Podcast #24: Explore Uncaptured IoT Value in the Basics - A Conversation with Rob Tiffany

Good day everyone, this is Ed Maguire, Insights Partner at Momenta Partners with another episode of our Edge Podcast Series, and today we have a very special guest, Rob Tiffany, who is something of a renaissance man in the Internet of Things world, he’s an innovator, he’s been a CTO, he’s a board member at multiple companies, he’s an author, he’s a speaker, and he also has a terrific website which I recommend called theinternetofthings.io it’s a blog, and he also does podcasts which are very much worth listening to. Rob, it’s great to have you join us. 

Hi, it’s great to be here Ed, thanks for having me. 

Rob, it would be great to get a bit of perspective for people who don’t know your story, share a bit of your background and what’s informed your view of what we call Connected Industry, or industrial IoT. 

Sure. Well, because I’m so old…! I got started in the space a long, long time ago, I’d say early-mid nineties, right after I got out of the military, I joined a start-up in Bellevue Washington, called Real-time Data. I guess this was ’94-’95 timeframe. The whole gist of that was connecting unintelligent vending machines, getting data from them to optimize their operations, no inventory, no route drivers, where to go, what to bring, stuff that I hear people talk about like its new today. 

I’m just a kid learning from a bunch of old guys at the time who really taught me everything, so you can imagine how much harder doing connected stuff was back then. You didn’t have smart anything, you had dumb vending machines, and I would say we had different teams; we had people who were looking at the vending machines from the different vendors, the classifications of them. You can almost think of this when we talk about digital twins today, imagine us with blueprints and stuff, you had the instance, the digital twin of certain machines, but we also had like a lot of times you think of the type, a class of machines that are all the same, we were happy to go do that stuff. 

We had a team that was doing the firmware development, so we had our black-box, our firmware writing all your stuff in CE, which would sit inside these dumb vending machines. We had cable connecting… you know, canning machines, coke machines, Pepsi, whatever, the way they push stuff out, and so we’re having to physically connect to all that stuff; find out how much money is in the coin tray, remember you didn’t use credit cards, you had to put in quarters.  

Then we had a whole other team, some of the best RF engineers in the world, we were creating our own modems, we were bouncing packets off 450 MHz business radio towers, we were using technologies like RAM mobile data that Blackberry would automatically start using years later.  

Then there was the computer side, how do you view what’s going on, on a PC in this case. So, just to give you a perspective, this is back in the Windows 3.1 timeframe, going into the Windows 95 timeframe, visual basic! Then putting these packets together and learning lots of hard lessons I got pounded into my brain, about efficiency at the thing-side, on machines. Efficiency in wire protocols, so we were having to create our own bidding coded packets because unlike today… I don’t want to necessarily compare it to what we’re doing in IoT now but, you had to pay by the byte, you had to be super-efficient. It wasn’t like the times when you started to get unlimited cellular data, and so everything was super-expensive, but a lot of times in life when you’re given insane constraints, that makes you be creative about what you’re doing, and you come up with better solutions than when you can be lazy, and you have unlimited bandwidth, and unlimited storage, and all of that. 

All those things put together to build this company, gosh I think we raised $60-million back then. Anyway, all of that was my trial by fire to teach me all this stuff which ultimately would become the Internet of Things, ultimately at some point we did start using the internet, obviously you’re using all kinds of proprietary networks and things; but the idea is the same, you’re moving data about a thing to somewhere else, analyzing it and making decisions. It was really remarkable the outcomes of that kind of stuff, the efficiencies, obviously the pitch to all the vending machine companies was, ‘You’re going to pay us a subscription to have this technology in your vending machines, but we promise you’re going to make more money, or save more money, than what you’re paying us for doing it’, and that absolutely played out as you might imagine. Routes got more efficient, soon you were in the distribution center pre-kitting exactly what needed to go and where. Before that, people were just guessing, got instinct, done; ‘I’m going to visit all the vending machines on my route today. I’m going to have a pick-up truck of all the possible things I can need’, going away from that to being targeted.  

Another cool thing which we learned along the way, lots of times you’re just bumbling and stumbling along, we learned about merchandising. We went from driving efficiencies in those companies, to helping them make more money, when you started to see in real-time what items… say you’re looking at a city and you’ve got vending machines in buildings and offices all over the city, so as it turns out all those usually have banks and machines together. You’d have people working in offices near those machines, and as it turns out in real-time you can start to determine the preferences of those folks, and you find out hey they like Snickers, or white donuts or whatever, and they’re eating more of those than normal. So, you’d start doubling up on those kinds of things, and you’d also find out in real-time what products were not interesting to folks, so you’d move them out. Suddenly, we’d found ways to make these folks more money, then we could sell that data back to the manufacturers of foods and drinks, and things like that. 

