Blaine: Joining me today is Rob Tiffany, Chief Technology Officer and Global Product Manager for Lumada IoT at Hitachi Vantara. Prior to his current position at Hitachi, Rob was also an IoT global technology lead at Microsoft, currently on the Washington State IoT council, and is also an advisory board member for the UK based Smart Cities world conference. Thanks for the time Rob.
Rob: Yeah thanks for having me! It’s good to be here.
Blaine: You bet and we’ll make it fun and interactive as always. So, Rob, that’s quite a background I laid out for you. Why don’t you tell us a bit more about yourself, your current focus and what brought you here, to be an expert in IoT among other things?
Rob: Well, gosh. I’ve been doing IoT a long time. It’ll totally date me, but I kind of got involved in something you’d call IoT back in the early mid 90s, actually. I worked at a startup called Real-Time Data when I got out of the military.
Blaine: You’re on this TV channel: The Real-Time Enterprise! So there you go.
Rob: That’s perfect. And the thinking of this startup was bringing these old vending machines to life. I hear about no shortage now that IoT is so in vogue. I hear all kinds of people using vending machines as their example for their use case for IoT: inventory and stuff like that.
We did it back then and it was really hard, as you can imagine. The machines were completely manual. We had like three different teams: you have people who are writing the black box firmware, cabling, retrofitting old machines. You had some of the top RF engineers in the world working on it. It was like the early days of wireless and wireless data. And it was expensive. You had to pay by the byte. And then you had folks building the application.
To date me, this was running on Windows 3.1 with a basic visual app that looked like you were looking at a vending machine.
Blaine: You bet – a digital twin from from the 90s.
Rob: Yeah, but you know what? We got it to work. It was really hard. All the value of real-time inventory, knowing where to send the route drivers what to bring and what not to bring, finding out what customers preferences were, doubling up on a certain customer in a certain building that really likes Snickers. You could add more Snickers.
You could see real time. So pretty soon you were like, “Oh! OK We’re making money and we’re saving money and doing all these great things that people talking about today around IoT.”.
Blaine: So why has it taken so long? I read something you quoted in your blog about IoT just now moving past the shiny object stage. If we’ve been doing this for the last 20 years, and I know we have to some degree, why is it only now just moving past the shiny object stage?
That’s a good question. IoT has certainly underperformed. We’re certainly nowhere near the McKinzie 11 trillion dollar thing that they talk about. IoT, machine/machine embedded systems, control systems, data systems have been around for decades. You’re right. On some level, none of this is new.
I don’t want to come off as some jaded guy but I really feel like this is all evolutionary. I think, maybe, why we talk about it so much in the last three, four, five years might be a perfect storm of things coming together. For instance, I work at Hitachi which is a giant, Industrial conglomerate you know like Siemens or GE. Those folks have been doing machine to machine stuff and had embedded control systems forever. It’s nothing new for them.
The reality is, though, it’s not something that was widespread for everyone else. A lot of this did come out of heavy industry and manufacturing. The machines were so expensive that they would have microcontrollers in them, they would have sensors, and they’d have their own closed loop systems for analytics on how well those things are doing. It’s pretty primitive, but better than nothing.
A few years ago, all of a sudden – I think it’s about money – the price sensors started to drop like a rock, the price and ease of using a microcontroller. I know consumers think of things like Arduino or Raspberry Pis, but microcontrollers got really cheap and more accessible.
Storage to store all this data, storage is going to zero. Cloud storage, I mean, nothing is cheaper than that. Another crosscurrent happened with analytics. Even if you could do all this other stuff, this advanced analytics, trying to do things like machine learning, AI, and all that fun stuff. That was out of the reach of just almost everybody. So you had some really high-end companies that have been doing analytics, SaaS, a whole bunch of people for forever. But not everybody knew how to work with that stuff.
Today, Anybody can go to apache.org and download some open source stuff. They can download Spark, Flink, and storm, you name it. There’s something new popping up every month it seems like. All of a sudden, all these advanced analytics tools are free and they’re open source, for the most part.
Now, you just need to have the know how from people like data scientists, but we’re even trying to lower that bar and bring it to more people there too. I think all those things coming together is making this exciting and viable. That’s why the talk is there and that’s why it’s taking off. But your question is still valid. We still really haven’t taken off in a big way like we thought we would. We’re in a world of POCs I feel like right now. A lot of pilots.
