Making Decisions Based on Predictive, Preventive Analytics

Data, applied effectively, can be useful for everything from bolt maintenance on a ship to changing personal behavior for better health. In this episode of Mastering Innovation on SiriusXM Channel 132, Business Radio Powered by The Wharton School, Steve Orrin, federal CTO of Intel Corporation, shares how he strives to ask the right questions and find solutions through analytics.

Orrin discusses his background as an innovator and an entrepreneur, both at small and large corporations. He explains what he believes large companies can do to better incorporate a fail fast mentality that leads to successful innovation.

An excerpt of the interview is transcribed below. Listen to more episodes here.


Steve Orrin (Federal CTO, Intel Corporation)

Harbir Singh: I read one of your articles, “5 steps to accelerating insights with predictive analytics.” What I found fascinating was that three out of the five points had to do with people and culture.

Steve Orrin It’s one of those things people have said for many years: Solving big problems is about people, process, and technology. The key is that you can’t forget any one. A lot of times, people focus on technology, especially in the tech industry, but they need to understand that ultimately, it has to be deployed into a real-world environment by people, and it has to enable those people to be successful in order to achieve its goals. Then, you have another side of it where people are saying, “Well, I’ve got to change my process to do analytics,” but they forget that they also need to enable that technology to be adoptable. It’s really about all three working in concert and not giving one more credence than the other.

Singh: Let’s talk a little bit about some use cases because I find this fascinating, and maybe it would help if we can choose something we can all relate to — a case where we did predictive analytics and the agency became or can become much more effective.

“Solving big problems is about people, process, and technology. The key is that you can’t forget any one.” – Steve Orrin

Orrin: There are a couple of areas looking at advanced analytics and predictive analytics. The classic places that many folks are looking at are around predictive maintenance and prescriptive maintenance, and being able to not only detect or diagnose a problem that you have, but also be able to plan for when you potentially could have a failure.

Singh: What application would it be? Would it be a health care services application? How would the consumer know? I understand the producer side, so I would love to chat. If you can add some color there, that would be very helpful.

Orrin: I can give you a good example of the use case where predictive maintenance really comes into play. The same kind of use case can be applied to health care and to other areas. If you think about an oil rig or a ship at sea, bolt corrosion is a problem. Over time, bolts that hold things together start to corrode. The first step of the process is that you have to be able to detect that. Today, before analytics, what you’d have is a diver or a drone going down to take pictures or video feeds, and someone would sit there looking at hours and hours of footage. The first step in the process was to use computer vision to identify or diagnose that bolt number 27 was corroded and schedule a dive team to go fix it. That tells you you’ve already got a problem though. What you really want to do is start looking at, “Can I predict when the bolts will fail?” to give me, as an example, that six months or a year visibility of when I should schedule to have the dive team go out as opposed to waiting till the fire has already hit. So, predictive maintenance is about understanding the conditions or the root causes that cause that to fail.

“Predictive maintenance is about understanding the conditions or the root causes that cause that to fail.” – Steve Orrin

Singh: I see. There’s a physical failure, namely, the rusting of a bolt, but there’s also an analytic confidence interval that, we better get to it before there’s more than a 25 percent — pick a number, percent — chance.

Orrin: Exactly. One of the ways you do that in the case of the bolt is by looking at alternative data sources. You started with looking at video feeds of the bolts. If you start looking at the weather, the temperature of the water, and the salinity of the water, then start including those into your analytics, you can see that in places that have higher salinity, the bolts corrode faster. Therefore, you can start to predict, based on the salinity of the water, when a bolt may fail. Then, the next stage is what is sometimes referred to as prescriptive, which is that last stage of, “I don’t want to know when a bolt is going to fail. I want to know how to prevent bolt failure.”

Singh: So, the preventive part?

Orrin: Exactly. Looking at, say, “You know what? You’re buying the wrong bolts for that part of the world because they’re not good inside those salinity environments.” Now, you change your acquisition process to be able to prevent the problem, as opposed to getting ready to fix the problem that’s going to happen. That’s a good example of how the maturity model of analytics goes for a case in health care: from diagnosing you have a problem, to being able to predict that in 10 years, you’re going to have this problem, to being able to tell you, “If you start eating this now, you won’t have this problem.”

About Our Guest

Steve Orrin is the federal CTO for Intel Corporation. He has held architectural and leadership positions at Intel, driving strategy and projects on identity, anti-malware, HTML5 security, cloud, and virtualization security since joining the company in 2005. Previously, Orrin held technology positions, as the CSO for Sarvega, CTO of Sanctum, CTO and co-founder of LockStar, and CTO at SynData Technologies. He is a recognized expert and frequent lecturer on enterprise security and was named one of InfoWorld’s Top 25 CTOs of 2004 and, in 2016, received Executive Mosaic’s Top CTO Executives Award. Orrin created the Trusted Compute Pools Secure Cloud Architecture and is the co-author of NIST’s IR-7904 “Trusted Geo-Location in the Cloud.” He is a fellow at the Center for Advanced Defense Studies and a guest researcher at NIST’s NCCoE. He is a member of INSA, ISACA, OASIS, IACR, and is a co-founder and officer of WASC.

Mastering Innovation is live on Thursdays at 4:00 p.m. ET. Listen to more episodes here.