So, that was my journey. 

It’s forward-thinking when you think about where we are today, the same problems are driving efficiencies and being able to improve sales, some of the same value propositions that people are arguing are new today in many cases. It sounds like you guys are really at the ground level in terms of connecting data with these physical assets, and then moving past the pure operational efficiencies into being able to be proactive, in terms of the merchandising. Could you talk a bit about what you would consider some of the most important technology innovations, or frankly some of the obstacles that you’ve faced along the way, that we’ve see evolve since which have really allowed this type of vision to become much easier over time? 

You know, the word ‘easy’ is right on the money. I might be one of those people who will sound, who knows how well this comes off, I really think IoT is easy, I really do. Once a year we have a big camping trip out in Eastern Washington with all the guys, most of them are all retired who were part of that team back then, all these guys came from NASA and Allied Signal, the same guys who built the black-boxes for aircrafts and stuff like that. We talk about how hard it was to do what we did back then, and how easy it is to do it today; you can imagine sitting around a campfire joking about, ‘Gosh if we were going to do that same thing all over again, we could be up and running in weeks’, because machines got smarter. I would say the biggest innovation, and I know we’re talking about IoT, and so we’re talking about the Internet, Internet protocol, but in my mind the thing that really blew the doors open and made a whole lot of things possible, really, was wireless technology. 

Wireless data networks by far in my mind have had bigger impact than any other thing, in making a lot of this stuff possible. Now, sure in factories and things where you use ethernet, and you see all kinds of weird ways to get data moved around. But the whole cellular industry taking off from its earliest days, getting early days of cellular data, and yeah it was expensive, but it got better and better, and it has facilitated so many industries. When I think of IoT in its simplest form, I’m remotely knowing something about a machine somewhere, or a process or whatever, and wireless technology whether it’s the growth of cellular networks all over the country and all over the world, or Wi-Fi, or things like that…big time. So, in those early days it was primitive, so the fact that I’m telling you we were creating our own modems, by hook or by crook we had to do whatever we could to move that data. The concepts are the same, but it was hard to do it. 

So, I’m going to stick to my guns, I think the whole smartphone revolution validates me along with most, especially outdoor stuff, when I think about IoT and using cellular networks and things like that. Of course, you’ve got low bandwidth things like SigFox and LoRaWAN and stuff like that, but that’s made things a lot easier. Technology got easier, software development, doing firmware development was so hard back then, as you can imagine, it’s not like its easy today. And then software on computers and things like that it’s got dramatically easier, the programming languages are easier to work with, higher level languages you can get a lot more accomplished a lot more quickly, just great stuff, huge innovations. 

What you touched on earlier was an important point about working with constraints initially, when the cost of storage or the cost of bandwidth is prohibitive, in many respects it forms a far more thoughtful design process when you’re thinking about what data you need to analyze, and how you need to apply that back to the business case. In many respects I think as we’ve got to the point where people aren’t even thinking about being thoughtful about what data that they save, now you’ve got Data Lakes and Data Oceans, Data Glaciers as it were!  

Can you talk a bit about the discipline of looking at a business problem, and then being able to apply the data that you’re pulling off physical assets, how that’s evolved through your career, and maybe some of the principles that you learned early-on, that may not be as widely applied today, given some of the luxuries of easy technologies, how have those principles informed the work that you’ve done since the early days? 

They’ve stuck with me the whole time, through different types of technology, mega trends, things that have happened unrelated to IoT and stuff like that. Do you remember, maybe it was ’98, ’99, 2000 timeframe when the whole web service SOAP thing got big? 

Absolutely, that was another great acronym, yeah. 

In my mind I was thinking SOAP became this big thing, and XML was the biggest thing in the world around technology, and suddenly everybody and their dog is building solutions that can remotely communicate with each other over the Internet, using SOAP web services and XML. I remember my mind it was trained as a kid, doing bit and coded packets, XOR, all of that kind of stuff, a bit twiddling, and how important that was, and then watching the pendulum swing to… and I don’t want to call these people lazy developers, or whatever, but they used what they had, and it was easy. So, of a sudden you’re seeing the use of the most bloated data formats and wire protocols I have ever seen. I know it’s geeky to talk about stuff like that on this podcast, but it’s still important, and it became more important when the mobile revolution really started taking off… 

Well then, you’ve got constraints again, right?  