Blaine: It’s interesting. I absolutely agree with that. Although, I have felt in the last, even in the last year, that things have really, finally started to move out of the POC stage and into the real world stage. I’ve been attending IoT conferences for a few years now, not nearly as long as you’ve been focused on the market, but I’ve absolutely sensed a shift there: a mind shift. Am I imagining things or do you think that’s starting to become real?
Rob: It is starting to become real but it’s certainly baby steps. Boy, there’s been no shortage of opportunistic IoT conferences popping up over the last several years. And of course, you’re going to hear that it’s all working great. I would say so far, the incremental value derived may be from some of these POCs still is not exceeding the cost to put a lot of these solutions together.
Even though I just spent the last few minutes talking about how cheap everything got, there are so many pillars that make up an IoT solution. When you put them all together, it’s very possible that they still exceed the cost, the ROI you’re looking for.
We live in this little echo chamber where we all read the same stuff and we talk to the same people and everybody keeps saying AI or deep learning. Other people who are not part of this club keep seeing IoT and AI together. Those regular folks are trying to run their business . They’ve got natural gas pipeline and that’s their business. They don’t know anything. When they hear that IoT and AI go together it scares the heck out of them.
What we all have to do a better job of is telling a simpler story; a story where value can be derived with very, very simple things. I would posit, value is derived just by connecting your things and remotely knowinging something. I know that doesn’t sound very sexy.
I say this all the time, and sometimes I wonder if we really are in the 21st century or not. So many businesses are still out in the field writing stuff on paper today all over the place. I see it all the time. I used to freak out. I don’t freak out anymore because it’s still as commonplace as it was decades ago.
You will still walk around power plants, nuclear reactors, factories, whatever with a clipboard and they’re reading gauges and they’re writing stuff down, and then they’re walking back to the office and they’re transcribing it into whatever system or record they have. That is still all going on.
I over use this example but I’ve heard it so many times. Oil company folks, I’ll talk to people where literally some guy jumps in a pickup truck, goes out to the oil tank field, and puts a big stick in the tank to see how much oil is left in the tank to check if they need to replenish it. I’m like, “You’ve got to be kidding me.” But, It’s true.
And so, just the simple act of adding a sensor to tell him how much oil is left in there. This requires the cheapest of wireless. Just remotely knowing that instantly saves you all this time and effort. I think it’s important to send a message and improve the message that there’s tons of value to be derived with simple things. There’s extra incremental value to be derived from simple analytics, even.
I know guys who are in energy savings IoT part of the space know. They are just focused on going into an apartment complex or some building and they rewire the HVAC system and they put their IoT system in there. With just simple if-this-then that analytics, the customer is saving 20 or 30 percent a year on their energy consumption. Customers love that stuff. You didn’t have to do ML, deep learning, or anything like that. I think it’s important that we do a better job of speaking and talking about solutions that average people can wrap their mind around when doing this stuff.
Blaine: I fundamentally agree. It’s about starting small and getting real value with something simple. To your point about AI and ML, I was at a conference the other week where an analyst firm showed a research study that was backed by a lot of data that 80 percent of equipment or machine malfunction is not predictable. It’s effectively random. You can apply a lot of very fancy machine learning, predictive algorithms, and everything else and that will help you with the 18 percent that actually could possibly be predictable. For the rest, it’s not. So, keep it simple.
And I think the value is partly in having the analytics. It’s also about being able to take an action in real time. Again, you sense that the machine on the oil rig is running out of oil or running low on fuel, and then you’re notified you can take an action before it runs out of fuel so you don’t have to shut down the operation. This is not rocket science. It’s use cases, but it’s very valuable.
Rob: Maybe a lot of people in our little club are worried about talking like that in public because it doesn’t sound very smart. They’re worried that if they’re not talking about deep learning every day, they’re viewed as an idiot. .
Blaine: At the same time, the pace of technology change and the drivers are hitting this space at an amazing speed. You mentioned AI and ML already. There’s augmented reality that everybody’s talking about right now, edge computing, distributed computing, 5G is on the horizon in the not too distant future, everything has to change and there are valid reasons to. Edge computing is a very valid concept. It makes a lot of sense.
