🎧 #044: John Petze on Tridium's early days, founding SkySpark, and 10 years of Project Haystack
“So I know that's your favorite question and you want me to answer it, but I I'd reword it: Why is the adoption of technology behind? I won't go to say the technologies behind because this industry invented a lot of firsts.”
The Nexus podcast (Apple | Spotify | YouTube | Other apps) is our chance to explore and learn with the brightest in our industry—together. The project is directly funded by listeners like you who have joined the Nexus Pro membership community.
You can join Nexus Pro to get a weekly-ish deep dive, access to the Nexus Vendor Landscape, and invites to exclusive events with a community of smart buildings nerds.
Episode 44 is a conversation with John Petze, Partner and Founder of SkyFoundry and Executive Director of Project Haystack.
We talked about John's illustrious career in smart buildings, which started long before that term was even coined. This is a fascinating history for those of you that weren't around back then.
And then we narrowed in on Tridium, SkyFoundry, and Project Haystack.
Mentions and Links
Project Haystack (0:58)
Mitco Mfg (2:34)
Andover Controls (4:03)
Echelon LonWorks (9:37)
Constellation Energy (15:31)
Cisco Mediator (17:00)
Brian Frank (18:22)
Talisen Technologies (20:07)
CBRE ESI (38:13)
GSA project (38:42)
Stephen Frank, NREL (1:04:14)
You can find John Petze on LinkedIn.
Thoughts, comments, reactions? Let us know in the comments.
John’s career history (2:02)
Why technology adoption is so far behind: split everything (21:02)
The history of Tridium (24:21)
SkyFoundry: how and why SkySpark came to be, and where it’s going (35:27)
The shiny object problem with graphics (58:24)
The past, present, and future of Project Haystack (1:01:05)
Note: transcript was created using an imperfect machine learning tool and lightly edited by a human (so you can get the gist). Please forgive errors!
James Dice: [00:00:03] hello friends, welcome to the nexus podcast. I'm your host James dice each week. I fire questions that the leaders of the smart buildings industry to try to figure out where we're headed and how we can get there faster without all the marketing fluff. I'm pushing my learning to the limit. And I'm so glad to have you here following along.
This episode of the podcast is brought to you by nexus pro nexus pro is an annual or monthly subscription where members get exclusive writing podcasts and invites to members only zoom gatherings. You can find info on how to join and support the Without further ado, please enjoy this episode, the nexus podcast. Episode 44 is a conversation with John petsy partner and founder of sky Foundry and executive director of project haystack. We talked about John's illustrious career in smart buildings, which started long before that term was even coined. This is a fascinating history for those of you that weren't around back then.
And then we narrowed in on tritium sky Foundry and project haystack without further ado, please enjoy it. Next is podcast episode 44. All right, John, welcome to the show. Can you introduce yourself?
John Petze: [00:01:19] Yeah. Hi, John petsy, I'm known as a co-founder and partner of a sky Foundry, the makers of sky spark, but, uh, I've had numerous, physicians over the years in the building automation industry. So happy, happy to be with you. And I do want to open with a comment that what you're doing is really valuable for the industry and for bringing new people in.
So I really applaud you for what you've been doing with nexus lab.
James Dice: [00:01:43] Thank you. Thank you. I appreciate that. And thank you for being one of the earliest members of the community. Appreciate that. so speaking of being early to the community, uh, I want to talk about your career history. So a lot of this, I didn't realize until you and I were emailing back and forth, so, uh, let's, jump through it.
So obviously we're going to get to sky spark where you started, but take us back to the beginning if you would.
John Petze: [00:02:08] Yeah. Okay. So, you know, I'm an engineering school graduate and I'm an industrial engineer, from WPI and, uh, you know, I was. To make things right? Plant equipment, metal, fabrication, welding, make products, get them in boxes, get them out the door.
I mean, that was my training and education. And, um, I ended up in the HVC industry from one of my early jobs, with a company who made HPAC equipment. The company's called Midco. And what's important about Midco is this was the original inventor of variable air volume technology held the original patent issued in 1962 or 1963.
He was a, immigrant who, uh, believe hungry. If I remember and got across the iron curtain when he was 14 years old and came to the U S he was quite a guy, brilliant guy. also his name is DeMita Gord chef. He was also, um, A unique person let's put it. Okay. But, um, you know, so I'm in there, you know, running manufacturing, engineering, operations manager, and you know, how do you build VAV boxes, um, pre production for reliability.
And then we get into VAV air handlers and, zero leakage dampers. He had a very famous ad where he was floating in his damper on a Lake and ad was, let's see the CEO of Honeywell do this. Right. So you're super, but this is interesting. Yeah. We're focused on energy efficiency behind the iron curtain, energy, you know, energy mattered.