Yes, because you think about the earliest days of the mobile revolution, so we’ll start moving through the nineties to the late nineties, and the early 2000’s, when you started to get those first smartphones, Blackberry’s, Windows mobile, things like that, some earlier Nokia devices, and you had GPRS, and that’s not terribly fast, it’s a 2088 modem I guess or something, early days. So, suddenly people were building these big bloated webservices, and wanting to communicate with smartphones back then, and they were wondering why the performance was so horrible. 

All that stuff stuck with me, and I remember mentoring, I’ve always been able to play some prominent roles in the software industry over the years, and really pounding how important it is to stay efficient, because having great experiences, ultimately we’d have this big apps revolution on smartphones, yet those early lessons from the earliest days, precursors to IoT, really hit home and made an impact with working with smartphones, working with apps, anything working. Then building large solutions for all kinds of big companies, or transport companies, really trying to pound into the heads of people who were building stuff; even if you don’t have those constraints you need to pretend that you do, you need to assume that whatever you’re targeting is some kind of 2088K modem, with the miserable battery life, slow processors and all that, and if you can be successful targeting that, everything else is gravy. 

I think all that stuff applies today in IoT just as well, because it’s almost Back to the Future for me! Or Return to the Past or whatever when I think about IoT. You don’t necessarily always have the same bandwidth issues unless you want to build your own networks doing LoRaWAN or things like that. But the sensors, the devices, the micro-controllers connected to machines, especially if its things in the field on a battery limited bandwidth, all those same lessons apply for sure. Yes, it’s important to stick to that, that discipline is just as important today as it was in the mid-nineties. 

Yes, it’s a new generation of technologists in a way, it’s like forgetting how to write with a pen and paper, because you’re so used to typing things in. You still need to know how to use those basic principles when the lights go out. 

Absolutely! 

I’d love to dial the clock forward a bit and get your view on the current state of the market. Just set the context a bit, if we go back to say 2013/14 where we had some big moves by GE for instance snapping up a number of companies and building up the Predix Platform. You had PTC making a raft of acquisitions, really starting to put industrial IoT on the map as a sector of promise, and there were a number of forecasts that were, how shall we say, they were optimistic, even talking about economic value-add. I think a lot of people jumped into the fray hoping that we would see an inflection, but the market didn’t unfold that way, it’s been much more of a linear progress, some overshoots and some corrections, but generally a lot of steady progress.  

I’d love to get your view on the evolution of the market since the big boom in expectations a few years ago, and whether some of those overly optimistic expectations have resulted in a bit of temerity amongst businesses in investing, and how you view the landscape today. 

Sure, also some of us may have a different view of this whole thing than others, a lot of people think IoT is some new think that took off maybe back in 2013, or whatever, and some of us look at the world who did end-to-end systems and scale of system, and all that, and we just think of it as an evolution, or it’s just a different way to do an old thing. But there was this perfect storm that got people excited, and part of that was confluence of things coming together, suddenly sensor are getting cheaper and microcontrollers are getting cheaper, and communications whether its cellular or whatever, networks are getting more pervasive and ubiquitous everywhere. Storage is going to zero, all kinds of things are starting to come together that made that more possible than maybe in the past, because people from the industrial world have been doing this forever. 

But the machine-to-machine technologies that have been around for decades were typically only used by manufacturers, or companies that had a lot of money, and could afford this kind of stuff. A lot of those packages came with their own analytics, there’s been plenty of analytic companies over the years and their other products were prohibitively expensive.  So, part of the perfect storm was, and it sounds like a cliché, is when suddenly you start to democratize this stuff, and things got cheap enough where anybody could play.  

Not only did all these elements that I’ve talked about get a lot cheaper which made it more possible, but also the analytic stuff, the analytics went from being the domain of a few companies to, ‘Hey, I can just go to Apache.org and download Spark, or Flank,’ or whatever, Hadoop, you name it.  Obviously I have to know how to work with it, but all this stuff, Open-source finally got to be good enough, and things that used to be really expensive in many cases became free, so cram all that stuff together and you do have an explosion, you have a zillion start-ups that showed up at the door of Venture Capitalists, and right there on their pitch deck they said, ‘Hey, Mackenzie says this is going to be an $11-trillion business’, or whatever! Now we have hundreds of IoT companies, hundreds of IoT platforms up there overnight doing this goldrush type of thing, because of that perfect storm, because of analysts saying, ‘Yeah, we’re going to have 50-billion devices connected, and it’s going to be the biggest thing ever, and we’re all going to get rich’. 