You’ve got a gigabyte of data coming off your oil rig every day and if the rig starts to shake and it’s going to blow up, you don’t want the data going to the cloud, being processed, and sometime later you take an action. It has to happen now.
Rob: Absolutely. Maybe the last four or five years of the IoT days let’s say there were a lot of smart people who liked to just tell people, “This is what this means. This is what it is.” If I asked 100 people what digital transformation is, I’m going to get a hundred different really authoritative answers and they’re going to look really confident while they’re saying it.
A whole lot of people said IoT has to include the cloud for quite a while. I spent a lot of my time at Microsoft. We built Azure IoT. There was this notion that there’s going to be this one too many thing; a cloud and then trillions of IPB 6 devices are going to talk to the cloud and it’s all going to be so easy.
Until, I went and talked to one manufacturing company and they said, “This is really great technology. Show me the version that runs right here on my factory floor.” Then you start tap dancing. You say, “We don’t have that.” And they say, “I don’t have time to shovel terabytes of data every day to your distant cloud to get munched on by your analytics and then maybe tell me something too late.” It needs to be right here and it needs to happen in milliseconds.
Whether I’m in Germany or Japan when I talk to a lot of manufacturers, what do they say? “The data doesn’t leave my factory. It’s not leaving these four walls. I don’t care about what you’re telling me about how everyone’s going to the cloud. I’m not going to the cloud.” So, you either leave those people behind or you need to meet those customers where they are.
Blaine: Of course. And 99.99999 percent of the data is actually meaningless. It just needs to be looked at, processed, and thrown away immediately. It’s the exceptions. It’s the important business events, manufacturing line events, or whatever it is that you need to take an action on in real time. So to put all that data up into the cloud, with or without 5G, is irrelevant. I think a lot of people are seeing 5G as the panacea. “Now we can get the trillion sensors all putting data up into the cloud.” But it doesn’t make sense with 5G either.
To be fair to the 5G folks that we’re becoming increasingly involved with, they have, inherently, in that architecture, thought of edge computing as an important tool as well. Even the 5G guys aren’t saying you’re going to put terabytes of data in the cloud every minute.
Rob: Right. Well as soon as you go into a building, all these discussions change. It doesn’t have to be a factory. As soon as I go into a hospital and they want to have all these life support systems and all these medical devices and people patients, we don’t just digitize machines. People are part of that too. You can have a digital twin for a patient and their vital signs just like you do that for machines. Well, they’re not interested in sending that stuff somewhere else either. They need millisecond response times too.
A lot of that stuff is wired. I think 5G is great. It’s going to be great for mobile. It’s great to reduce latency and there’s a lot of good stuff there. But when you’re inside buildings, a lot these buildings are wired. They’ve got Wi-Fi and they’ve got all kinds of stuff which might serve them well inside that building.
Blaine: So we have covered a lot of ground: AI, ML, edge computing, 5G – all good stuff. Let’s bring it back to something you said a minute ago. You threw out the term digital transformation and then ran away from it fairly quickly. Let’s go back to digital transformation, the relationship between that. And earlier, you were talking about the simple stories that we need to be able to tell, the simple use cases we need to solve. At the same time, digital transformation is probably one of the most commonly used terms that gets thrown around right now. So what’s your overall thought on the concept of digital transformation?
Rob: I think it’s similar – and let’s not scare the customer. Let’s let them know it’s OK to do digitisation first, if they haven’t already. Companies have been doing digital transformation, digitization of things, for decades. In the 80s and the earlier, before it was called the IT department, it was data processing. You had people sitting there typing and stuff on a piece of paper and making stuff in the real world digital in mainframes. They may tell you back then that was digital transformation. People may have told you in the early 90s when they came out with OCR technology to scan documents into a digital form , “Look at this! This is amazing. It’s digital transformation.”
I think it’s OK to let the customers do baby steps and I think any company who’s involved in that space, or if you’re a consultant trying to guide a customer, let them do those baby steps. If they haven’t digitized lots of things, let them do that. A lot of people say, “The difference between digitisation and digital transformation is you’re just reproducing an old process or an old way of doing things and making a digital version of the same thing. There is still value in that. Even though you didn’t radically change the process, there’s still value in that first step, just like I’ll say there’s value in simple analytics with IoT.