It was a diehard theme of energy efficiency. That's why he invented VAV and zero leakage dampers. And these air handlers pioneered a whole new type of VAV air handler that was more efficient than any other ones. So I'm a, um, you know, I'm operations manager running the production facility where we have, you know, union employees and we're making stuff with metal.
Wow. Okay. So that's where I get started. And, um, we won't go into all the details, but that, you know, That's fast. Thanks, transition there. And, uh, I ended up getting, taken an interview at Andover controls, right. Um, you know, through head on and went out there and I had an interview and, you know, I didn't really know what they did.
The only exposure I had to controls at Midco was mounting Robert, Sean pneumatic actuators on the VAV boxes. And then one day, these really weird linear screw actuators showed up from and we didn't have any power supplies and all these things up on the production line. Right. And the point of this is, you know, I've had kind of a unique opportunity to create a, be involved in the earliest days of applying digital technology to HPAC assistance.
You know, not because I planned it, it happens right. Yeah. So I go to handover and I get the long story short. I got hired as, um, the very first application engineer they had. So when I got there, the engineers were still answering the phone from customers. Okay. And they gave me the job and they gave me a desk and they gave me a yellow pad of paper and a pencil and a Brown phone.
And they said, now you answer the phone from customers. Okay. And they, and then they said, and we know you don't know anything, kid, no problem. You write it down. You come to us and we'll, we'll tell you the answer. Right. And then the guy, the VP looked at me and it was. Semi a smile, maybe partly a snare. And he said, one time, John, meaning we're giving you the it's a one time.
You don't get to come back at them. You don't get to come.
Right. And it was originally the sun keeper. I was programming song keepers and supporting them. But then it was the AC two 56, which really put Andover on the map anyways. Um, so now I'm doing controls with computers, right? And the only exposure I had to computer in college was, uh, the mainframe in the basement of the library where you had to go with your Fortran punch cards.
Okay. And I was convinced that if that's what it took to program computers, they would never be used. We're going to work because I couldn't type. Yeah, anyways, I left and over to go to their biggest dealer at the time on the West coast, coast, calmer mechanical. And this is interesting. Now I went into the field and we did a really unique project.
It was over 350 bank branches for security Pacific bank. The longer in existence in California with the small ACA and 300 baud modems and a TRS 82 program, 350. And you had manually type these things out right. All over California and stuff. So, you know, the type of thing of being in a basement of a building on a white bucket program and control systems that happens that's reality.
That's real. Okay.
James Dice: [00:06:34] And what year would this have been?
John Petze: [00:06:35] This two 84. Okay. Maybe four.
James Dice: [00:06:40] So I was just wondering, because before I was born, we haven't gotten
John Petze: [00:06:45] one yet. Okay, go ahead. The math will get tough here, but anyways, so that was a big handlebar dealer. And I left there. I went back to Andover, so it's my second time back.
And I ran what was now technical support and documentation and repair and all that type of stuff. Um, And then I left there, in 86 to join telephone and, joined Teletrac, you know, where they think about Teletrac all that company was owned by. Dean came in the inventor of the segway, um, the inventor original inventor of automated infusion pumps for delivery of drugs.
Um, you know, self-made millionaire, incredibly unique guy. One of the smartest people I ever encountered in my life, absolutely brilliant guy, but he had wandered into this, building automation thing. Let's just say it wasn't his ballywick. Um, she hired a team, small team, and I came up to run the company and we got some engineers and there was some already there and they were computer.
And computer science guys, they weren't controls guys. And they, the unique thing that evolved out of that, you know, it's kind of like get people together for a band and you don't want to necessarily musical come out, right? This wasn't the original plan, but what came out of it was these guys, new standard computer and PC technology.
And we built the first control system based on true standard computer industry, hardboard hardware, and IBM XT motherboard with a real-time operating system. Right. And which then became an 80 and you know, and the 3d six 46, five 86, the whole thing, and a standard programming language. We didn't invent when we use C.
And built libraries. So we had standard based product standard base network, and we had high-speed multi-user network and a standard programming language. Wow. What's a very different product line in the marketplace. And the company was, you know, it was reasonably successful and, you know, we had lots of deployments, Sydney opera house did was all telephone calls, like unbelievable, right.
For a small company out that was in New Hampshire. So
James Dice: [00:08:46] anyways, Stumbled upon a tele trial controlled building. What happened
John Petze: [00:08:51] to them? What, what, okay, so they, um, they finally got acquired by Phillips and merged with Cylon and, uh, so they still are there, you know, there were, a couple of generation, the product generations, I was able for we'll call the integrator line and then they moved on and they did other things.
And so that company got absorbed into Phillips and stuff. So, you know, it didn't go away. It just got acquired and merged and that type of thing. Um, but anyways, I left there after a while and I went to a company called solutions, direct a two person startup, based out of Atlanta, Georgia. We were the very first authorized.
