Jon Sobel has not been entrenched in manufacturing all of his career. He has, in fact, managed to string together what he calls a “fortunate series of opportunities to get involved in new technologies just as they were starting” — similar to Forrest Gump, but with far more skill, far less dumb luck and far, far less distance running.
Sobel worked for Yahoo! and CBS in the late 1990s and 2000s, the former as consumer Internet transitioned from a niche product to an everyday necessity, the latter as video started to dominate the medium. Later in the 2000s, he worked for SourceForge during the rise of open source and Tesla Motors as electric vehicles started to become more and more of a reality.
His current venture is Sight Machine, which launched in early 2012 and blends manufacturing with Big Data and the Internet of Things, bringing analytics to the factory floor. The company has offices in the San Francisco and Detroit areas — a physical mix of traditional manufacturing and technology — and aims to solve what it targets as the most critical problems for manufacturers.
IndustryWeek: One of the stated goals of Sight Machine is to “solve manufacturers’ most critical problems.” In the last four years, what are some of those critical problems? Individual companies know what their problems are, but maybe not what the industry’s problems are as a whole.
Jon Sobel: On an industry level, it’s a highly-coordinated set of activities, not just within a plant but across many factories and many companies. One of the problems that is increasingly important to resolve is the ability to have transparency and connectedness between far-flung operations. Pressures on manufacturers to deal with regulations, much faster design cycles and global competition — all the things we talk about — require manufacturers to have a level of transparency that most people outside of the industry assume they already have.
When we talk with people about what we’re doing, the most common response from people outside the industry is, “I thought they’ve had this for years.” How can you run factories without really knowing what’s going on in your own factory or supply chain?” There’s tribal knowledge and an implicit understanding of what it takes to make factories work, but that’s not going to cut it in the future. One of the most important challenges is turning real-time understanding and transparency into highly-connected distributed operations.
IW: The whole concept of Big Data is sort of like hoarding: Whether it’s newspapers, or soup cans, you can’t find what you’re looking for, you can’t walk through your home. Companies are streamlining their hoarding then, for lack of a better analogy?
JS: Exactly. Manufacturers are very articulate about this. We spoke with a leader in Detroit last summer, and he said he had twice the order of magnitude of data that he had five years ago, and he needed three times the order of magnitude to understand it. If you’re awash in data, unless you can make it useful, it’s a burden and it does you no good.
We describe factory technology as automation and control. All this was built to enable machinery and processes to run predictably and to give manufacturers control over scheduling those assets. This wasn’t designed to be an IT network but it produces valuable data. Now we have the tools and some companies are realizing the best sources of insight come from the production department. It’s an untapped area of opportunity, and now they’re applying on the plant floor.
IW: I’m sure it’s different from company to company — from floor to floor, even — but is there a right approach to this? Is there a wrong approach?
JS: A wrong approach is to start with the technology and say, ‘Let’s try a bunch of Big Data technology, let’s build a data set and see what we can find out.’ You don’t have specific business needs, you don’t know what problems you need to solve. Equally important, those projects almost always fail because they don’t have a person inside the organization who owns the outcome and has the influence to make it successful. … We’ve seen some companies that just want to try the technology and don’t care about solving a problem. That almost always fails because nobody cares. It doesn’t do anything.
A right approach is to start with the business problem. What we always say to folks is, ‘Let’s prove that this technology can solve a problem you haven’t been able to solve before — otherwise, there’s no reason to use it — and let’s find something that’s really bothering you and see if we can’t fix it.’ If you can identify somebody in the company who cares about this problem and really wants to solve it, you’ll have the momentum you need to break through various organizational challenges and get it done. And once the problem is solved, people will be able to see something concrete — results, improvements, whatever the problem is — and that’s when you really start to engage.
IW: Sounds so simple: Having something to have it and twiddling your thumbs leads to nothing.
JS: Exactly, and a lot of people do that. The Big Data industry has put out the theme — and it is true — is that you don’t know what you don’t know. You should never start a project wanting to figure out what you don’t know. Let’s start with something practical and concrete.
IW: We’ve all seen the exponential growth charts showing where IoT and IIoT are projected to be in the next five or 10 years. Where can it actually go, and how can it evolve on the floor for manufacturers?
JS: We believe industry is at the beginning of a very long transformative cycle. The worldwide advertising market, which is where the Googles and Facebooks have had so much success, is about a $500 billion global market per year. The worldwide manufacturing sector is $11 trillion. Health care is $7 trillion. Education is $4 trillion. You’re talking about industries in realms that are orders of magnitude larger than the industries where the Internet has already been supplied. It seems that there is abundant opportunity for great technology, great companies and really exciting uses. And we’re only at the beginning.
IW: A lot of companies are putting together Big Data or IoT teams, or both.
JS: We frequently talk with companies that are putting together Big Data teams at the top of the house. They’re making this a strategic priority, and I think it’s for all of these reasons. It’s remarkably consistent how they’re approaching it and the questions they’re asking. They’re all looking for startups outside of their company who can bring technology they can try. They’re intentionally seeking a bunch of proof of concept projects with long-term thinking about technology architecture and business models that are different so they can get going on this. Two years ago, we were driving around, knocking on doors, trying to get people engaged in an IoT conversation. Now, companies have those teams or want to put them together.
IW: Your own team is geographically divided between San Francisco and Detroit. How does that work, with one group in the center of tech and the other in the center of traditional manufacturing?
JS: We are roughly 50-50. Our best asset as a company is our team and the fact that our team is both a manufacturing team and a high tech team. That is the result is being from and in both places. We have encountered concern from potential investors about having split offices. In our minds, it’s the reverse: We don’t see how you could do this without having one foot firmly planted where the manufacturers are, and one foot in Silicon Valley. We have the hoodies and the people who can walk the floor, and you need both to do this right.
We believe team and culture are the competitive advantage in this kind of enterprise. There’s so much translating and understanding that has to occur to get this started for people. Most of our management team is from the Midwest and I don’t see how we could do it any other way. ... I grew up near St. Louis and I’ve been out here 25 years. The cultural differences between the Midwest and the West coast, and between manufacturing and high tech are real, but there is so much opportunity there, it could transform industry.