Like anything in life, digitize, get that early value, and then sit back and look. Now that these things are digital, is there a better way to be doing this rather than mimicking the old analog ways?
Blaine: I agree with the distinction you’re making between digitization which is fundamentally more about taking an existing processor, system, a model, and bringing it into the digital age with digital technologies. Digital transformation is about doing something new or something different; enabling a new business model, a new operating model, whatever the case is.
You’re right. People In the market are sometimes confused between the two. One is not the other. Let’s focus on real digital business transformation, something that is more transformative to the business. I know one of the discussions right now is should these projects be owned by IT or OT; the IT side or the operations side, the business side? What’s your thought on who should own the X in the enterprise?
Rob: Like when you keep hearing about, “Oh we need to have a Chief Digital Officer”?
Blaine: Well, that’s one of the possible attempts to answer that question: neither, actually. We’re going to have this person who is both IT and OT and they’re going to own it. That’s one possible solution.
Rob: I remember hearing that we needed to have Chief Mobility Officers. I hear what you’re saying. I think if IT was more empowered these days than it actually is, it would make sense for someone in IT. You’ve always had CIOs and you’ve had CTOs kind of doing their thing and then all kinds of other people right beneath them. You can make the case is a Chief Digital Officer that different?
I won’t name any names, but lots of companies have been trying to do the IoT thing and they started in their own company first. They tried to have Chief Digital, or whatever you want to call it, Officers and different business units to help them embrace and grease the skids for this new digital technology or IoT technology. There’s always lots of pushback in the business units. It’s human nature. People want to do things the way they’ve always done things, right or wrong. A few people will be early adopters or they want to get on board and do that kind of stuff.
The last decade and a half, we have we really seen IT lose its power.
Blaine: Why is that? It seems counter-intuitive to me because with, of course, the technology revolution is just continuing to accelerate. And so, that should imply that IT should be getting more and more and more power in the organization as everything becomes digital. They are the only ones that can make sense of all of this. Why have they lost their power? Explain that to me.
Rob: I’m going to give you a definitive answer.
Blaine: Ok I’m ready!
Rob: It’s because I worked at Microsoft so long.
Blaine: Microsoft caused it? Is that what happened?
Rob: No. I’ll tell you who caused it. I’ll tell you one thing, the ability to accurately see it was really relevant to being at Microsoft for so long because we absolutely will watch. Microsoft always sold to the CIO. That was what you always sold to all of them. “Hey how many copies of Windows do you want to buy today?” That kind of stuff.
I watched this happen inside Microsoft. They started to realize they used to talk 100 percent to IT, IT directors, managers, and the CIO. We’re noticing that we’re not hitting our numbers. We’re noticing things aren’t working like they used to. There’s a problem here. You started to see this pivot. You learned about the pivot from sales training. You need to start talking to the chief sales guy or the chief CMO, all the business units. A lot of guys scratched their head and said, “Why is that. Why do they care about technology?”
I’m going to go ahead and blame salesforce.com if I have to blame someone, but this is actually a compliment. SaaS Software is what started the change. This is just my opinion. I could be in left field. I think when salesforce delivered a SaaS offering, certainly the biggest, most well-known SaaS offering to date and maybe the first, all of a sudden, what did you see? You started seeing business units going around IT and signing up because it was so easy to just go online and sign up and start consuming the service.
Also, you had this thing where the business units really got sick of them. A lot of times, you’d hear them call IT “the Department of No” And so, a lot of them said, “You know what? We’re going around you.” Salesforce really got that engine going. It really facilitated it. So, I think that salesforce kicked it off and that whole SaaS movement really started moving. It started empowering business as they go, “Hey we actually have the budget and this is exactly what we want: this web-based delivered thing and it’s really empowering our sales force and we’re loving it.” And so I mean that’s kind of what kicked off.
And then the clouds made things even weirder. Then, you had to retrain salespeople again. They thought, “I’m going to sell this product. High five and I’m out of here.” Then you find out it’s about consumption and consuming cloud services whether it’s Azure, AWS, Google or whatever. It’s interesting that I’m sitting here telling you that observing salesforce is my leading indicator on what’s working and not, but it’s certainly a good clue.
