quality, but without replacing everything? Through a platform approach to knowledge automation.
Geoffrey: I am pleased to be in conversation today with Kenton Gray, who is the Chief Technology Officer for Datagration.
We’re going to discuss a vexing problem that the industry oil and gas industry needs to confront. Throughout the value chain, from upstream to downstream, companies have to make various choices for different technologies to solve industrial problems, and then work with the information from those technologies to support business decisions. The resulting architecture is often convoluted and hard to change. Datagration is fixing this problem.
Kenton, let’s begin with your background, and how did you landed in Datagration?
Kenton: I am a Texan born and raised. As a child, I was always very excited by tech. I remember, when I was a kid, my mom would take me to the library, and I would go and check out every computer book I could find. I was obsessed with computers. I really liked games, and I found out that there’re books full of pages of code that you can go and type out, and then you had a little game at the end of that.
Geoffrey: A kind of self-taught game programmer. Where did this lead you?
Kenton: I got my first paid gig when I was about 13 and did a bunch of web development. There were some interesting brands like the comedian Pauly Shore, Dr. Dre and Dr. Scholl’s. The two doctors, I called them. At the time, I really didn’t think tech was going to be my career path. I thought I was going to be a writer. I went to school for English. Then I realized, Oh, well I’m doing this web building thing and I’m getting pretty successful. Maybe I should just stick with tech and not really try to do the starving writer thing.
Geoffrey: Coincidentally, my first gig doing software coding was at university, working on the IBM PC. Tell me about your current role, as CTO at Datagration.
Kenton: It’s a broad role. From making sure that our computers are secure, that our internal business processes are proper from technology perspective. Avoiding viruses and ransomware for example. More important is ensuring that the product and the technology is secure and stable, performing well and constantly evolving. We’re adding new features and functionality all the time. I bring together our really talented development team and engineering team, and help them understand the challenges each other are facing.
Geoffrey: You have both an internal role, provisioning the computer capability and support necessary for Datagration to run as a business, and an external role overseeing the product set that Datagration takes to market. Usually that’s a two headed role, because the internal role tends to be focused on cost and basic service provision, while the external role is much more of a creative, customer-facing role. To have both roles is really novel.
Kenton: In some ways, yes. But it is more common in Software-as-a-Service applications, because the security of your company reflects the security of your product. If you don’t have your internal controls in place, then your product can’t be trusted. For example, in the pandemic many companies have been struggling because they had all their secure servers on-site, not available from home, and then all of a sudden, they need to support remote workers.
Geoffrey: This is likely a permanent change. Now let’s turn to how you’re helping oil and gas companies. You’ve identified a problem that is endemic to the industry–many unconnected and siloed data sources. Can you help me unpack this issue?
Kenton: It’s a really common problem that I think people see again and again, which is that engineers are tasked to prove and predict what events are likely to happen in the future. But they’re spending so much of their time gathering data, parsing Excel spreadsheets, trying to learn how to program in Python to simplify data management, and they’re having to do quality control on data. All those kind of things take them away from their role. They spend so much time on data management that they can’t actually solve the problem, because they’re trying to get access to the provisioning of data. What we’re trying to do is remove those roadblocks and let engineers engineer. We do have specific oil and gas areas that we focus on, different solutions that we try to solve for–work over candidate selection, portfolio optimization, unconventional economics, and so on. In the end, it’s all about empowering those engineers.
Geoffrey: Woodside Energy once estimated that 40% of their engineer’s time was being spent on searching for the right information to use in analysis. What data types do they need? What is the source of the information?
Kenton: It’s a bit of everything. Excel spreadsheets, for one. SQL databases are another. A lot of companies will have different SQL databases in each field, different structures of data at different locations. You also have historians, or PI systems or SCADA data, those are really common. Furthermore, you don’t do engineering for fun. There’s a business outcome you want to get to. It doesn’t do you any good if it takes you six months of engineering to figure out what you want to do in Q1. So, being able to bring all that data from all those different systems, and be able to query it together and see the relationships between them in one place is powerful.
Geoffrey: What has to happen for that to work?
Kenton: The traditional way to come at this problem is yet another SQL database, but this method fixes the way data is accessed, whereas the PetroVisor platform is an unstructured database. Everything that we’re bringing in can appear, or disappear and that’s totally fine. We understand that you don’t always have every piece of data. What we do is we link the data we have, and use machine learning to fill in gaps. We also have specific types of data that are very relevant to an engineer at an oil and gas company, like pressure, volume, temperature, depth measurements, porosity, permeability, and we understand how those work.
Geoffrey: How are new Artificial Intelligence tools making this easier to do?
Kenton: A big part of our application is what we call data mapping or data integration. Step one is to determine where the data is coming from? Is it a SQL databases, or an Excel spreadsheet? Step two is you map it. For example, oil production is measured in barrels if you’re in America, or tons in Europe. And then you can reuse those mappings over and over again. If you have the same systems, you can use that same mapping, or you can create a different one if the data is structured slightly differently. Then, we look at the entity name and we use “fuzzy matching” to improve the data. For example a field might say Well 01, Well 1, or Well_01. All those things are obviously the same thing, which a human can determine, even if the data is different. We have a lot of tools that automatically detects those differences in taxonomy which we clean it up.. We can apply filters to help shave off anomalies in the data, such as when a measurement is dramatically incorrect , or you can define your own rules or your own gap billing.