Also, especially if you think about the platform side of IoT, connecting to things, doing edge stuff, doing server stuff, maybe cloud stuff; that technology looked familiar to lots of people, anyone who has ever been a developer, or an architect said, ‘You know what? That looks like middleware that I’ve built a hundred times before, I can do that’, and so I think that led to it. A lot of people said, ‘Well, I’m going to build it, I’m going to connect to stuff and we’re going to do this thing’.  I think that’s how the horserace got started there, and how we got to where we are with so many players, and so much promise, that you’re right, we’ve had a failure to launch a little bit. Where we thought where we would be is certainly not where we are, no doubt about it. 

One of the challenges of course is that we have a lot of these platforms, these application enablement platforms, a proliferation of them which in many respects we’re starting to see a bit of a shake-out, as you have the infrastructure players like Azure and Amazon web services taking care of a lot of the plumbing. But there’s also another dimension which is the unique challenges and opportunities working with traditional operational technology companies. I’d love to get your perspective on maybe some of the challenges, the opportunities and some lessons that successful companies who have been able to bridge that, both the technology and the cultural gaps between your traditional non-software mindset industrial companies, and then the rapid agile ship first, and move fast, and break things mindset. Any thoughts around those dynamics? 

I’ve a lot of thoughts, we should just talk about Hitachi, for the last couple years of my life at Hitachi building Lumada. You’re knee-deep, you’re working with one of the large world’s giant industrial conglomerates; you’ve got Hitachi, GE, Siemens, ABB, all kinds of companies like that all over the world, and they’re one of the biggest players. So, keys to success or lessons learned in that space, because suddenly, you’re a dominant operational technology company that can really walk the talk too. I’m hearing lots of companies these days, they’ve been looking at the big pie chart of IoT, the different segments, and trying to figure out, ‘Hey, where am I going to make money in this space?’  

A lot of people try the consumer thing and that hadn’t really worked out, because there’s a lot of ‘nice to have’ stuff, and I’m seeing more and more companies over the last three-years maybe going, ‘Hey, I think this industrial space is where all the money is’. They’re right, it is, and in the industrial space they’ve been doing this already for a long time, so it’s not a Johnny-come-lately sort of thing. They also have more to lose which is a key element to being successful if you’re an IoT company, delivering a platform or whatever. 

These OT guys are not big fans of IT guys for starters. Lots of people talk about blending IT and OT, but I’ve had the luxury of spending lots of times in Japanese factories for the last couple of years, and working hand-in-hand with folks, control systems, really making it real, and there’s a lot of things to glean from all that. We built this industrial IoT collection of technologies called Lumada, and the success of that product, and our success with the new Gartner Magic Quadrant for that, it came from being in the trenches working hand-in-hand with all the OT guys in the factories, and the bullet trains that Hitachi makes, and things like that, making it real. There’s lots of decisions you have to make, like it’s easy to talk about all the cloud providers, Azure, AWS and Google jumping in the IT space. I was a part of that too, I was part of the Azure IoT team.  

There was an early belief, it was wrong, that IoT somehow required the cloud, I heard a lot of people trying to convince other people that that was the case, and that there would be this IPV6-1 to many relationships between everything on the planet connecting to the clouds. All you have to do is go to talk to someone in a factory, powerplant, oil refinery or nuclear reactor, or whatever, and you find out that you’re just dead wrong, really fast. You find out there’s lots of people who say, ‘Data doesn’t leave my factory’, and suddenly you can either resist and say, ‘No-no, didn’t you know, didn’t you read? Everybody’s going to the cloud’, and they’re not interested. So, there is this highly paranoid, highly suspicious air-gapped world, and when I say suspicious, I mean there’s lots of companies in different countries around the world, who don’t want their data from their factories getting outside their factory, because they think they’re competitive, they’re going to find out what they’re doing, or find out they have a competitive edge or something like that, whether it’s real or perceived data sovereignty inside the factory. 