Blaine: Very interesting. Do you think that trend is just going to continue or is IT going to find a way to claw back more relevance based on maybe attempting to embed more deeply into the business units or being led by strategies to keep people as more strategic mandates like CEOs? What do you think? Or is this just going to continue to the point where, as technology becomes more simple to implement and the complexity abstracted away, maybe IT almost disappears? I can almost see two possible outcomes.
Rob: Yeah it could go either way. You’re absolutely right. With fewer and fewer things having to be managed on premise by some kind of department you can see how the need for those folks goes down. Clawing back, ITs going to have to fight back to get a seat at the table I think. IT leaders, the CIOs, CTOs, and everybody else, they’re going to have to prove to the CEO that they’re strategic thinkers and that they are the ones who could really do this digital transformation and really hit the bottom line.
Too many people are focused on technology for technology’s sake. That is just not what these guys are interested in hearing. It’s got to be about money. We’re here to make money. That is the only reason we have this business. If you’re here to help us make more money by making things better for customers, delivering right-time experiences to customers on their mobile phones, and all the things that go with that, then we’re all in. But if you’re going to sit here and just talk about what kind of servers you need to buy from the data centers to run the e-mail server, we’re not interested. It’s on them. It’s on their leadership and in the caliber of people that are hired in those organizations to really fight back.
Blaine: I fundamentally agree. I was at a manufacturing conference a couple of weeks ago that was focused on the OT side, the operations side not the IT side. I was hosting a lunch roundtable on the topic of turning your manufacturing organization into a real-time operation. I asked, “Who owns this transformation of your manufacturing operation in your organization? Is it you guys or is it the IT side?”
You can imagine what the response was. Of course, they say it’s them. But, when I asked why it was them, the vociferous negativity around their IT organizations, and these are some large multinational companies – well some medium sized businesses but including some large brands that everyone here would know -, I was shocked. Like you said it, they see IT as as the “Department of No”, having a long lead time and a backlog to get on and they need these projects done now not six months or a year from now.
The lack of of trust and potential for collaboration between IT and OT became really visceral to me. We have these discussions all the time, but to hear it from these guys was like wow! It was very, very convincing to me. The business side and the OT side is going to continue down the path of taking more and more ownership and control enabled by easier to use technologies that have abstracted away the complexity. And IT, they could fight. You said IT needs to fight. The other thing they could do is just figure out how to collaborate a lot better.
Rob: They definitely need to be the ‘Department of Yes’. “How can I help you. I want to make your life easier.”.
Blaine: They do. They do. And the ‘Department of Ideas’ as well, not only about saying, Yes. I will do that for you, but have you thought about doing this?” Because of the operation side, they might not be aware of some of the implications of what could be done with 5G, augmented reality, edge computing, or real time processing. They have no concept of what could be done.if IT is immersed in these things. They’re in the bubble like you and I are. They know what’s happening. And so it could help formulate that into a business case or a use case for the business side that that business would go, “Wow. I don’t even know we could do that!”.
Rob: Right. An OT guy would probably say, “We already have edge computing. It’s called a PLC.”.
Blaine: Well let me change gears one last time and we’ll wrap it up. I’m intrigued. You talked a little bit earlier about machine learning and AI, I think mostly in the context of predictive analytics and that kind of thing. But there’s a whole other discussion going on right now about what is commonly called, and we talk a lot about it, human-machine collaboration; the fact that systems and people, increasingly intelligent systems need to be able to work effectively to make people more powerful and effective versus just this notion of just replacing all the people out of the system or out of an organization. Are you on the side of eventually, we will all just be sitting in our easy chair because the robots will be doing everything for us or do you see that there’s this future of humans and machines collaborating together maybe like OT and IT should be doing more?
Rob: I’m saying power to the people, dude! We can’t let the machines win. You’re right that the best scenario is collaboration. Like anything, if you push too hard on this machine only intelligence, there will be a backlash. But it’s little by little like the frog boiling in the pot kind of thing. If it’s so radically changed with robots and AI and stuff like that, you’ll see pushback happen and it can all blow up on our faces.
I think people should be actively working and speaking about human-machine collaboration because there’s too many people talking about, “No the machines are going to get more intelligent. We’re going to hit that singularity and it’s game over. We hope things work out for the rest of you folks.” Working together is huge. We’ve been doing that in different ways with computers, just generically speaking, making people more productive. So, I think that’s got to continue on. How can you aid a person?