Ashland lawn works distributor in the U S they had set a couple up internationally. We're the first one in the us. We sold lawn gear, but more than echelon products, we accumulated, I believe the very first product line from all these companies who were building a lawn based sensor, a lawn based this along, but, you know, nobody had a lawn base full automation system, but there are a lot of people who had launched lawn based things, and we pulled it all together.
And in January of 97, we had a store on the internet. We could buy stuff. That's pretty early. Yeah. Okay. Yeah. Coding the HTML on, uh, on the website and all that type of stuff. Um, can you give
James Dice: [00:10:12] people text on lawn? I think there are some beginners that may not have even some on that yet.
John Petze: [00:10:19] So can you created a Silicon Valley company created a control system network based on real advanced networking technology, kind of a, you know, like.
Ethernet like, and whatever different boot course and offered that to the market as a commercial product that was going to standardize communications and control systems to get away from this highly fragmented, highly proprietary, you know, backnet was still a gleam. It was just emerging at the time.
Right. So, so there was a huge back that there was huge war protocol, Wars of the nineties. It was BACnet and lawn, which camp are you in? James? You know, which was, you know, and, um, I was
James Dice: [00:10:58] still, it
John Petze: [00:10:59] wasn't in the cap yet right now. Yeah. Um, but you know, I had always believed in standards, so it was just a kind of a thing.
And that's why the whole idea of standard technology InteleTravel and then standard networking. And so, you know, it was. Big proponent of it at the time. And anyway, so, um, one back that won the war back then. Absolutely. Backnet won the war. I don't think that there's still law and systems out there. You don't hear that much about it.
And you know, there's a lot of factors. I think the highly, um, commercial approach that Exelon took, They really put a lot of restrictions on the companies who are gonna use their technology, um, and made it difficult. well, you know, I went back to Andover and we implemented some lawn products, but the big challenge was you couldn't make them use a programmable easily.
Hmm. Okay. They really had a mindset of fixed function, configurable terminal human mentality. Right. Well, great. So now I can do just a little stuff. Right. And I think that was in my opinion, that was a big barrier to their adoption. So anyways, um, but then I go back to Andover and you know, one of the things I'm very proud of is I am tied for the record as a three-time repeat offender, no one will ever beat it.
Um, so there were three of us who were three-time repeat offenders. So anyway, a lot of people leaving and going back, it sounds like, yeah. Yeah. So I ran the systems group there and then I took over as VP of product development and during the final stages of continuum, and now this web technology is emerging ransom, you know, development projects on and, and stuff.
And I go to a trade show in the Hague, a lawn work show because we were showing off our lawn work stuff. And I met these guys who was showing me this box it's Brown box. That's all. Web-based right. Okay. Let me Cherie, Frank and go Rockwell up. Tridium okay. And one of them, they were there because they will have one of the only independent integrations to LonWorks networking that wasn't based on technology, you know, buying stuff from echelon.
They had network management tools that you needed with this network and all that, that was very unique for someone to have independently developed one. Um, and they supported back then. Okay. So both right. Anyways, I was really impressed and, you know, brought back the thoughts with, web based stuff, and all of that.
And, and, uh, months later I got a call from Jerry, why don't you come on, join us. And I'm like, Um, kinda like, you know, busy here, your product development and over controls, you know? Nah, you got to come down and talk to us, someone done, I joined them and they were local to you, right? No, they were, you know, Andover's up in Massachusetts, they were down in Richmond, Virginia.
So I went down there and moved the whole, family, pulled the plug, did all that went down and joined them. There were about 33 people when I joined and they were just getting ready to close that first, uh, series a VC round. Got it. And I came in into Tridium and, uh, when I resigned and, uh, you know, the CEO.
Andover had been really good to me and I resigned, I was a fan, a little painful and Hey, what are you going to do? Well, we're going to make a software platform. We're going to sell it to control. Yeah. And then John, I hate to see you throw your career away, man. I mean, you're, you're a really bright young guy.
I mean, no one will ever buy software from another company a thousand. Yeah. Seriously, that, that, that actually happened. That's true. And, uh, so that was a great challenge, right. Anyways, what they are spending, key years there, uh, left, you know, joined him in 2000 left in 2006 after they'd been acquired by Honeywell.
Okay. So I stayed on a while after that and then left and, uh, mainly cause I got recruited to go join a, another company called purveyors. Okay. Security technology, user authentication, based on biometrics that worked with your computer, worked with your door locks could work with cars. it was really.
Forward thinking stuff. Um, but the security industry's very different. So we'll leave it at that for the moment we haven't lost a lot of time. Uh, my joke, I think to you in the email was yeah, security technology danger will run. I had to add to Google that, anyways, that has spent, um, fears there, that company didn't make it.