Most OT people I spend time with, not only do they have a high distrust of IT; you know how we often talk about obviously security being one of the biggest blockers to IoT taking off, the OT guys think that IT causes all the security problems! Now, we can look at OT and machines and say, ‘Well, a lot of those don’t have any security at all’, but it doesn’t mean that those guys are going to believe you, or whatever. It’s weird, and so you have to have a deep partnership and be on the same page, and it can’t be confrontational, or the IT technology people thinking, ‘Ah, we’re smarter, we know what to do’, it’s just a non-starter. And so Lumada came from that joint deal, Hitachi has been an OT company for more than 100 years, and it’s been an IT company for more than 50 years; Hitachi started making mainframes a zillion years ago! And has been in the tech business for a long time. It was great having that blend from the get-go and working hand-in-hand with control systems.  

Whilst we race ahead and talk about AI, blockchain and all this other stuff, it’s important to realize, that most of these machines are still not connected today. We’re talking a good game about what you should be doing, I can’t totally back this up, but I’d say, 80-90-95 whatever-plus percent of all these things, these end points, these machines in the industrial space, they’re not connected to any IoT system at all, still today, after all this hype and after all this talk. A lot of times an MTM system will come with its own basic analytics package, like you’ll see a Windows ’98 machine on a factory floor showing you some charts or whatever, you’ll see some of that, but the reality is we’ve got ahead of ourselves, most of the stuff is still not connected, and that’s a problem. So, being able to connect to industrial machines, it’s not like doing JSON calls to a Raspberry Pi, right? Its working with industrial machines, and a lot of these are still talking over RS-232 serial, they’re doing Modbus and BACnet, a zillion other protocols that most of these platform players out there have never even heard of.  

I still think connecting and interpreting, like let’s say you’ve got an Edge Gateway, Lumada had stuff to talk to endpoints that were smart, directly; machines that had their own microcontrollers, and CPU storage RAM networking, and they could talk all the way back to Lumada Core, and that’s great. We also have Lumada Edge, Gateway Components that would help to talk into machines who maybe needed a little help, we have to integrate and use LPC or something like that, and translate, and know that hey, this strange register on this serial port means temperature! Or whatever. And so, doing all that hard work at mapping, I think people are continuing to overlook the importance of just getting connected, and connecting the machines that are most valuable that have the biggest payoff for customers, and getting them monitored, are the very hardest to talk to. That will change over time, but as you probably know a lot of factory equipment, industrial robots, things like that, a lot of these machines are in service for decades. Companies don’t just rip and replace because you want them to have a Raspberry Pi, or Arduino embedded in them. 

Yes, its decades-long capital replacement cycles. 

Absolutely, absolutely. I’m a really pragmatic person, I’m really interested in getting the basics right instead of racing to things that most people can’t understand, and so being able to connect to those high-value machines… and something I said earlier, I talk about people have something to lose; the reason there’s success for IoT companies who have the right technology, have the right people, have the background in the industrial world, they can be more successful because as you can imagine if you’re running a factory and you’ve got all these machines, you know that if a sub-system of some machine starts to fail and it has a cascading effect to the rest of it, it’s enough to shut down an assembly line or something like that. Or, a pump-jack, an oil rig fails for some reason, any time you have these machines where downtime is equivalent to potentially hundreds of thousands of dollars per hour, or per day, those are the folks who are absolutely interested in signing a check for your industrial IoT technology, to help head that off at the pass. As opposed to one of the consumer things which I consider ‘nice to haves’. 

Can you point to any industries that you feel are particularly effective in harnessing events, technologies, or connected technologies, and if there are any common lessons from the folks that get it, versus any other industries that seem to be lagging, or missing the boat a bit? 

Well things that don’t always align as well with IoT as we might like them to, like financial services, I think the financial services industry is probably one of the most technically advanced industries in the world, in stock exchanges and things like that. We sometimes think of healthcare as a laggard in some ways, but they’re hi-tech in other ways, a lot of what we think of as IoT just came from the manufacturing sector for things they’ve been doing, or utilities or whatever that had scatter systems. So, they were technologically savvy, but maybe they continued to use the same old stuff that they’ve had for maybe 10 or 20 years, which may make you think they’re a laggard. 