Geoffrey: Are those rules and filters shareable across different engineers and companies, so that you don’t have to reinvent over and over again?
Kenton: Everything in the PetroVisor platform is designed to be collaborative. Everything is permission driven. We don’t just show everyone everything, you need the right permissions and roles. But once you have those, then you can see exactly how everything’s configured, you can reuse what other people are doing. If someone leaves the company, or retires, that knowledge didn’t just disappear, it’s still embedded in the platform.
Geoffrey: In essence, you’re capturing industrial knowledge and know-how in a platform that survives when people leave. Is there a way to take common processes and make them standardized, or copyable, or repeatable?
Kenton: What we call it knowledge automation. There are many people in companies who are doing the same thing every day. For example, one of our customers, a European oil company, were conducting a regular work-over candidate selection. Once a week every quarter, their team of engineers and business people would meet in a conference room and determine their work-over plan. What are work-over candidates? What problems do we see here? What are we going to spend our next dollar on? We worked together to codify the analysis using p-sharp, which is our scripting language, and PetroVisor. We were able to move all their calculations to where it was running on a nightly basis. They’d wake up in the morning and get a Teams notification saying: “we’ve identified a new work-over candidate” or, “we detected a skin buildup problem on this well”. All that knowledge that they had was automated and put into the platform, while it didn’t really change their way of working.
Geoffrey: Through this knowledge automation, could you also bring that kind of engineering excellence down to wells which would not normally qualify for in-depth attention?
Kenton: Absolutely. In their case, the focus was with brownfield assets. They had many shut-in wells, or inactive abandoned wells. When they started running with our platform, they discovered a lot of potential in these older wells. Oil companies naturally focus on their high performing wells, but don’t have capacity to study those that are producing at a lower rate or inefficiently. We were able to identify that, and create a massive opportunity for shut in wells.
Geoffrey: Could this be a game-changer for many oil companies?
Kenton: Certainly. The environmental aspect of the industry is also really under pressure. Everyone’s looking at the environment, social and governance (ESG), and what that means for oil and gas. Obviously, exploiting existing wells and infrastructure is a great way to be more environmentally friendly, but everyone’s setting very aggressive goals. Everyone’s saying that we need carbon neutrality by 2030. But it’s very difficult to do that without knowing what your actual output is. How are you measuring that, and how are you weighing it? One of the things that we’re excited about is being part of that solution, delivering those metrics, and then identifying ways that you can do more with what you have. We can help companies avoid creating excess waste or using excess water.
Geoffrey: Companies will have made investments in solutions that they intend to keep, such as Spotfire. What is your approach to helping companies preserve some of those key investments?
Kenton: One of the things that we believe is the separation of the business logic and the engineering work versus the visualization. A lot of companies love Spotfire. But they start turning it into a programming language and doing all their calculations through it, and then eventually it becomes slow, maybe even crashes. In PetroVisor, we’re doing all those calculations for you. What you do with those, and how you visualize them, is up to the customer. We have connectors into Power BI and Spotfire and Tableau. We also have some built in visualizations inside of the PetroVisor platform. For those who rely on Spotfire, PetroVisor can just be a title on Spotfire. You can preserve everything you’ve built over the past ten year.
Geoffrey: How does the system handle our new way of remote working?
Kenton: One of the more interesting outputs, when everyone’s remote, is our integration with Microsoft Teams. We’re able to notify via Teams when problems are detected, a change occurs, or if there’s an issue with a workflow. All those things are configurable to be a Teams notification, all the dashboards can be embedded into your Teams channel. When people wake up, they just go and look at their Teams. There’s no need to visit yet another system.
Geoffrey: What is the market’s reaction to this concept of a data platform?
Kenton: It always takes people a bit of time to understand what we’re doing because it’s too easy to conclude that we’re a data lake, or a database provider, or point solution. I understand where that comes from, as we share features of those solutions, but we’re different. We help solve these problems in a very open way. The PetroVisor platform is transparent. We don’t take your data and give you the answer. Everything that we do is completely reviewable, open to the end user so they can go and see our analysis and work. How did you decide there was a water funding issue on that well? The customer can take a look at our analysis, and using their 20 years of experience in managing water, they can replace our calculations with their own. They still capture all the benefits of the entire solution. We give them a head start, and they can take our workflow, be 90% done with a job, and reduce a year long project to a couple months.
Geoffrey: What about the use of tools like machine learning?
Kenton: At first, everyone kind of was very excited about machine learning or AI. But soon, they realized that they have to trust the machine learning algorithm to get the answer correctly. Even on that, we have a different approach. Through the interface of the application, you can configure that machine learning algorithm, training it on your data. You can see how it’s calculating, and opt to use the physics-based way of doing this, rather than using machine learning. On the other hand, maybe you can find use for machine learning and leave it in as well.
Geoffrey: So by combining the human experience in the industry with the technology power of these platforms you unlock value?
Kenton: The fact is that every one of these companies has a ton of really smart people working for them, and I have a ton of smart engineers on my team. What these companies have is decades of knowledge, things they’ve seen that we couldn’t dream of. If we can combine our technological head start with their expertise, it’s a win-win.
Check out my book, ‘Bits, Bytes, and Barrels: The Digital Transformation of Oil and Gas’, available on Amazon and other on-line bookshops.
Take Digital Oil and Gas, the one-day on-line digital oil and gas awareness course.