The Mobility Journey: Oil & Gas Should Think Of ROI As "Return On Innovation"

Tim Haïdar

In this exclusive interview for Oil & Gas IQ, Satyam Priyadarshy, Chief Data Scientist at Halliburton, speaks to us about the discipline of data science, and why the modern CIO must harness this for companies to win in the big data revolution.

SP Satyam Priyadarshy, Chief Data Scientist, Halliburton

TH Tim Haðdar, Editor in Chief, Oil & Gas IQ

Satyam, many thanks for joining us today. My first question for you is, what is your definition of a data scientist?

SP A data scientist is a person who’s trying to find patterns in data to solve business problems. A person who understands what kind of data exists within the company, what data can be connected, which datasets are not connected and how they can be leveraged to find new correlations.

So, in some sense, a data scientist is a person who brings together in-depth scientific
knowledge, -- technology experience and extensive business acumen to solve complex problems. You can have programmers write algorithms, you can find software which allows anybody to run a neural network on any dataset. But does that make any sense to the person if they don’t know the context they are operating in? If I give you a dataset and not tell you what business it is, and you will find some pattern and predict some variable but how does it relate to anything useful?

TH And not all data is good data, you need to have a regimented idea of what you actually need to use and what is valuable to you. Data just for data’s sake doesn’t make any sense.

Absolutely. If there is no value in the data then you can have as much as you want but it is still worth nothing. So value in the data has to be there. And that value can be realised in different ways.

For example, your cell phone is generating a lot of data as we are talking but it has no significant value to you really. However, it has significant value to the provider, the device manufacturer, the apps that are running on your phone. So even if you have data in your pocket, it might hold little or no value for you, but may be extremely valuable to somebody else. In that sense, all the data is important, but it has to be in the right area.

TH Now let’s talk about mobility. What are the challenges relating to mobile technology for the upstream industry?

SP Well, if we look at downstream, it is pretty much like any consumer play and midstream is more like an ERP play. But upstream is more scientific. A lot of predictive analytics needs to be done before you can go and actually start doing things. So, let’s take exploration and the data coming in from the seismic studies. There is really no need for it to be analysed in real time because when you study seismic data you get this 5km area or 20km area.You don’t have to immediately say, this is the place I really need to drill. It has to be built up and analysed properly and thoroughly.

Mobile technologies are helpful because you can speed up the process and look at certain things and say, let’s collect a little bit more data in this area or that area.

Let’s say you now went from exploration into the drilling phase - it certainly makes sense to have a mobile device or good standard mobile strategies, because you can do things like optimise your supply chain, optimise equipment movement and you can see all those things in real time coming in from various parts of the business. Because this is a highly complex operation, you usually have to wait for paper to move. With mobility you can actually do much better. On the traditional supply chain side of things you can augment process improvement, logistics, people management et cetera. There is a huge play for mobility there.

When it comes to the physical drilling, the predictions you make can turn into profits. You can analyse data as it comes in from the wellhead to give you some powerful insights as well. Mobile devices are key for that since the man on the drill floor is looking at the same info as the senior engineer in head office. Questions can be answered immediately, decisions can be made, and costs can be cut in situ.

So then it brings up the question about connectivity. If you’re in a remote offshore region - maybe you might be on an FPSO off the coast of Brazil, or a drilling rig west of Shetland - what kind of impact is the lack of connectivity going to have on mobile device usage?

Connectivity is improving day by day in this industry. Companies are now deploying fibre optic and compressing the data so that a large amount of data is not moved. I think M2M communication is going to be optimised soon and satellite connections are getting much faster. As communication technology improves, the traditional mobile connectivity to these carriers will be leveraged even more.

Now, there’s one question that’s got to be asked in a time where we’ve seen the lowest WTI crude oil price for six years and Brent crude is hovering around $53 a barrel. Do you think oil prices are going to stop mobility programmes in their tracks?

Coming from a background in multiple different industries, I think it is imperative that the oil and gas industry uses mobility to its maximum potential. This is a great time to leverage a mobile offering because it will absolutely help in reducing costs. That statement is true based on what we have seen it in other industries. The barrier to this is the way that most people think of ROI as meaning return on investment.

We should no longer be thinking of ROI in those terms but as a return on innovation. Maybe it is an investment this year on people, processes and culture but in the long run, you will make your returns on the innovation that you undertake. So rather than just looking at purely a short-term investment return, you have to be thinking about the bigger picture.

That’s an interesting way of putting it. Can you expand a little further on that definition of ROI, then?

As we know, the same methods will only lead you to the same results. Innovation is about finding something new. If you connect multiple datasets which you have never connected before, you’re going to find new patterns and those new patterns are innovations.

When you see new patterns emerge, you’re asking the question, why do I see these patterns in the data and what does it mean in the physical profit sense? Here’s an example. People have been studying a problem in the upstream domain called stuck pipe. There are roughly 400 papers published in this area and everybody has been looking at this problem with a very narrow, siloed version of the data. For example, you may have one well and you decide to use four hours’ worth of data to build a predictive model.

These 400 odd papers have been published using various machine learning algorithms, neural networks, fuzzy logic and whatnot but nobody has been able to solve this problem and it’s roughly a $2 billion a year problem.

You use the data but you have never actually looked at the problem holistically. If I looked at this problem from a holistic point of view and connected the data from multiple places, you approach this problem in a much different way and look at patterns in data that nobody else has seen before. Now I can say that this is not actually a mechanical problem nor is it actually a chemical problem, it is something else. That data was never analysed. So, if I input all the historical data and get new patterns then I ask different questions. When I ask different questions, I get a new innovative solution to it.

The return on that is going to be very significant. If this solution is implemented across the industry then everybody benefits from it. Initially, I’m investing time and energy to actually solve this problem, trying to formulate it. But if it solves the problem, over five years you could save roughly 9-10 billion dollars.

That is why I use the term return on innovation rather than return on investment. Return on investment is more a traditional CIO view. Let’s buy this hardware, this software and we are done. We have invested $5 million, now I need to really write it off in ten years so that I look good. I believe that way of thinking is becoming outdated.

Satyam, many thanks for your time today.

SP Thanks so much, Tim. Bye now.