When you think about why someone wants to do IoT, people talk about a lot of reasons why, but its super-important to cut to the chase, why are all these companies in business? They’re in business to make money, and so you really need to focus on, is this technology going to make me money? Is it going to save me money? That kind of thing. There’s depreciation, and stuff like that, so a company who has had some technology which maybe seemed advanced a few decades ago might seem not as advanced today, but a lot of them are like… I don’t want to say it’s like if it ain’t broke, don’t fix it but, they’re getting enough incremental value, that they’re not seeing enough extra value from a new system. For instance, as you’re aware we’re in proof of concept hell right now, everybody is doing POCs, where you don’t have enough stuff happening in production across all companies that are in the IT space, all customers, that’s well known right now, and people are wondering why are we not breaking out? Ultimately, as you can imagine, it does come back to getting additional incremental value from that POC that’s worth it.  

So, a while back in our conversation we talked about IoT exploding maybe 2013 when things get going, and a perfect storm of things getting cheaper, but you could still make the case that in spite of all those things getting cheaper, they may not be cheap enough still depending on the use case, when the sensors, the MCUs, whatever, wireless or wired , whatever connectivity, whatever cloud or data center storage, all those things, you have to add all those things together. In some cases, unfortunately, that dollar amount exceeds the ROI that you could get for doing it. So, some companies go, ‘Well, you know what, it’s kind of cool but it wasn’t worth doing’, you know what I mean? 

Absolutely. A lot of technologies that have been in the incubation phase for a while, are now starting to be ready for primetime, and this is one of the areas where we’ve been focused on a lot. Certainly, when you look at what people call AI, or cognitive, or machine learning, we’ve had some major breakthroughs in ease of use, and accessibility with tools like TensorFlow, and of course VR and AR are moving up the usability curve, we have a few other things, collaborative robotics, soft robotics, different types of automation, and even blockchain. I would love to get your thoughts on some of the new tools and techniques that you think might have a meaningful impact in accelerating time to value, of connected technologies. 

You’re totally going to hate me! I really have a contrarian view on this to be honest with you. 

Bring it on man, we love contrarian views! 

Okay! I think it’s all great, I’m still trying to find a place to sprinkle blockchain and IoT, I know some people are doing it, a lot of people say that they have to be together somehow. I’m happy to find the right use case where it makes sense, blockchain is the world’s slowest distributed database, most of its use cases turn out to be the exact same use cases we’ve been using relational databases for, for decades. Sometimes when you think about the high throughput, high velocity world of IoT, that doesn’t align with the slow as mud world of blockchain trying to get things done, and the extreme energy usage that blockchain requires.  

So, I’m always openminded to new stuff, but I don’t want to do it for the sake of doing it. AI scares the heck out of people. This group that we’re in, this IoT/AI/machine learning/data scientist/ whatever group that we’re all living in, we’re all talking to each other, and we’re all reading the same websites, and the same articles, I would posit that we’re our own little echo chamber and we think it’s normal. It’s important to get into your car and drive away from Silicon Valley, or Seattle, or wherever, and drive out into the rest of the country and talk to regular folks who are trying to get stuff done. They are scared to death, because we are conflating IoT and things like AI, and we’re saying them in the same sentence, and the average folks who are just trying to send natural gas through a pipeline, or do their stuff, they’re freaked out by it, and we’re at a problem where we’ve still hardly got anything connected. We still need to solve that, and that’s still the most important thing in my mind, it’s the least sexy thing, it is the lowest value part of IoT, and it’s the part we still haven’t figured out in a big way yet.  

When folks hear AI, deep-learning and stuff like that, and they glum it all together in their minds, it really puts them off to be honest with you. I’m not just saying this, obviously I talk to real customers all the time! 

Yeah, in fact last week I interviewed Troy Oliver who does a lot of meet-ups and community events for regular non-tech people, and it’s clear that there’s been a lot of thud around AI, and they’re selling a lot of vision, but I think we’re still super-early-on with… 

Yes, we as an industry, as thought leaders, as people participating in this IoT world, we need to do a better job of conveying a simpler message that everyone can understand, and convey to customers, that there is tremendous value in having your machines connected. Just the act of remotely knowing something, instantaneously has value before you apply any kind of analytics at all. So much of the world today is people taking logs, walking around powerplants, refineries, driving out into the middle of nowhere in pickup trucks to look at oil tank fields, or different things like that. Flying on planes to find out stuff, people doing service work out in the field, they’re capturing data, either driving back to the office and transcribing it in computer systems; all this stuff that you thought was part of the 20th Century and couldn’t possibly be still around anymore, it is still here, and so is paper and pencil. 