There’s lots of different ways, too. When I think of robotics, you see exoskeleton type things where a person can lift more than they normally could working in a distribution center. I think just the right amount of intelligence, and of course you can’t stop the flow with that. It’s just going to increase and increase. I think the positioning of it to be assistive to people doing their jobs is a huge thing. You really have to be mindful of that.
And you know what? If you’re not mindful of that in the US, other countries and other parts of the world will remind you quickly. Look what’s going on with this GDPR thing. Europe is telling everyone else, “hey this privacy thing is out of control and this isn’t going to happen.” Well, Europe and other places, Canada will also probably try to slap us down and say, “Hey this whole thing where you’re just going to have robots and AI taking over human jobs, we’re not onboard with that and we’re not going to do it.” I think people need to be thoughtful about how they’re proceeding to make sure that it’s a hand-in-hand type thing. I’m sure that there’s a million examples of how to make that work, but I think that’s got to be the mindset instead of the be all or nothing deal.
Blaine: I agree and it’s thoughtful for the moral, societal reasons and also thoughtfulness in terms of it’s just the right answer to figure out how to have technology and people work more effectively. I was watching old episodes of the of the Americans. I’m catching up on all the theories you know right now. For those that might not be familiar, it’s set in the 80s. There was a scene last night where the FBI is in this conference room trying to figure out what the Russians are doing and they have all these paper documents printed out. There’s not a computer, not a smartphone, nothing anywhere.
I watched that scene and thought, “Oh my goodness. We must be ten times more productive right now than they were back then flipping through these reports to try to find one fact that you’re looking for that took a guy an hour versus searching in three seconds.” Now, that’s more about what I would call human-machine interface. It’s the old notion of how people work with a UI. I think increasingly intelligent machines will enable a true collaboration. A software system, whether it’s an actual robot or just the software system, can leverage the experience, the intelligence, the wisdom of people that will take a long time for the software systems to to ever get there.
Rob: I think people bumble and stumble along because it’s people building this technology. So you’re right. In the 80s, the interface was probably paper or some DOS machine or something like that. But you know what, just because we got more advanced in the 90s – think about 9/11 – we had databases, but they were all siloed from each other. And so, you still couldn’t connect the dots on data that was going on between different government agencies. They actually may have had the data and more advanced way of getting it, but they didn’t talk to each other. So we learned from experiences.
Blaine: We’ve got a long way to go. Well, to wrap it do you have any key takeaways or tips for a business leader who is trying to figure out how to drive some real-time transformation in their business, digital transformation with all the stuff we talked around that topic?
Rob: Yes. Business leaders need to get everybody on board, bring all their employees with them, and get the best ideas. The transformation is, a lot of times, it’s a process transformation or things like that. A lot of people know about things getting digital. You know it’s funny. What is a semi new thing that we’re talking about, mostly in industrial IoT? It’s the concept of a digital twin.
And yet, the creation of a digital twin, because we’ve had the internet people and web browsers and servers and then we have web 2.0, people collaborating better, now we’ve got this internet of machines, of things. We’ve got to take baby steps there. So, that first step is we’re going to just digitize machines as digital twins and then we’ll figure out how those twins can work together with people processes to make things more efficient, make better choices.
You’ve got to bring everybody with you. You can’t be in a back room coming up with a game plan if you’re a business leader at a company. If you’re a company that makes products to help digital transformation or you’re consultants, you absolutely need to get out of the bubble that we’re living in and drive in your car. You need to drive out to where these people are, walk in their shoes, see what their life is like, and see if it is really that messed up. Is it just a little bit?
Blaine: Well I think that’s good advice, and I’m going to get out of the office right after our chat and do exactly that. Rob, thank you so much for joining us today. It was a really insightful conversation, lots of great topics there. I really appreciate it.
Rob: My pleasure. Thanks for having me.
Blaine: You’re welcome. For those interested in hearing more of Rob’s thoughts, you can follow him @robtiffany on Twitter and also check out Rob’s blogs at robtiffany.com. You can reach out to VANTIQ anytime at firstname.lastname@example.org. Bye bye for now.