So, all that costs time, money, inaccuracies and things like that. Being able to have sensors, video, wireless, put together to do all that stuff, then remotely know it instantly, there’s just dramatic tremendous value, and people need to recognize that. They need to be able to think, ‘Yeah, I’m going to do this IoT thing, because I’m going to save this much money, and this much time. I’m going to gain these efficiencies, I’m going to remove human error in transcribing stuff’, lots of stuff. This is the easiest stuff there is in IoT. The very basics of IoT is just remote monitoring, it’s not sexy, it’s not AI, it’s just, ‘Yep I know it’s gone on over there’. Well, we’ve been doing it the hard way for a long time, and we continue to. 

So, if you can get to business decisionmakers and say, ‘We’re going to do this simple thing, and here’s the use case, and here’s the ROI. Here’s what you’re going to save, and here’s how things are going to be more efficient’, that’s a great message, and it didn’t include anything about deep learning! 

And that gets to a question which I always like to ask, which is about disconnects in the market, and what you’ve just articulated is that technologists tend to get enamored of bright shiny objects, and the latest and greatest things, but I think the point that you are making is really important; it’s that there’s a hugely underserved opportunity to bring fundamental value to businesses, just by providing basic connectivity and visibility into operations. So, it’s your view that many companies may be over-selling events, technology, and maybe underselling, or under addressing the real problems that can be effectively addressed in a straightforward manner. 

They absolutely are, they are overselling, they are enamored. Or, they think that they need to talk about these things, they think they need to talk about machine learning and different algorithms, having an army of data scientists and training models, training models at the edge and all that kind of stuff. They think they have to say all that to make their IoT platform, or whatever, seem valuable enough to a customer. It’s just not true, and if it’s true or not, I can be left-field, but I can sit back and look at the reality of the situation, none of this stuff is going to matter if we don’t get these things connected! So, people need to do a better job of articulating value, ROI, how much money they’re going to save, things like that. How they’re going to do better serving their customers with simple stuff, and so, hit it hard with just remotely knowing, and compare that to how you remotely know things today, and what that costs you. Then, when you’re ready, later-on the next thing, layer on simple-simple-simple analytics, it’s this and that, right? KPIs, its easy stuff, it’s not rocket science.  

As you might know, most companies today, and they might be embarrassed to admit this, they still run their business on Excel. Excel is still the number one analytics tool on the planet. The second-place player is not even close. So, talk to them in a language they can understand, talk to them about KPIs, talk to them about things they can do in Excel, because I guarantee you, giant companies that we can all think of, and would be horrified to think that they run their business on Excel, but they do! You’d be amazed the kind of stuff you can do in Excel, you probably know, you’ve a financial background.  

Oh yeah. 

Talk like that, when you talk about analytics, just layer on, ‘Okay now we’re going to react to some KPIs here, we’ve got these values coming in…’ 

I know it sounds simple, and it seems like it’s almost embarrassing to say this, because surely, we’re so far past this, but I think we need a better job, and so say connecting is going to get you this much value, simple KPI analytics is going to get you this much more value. I read stuff all the time, you talk about energy savings and things like that, I know folks who are in that space doing IoT, managing HVAC systems, and guess what? They can go to a customer and say, ‘Yep, I’m going to save you 20 or 30 percent a year on your electricity for your building, or apartment complex…’ or whatever, and guess what? All it took was being connected, having some updated stuff for thermostats, and just really the dumbest simple, this and that type KPI analytics. It’s even hard to call it analytics. And at the end of the day the customer will go, ‘Awesome, all I heard was blah-blah-blah… save 30 percent a year’! 

It’s an important message, pay attention to your fundamentals, I guess it’s a little tough love not to chase the sexiest new idea. I think that’s a valuable message. 

One question I do have though, as you look forward over the next five or ten years, are there any key catalysts, or enabling technologies, whether it be 5G or anything else, that you think could make a big difference in how the market evolves? And the last twist to that is, what you’re most optimistic about as we look forward? 

5G is going to take a while to rollout, if you know the concepts behind it and lots of small cells to get you lower latency, and things like that. So, as you can imagine, it’s not going to be like the change from 3G to 4G or LTE, and things like that, early 5G is still using LTE under the covers for a lot of stuff. I don’t think there’s anything magic about 5G making IoT better; once it rolled out things will pop up, things that are latency sensitive, where we want to do something remotely instead of onboard well that will certainly be helpful. 

I’m glad that there’s lots of talk about the edge! One of the first decisions I made in creating Lumada was, Lumada’s got to be portable, and all you have to do is go and talk to customers and you’ll find out that in lots of use cases, again, they cannot handle the latency of going to a distant cloud, or the possibility that the Internet goes down, or whatever. I think there’s so much irony in cloud providers saying, ‘Well the edge is really an extension of the cloud to the edge’, that’s just ridiculous. The edge is the default location, the edge is already here, the edge is what we’ve been doing this whole time. The cloud is something new, I’m not bashing the cloud because I think there’s a lot of great usefulness for cloud stuff, but the edge is the default location and its where its already at. 

We’ve been doing that forever, that’s why we had PCs and servers on prim, or in a factory, that’s what customers say, ‘I want all these things in my factory, not even in the data center’. So, we made sure Lumada was portable from the get-go, Lumada runs right at the edge, right where the machines are because that’s what our customers wanted, and it can run air gaps. So, that’s cool.  

Newer technologies; I think a few things, it’s not necessarily a new technology, we need to do a better job of using stuff that’s already out there, or not racing to create yet another thing. Even though we have things that are a lot cheaper now than they ever were before, they need to get even cheaper so that if someone wants to make an IoT product, some kind of machine, and they want to attach a micro-controller and everything, all the other stuff they need in order to make it a connected product, all those things together; connectivity, storage, servers, MCUs, all that kind of stuff, still need to get dramatically cheaper so that it is a tiny fraction of the cost of the thing, the product they’re making, so it’s worth doing. Because people are going to be fed-up of paying 100 percent more, or 500 percent more for a connected product, than they are for a disconnected product, or at least one you can plug into a serial port. 

Another big thing, and this is geeky. I spend quite a lot of time thinking about… you what read folks say are the big blockers, to adoption, sometimes its cost, sometimes they say it’s too complex, they say their IT folks do not have the skillsets, they talk about security, and so I think it’s important to tackle those things head-on. I’ve been caught up in things just like everybody else, so there’s a lot of IoT snobs that all they do is talk about wire protocols, and why MQTT is the most efficient wide area thing to do, pushing stuff back and sending commands, or whatever. So, I’ve lived that for a long time with Microsoft, Hitachi and all that stuff, my realization now is, to me it’s more important to make sure the space succeeds, than to use the very-very best technologies, if that makes sense? I’m a total believer in that we need to abandon all the crazy stuff, I’m a fan of just using the existing W3C web technologies throughout there to be honest with you, for long-range stuff.  

I guess if you compared REST, HGP, REST to MQTT, REST gets trounced, but HGP is the most pervasive widely used, widely tested thing on the planet, nothing is even close. So, I’m a big fan of open standards, but I’m also a big fan of well-used highly pervasive standards that everyone knows, and everyone understands. So, I think I’ve evolved over time from always using the best of breed, efficient, whatever edge-case type things, to thinking it’s more important for us all to succeed than to focus on these geeky things.  So, I think we shouldn’t be building the next protocol! 

But to focus on business value, right, which is what it all comes down to? 

Absolutely. 

One final question which I always like to ask, if you have a recommended book or resource that you could share, anything that you feel strongly about? 

Yes, and this will align with what I’ve just said. I’ve been thinking of this independently, but I know some guys of the W3C who’ve been working on this notion of the web of things, and I think Mozilla’s in on it now too. A couple of guys who work in the W3C, Dominic, and Vlad, they wrote a book called, ‘Building the Web of Things’, and they basically outline how you can do every last thing that you want to do, in IoT, using just basic W3C standards, they’re used everywhere. 

In fact, I interviewed Dominic for one of our blog interviews. Oh, that’s great, exciting. 

Yeah, I evolved to that view a while ago, and then I connected with those guys, then they came out with their book recently, and I read the book. You’re right let’s focus on the business value, and let’s not get so caught up on all the geeky stuff. Let’s use something that’s out there that’s pervasive that everybody knows. So, ‘Building the Web of Things’, really illustrates that. There’s one of the IoT platforms, I think it’s called, EVRYTHNG, It’s an IoT platform, it espouses a lot of those concepts. 

He’s got some great principles there. That really ties things together nicely Rob, and with that, I think we’re running out of time, but again I want to thank you for taking the time. 

Rob Tiffany who again is an IoT Renaissance man, please do checkout his blog site, theinternetofthings.io some great content there. This is Ed Maguire the Insights Partner at Momenta with another episode of edge. Rob, I want to sincerely thank you again for taking time to speak with us, this was really a lot of fun and very insightful. 

Absolutely, thanks for having me on Ed.