In fact, some of the patents got sold to Apple because the whole idea of authenticating on an intelligent device, they had good patents on that and all that stuff anyways. Um, so I went independent for a while after that. Um, and, I was a consultant to constellation energy on their real time demand response, end to end platform.
And what end to end was key there, they were in the deregulated side of the market. They ran the worlds of the U S has largest trading desk for demand and, you know, and consumption. They had a trading desk on the third floor, which was like wall street. Right. You know, and. The whole goal was they could arbitrage demand response with where they could sell it.
And so I went in as a systems design consultant, along with Onno Shelton, and we designed the architecture and system and requirements. And then they had an internal and external team of people build that to the point that they would, recruit, people that joined the program. It was called virtual Y and they would, uh, dispatch automatically instead of calling you on the phone and saying, Hey, why don't you shut something down?
You've agreed that I can do things to your buildings and you get paid for it. But of course they were arbitration on the back end and they right. They're making money off. This was a working system business model for them. I'm not sure of the status of it today. It was called virtual. What? That is a 90.
no, rather, let's see. Where are we now? Uh, 2000 and. Nine. Okay. 2009, right? Uh, Provera Sunday is 2009. Okay. And so I'm consulting to them, but not full-time and doing some other gigs. And I'm trying to get a, someone introduced me to people at Cisco and I'm hoping to get a consulting gig there. They bought a product line.
They're going to get into smart and connected buildings. They need a consultant. Great. And instead they flew out, met me and said, well, we don't want a consultant one higher.
James Dice: [00:17:13] Okay. And this is like the mediator of that
John Petze: [00:17:15] when they purchased immediate is the product they hadn't bought in. But they're the.
Initiative was bigger than just the media. It was called smarter connected buildings. Right. Um, which was, under, smart and connected infrastructure, you know, so there was smart cities, smart buildings, whole idea, which I thought was compelling was that control automation data. It'll be just part of the fabric of the network.
These everything has to be network. So, and you know, you've got, have you do the processing power and all the network environment? Why not? It's control plane as well. Hmm. Okay. And also that they, you know, we're going to reinvent this Byzantine backwards industry and completely change it overnight. Like they had changed the phone industry with IP phones and they'd been so successful with that, that in my opinion, that was a great blind spot for them.
They assumed, every industry could be reinvented like that in the same way that didn't happen. As I think we all know. Yeah. Okay. All right. And, um, so, I left the air in between constellation and Cisco, Brian and Frank, and I get back together. He had now left tritium. I'd left no six. He lived in a way we'd get together, start talking.
Hey, what are you doing? And then that's you know, we worked on a business plan together. What's next? Okay. Okay. So we started it before I went to Cisco, but then I was dedicated to that. I left Cisco. I came back to him and he says, come on down. I've been building software, building stuff. He's been building stuff based on the early stuff that we had thought about and everything.
And I really hadn't been talking to him when I was at Cisco and that was guy found me. And it's like, I get out and talk to him. He goes, this is like, look at this data stuff, data on data. What do you think? Wow, it looks really interesting. I think we can pay the mortgages. Wow. Let's give it a shot. And so we get started in, um, the company was formed in a way Oh nine, but we went to market with the first product, the fall of 2010, the first version of sky spark.
and you know, so we're certainly early in data and analytics. We always have to give credit to Jim Lee who was doing it, out of symmetrics. And I don't think anybody, I certainly didn't really under Stan what he was doing. So we got into this and, uh, Brian know, created a software platform, that is, you know, the foundation of sky spark and.
And that's where we are today. Yeah. And, uh, we've been out in the market for 11 years now. Yeah.
James Dice: [00:19:43] And we finally just got to where, uh, young James just graduated from his last semester of college. So, uh, that was, that whole story was like right before my career began. So just to catch everyone up on you and I, so I started using sky's park when I was, uh, say two, three years out of school.
So they were, um, Thompson technologies was in St. Louis. I was in St. Louis and they were a reseller and, you know, sky sport was kind of powering the backend of their platform. And so that was when I started using sky's park. Um, and then obviously we started working together once I became a reseller a couple of years after that.
And, uh, yeah,
John Petze: [00:20:25] I didn't know the early part of that story, honestly. Yeah. Yeah. There
James Dice: [00:20:28] were, um, Talisen was providing the software and we were using it. Um, I was at a mechanical contractor and we were using their software and a couple of different,
John Petze: [00:20:36] uh, big deployments. Yeah. It's a tangled web, huh?
James Dice: [00:20:42] Yes. Yes it is.
so I want to dig into, uh, Tridium and sky sparks specifically, but let's start with, uh, given all of that history that you know about I mean, you've mentioned different, different facets of it, hardware versus software, uh, the trouble Cisco had, sort of like changing the industries.
Why do you think this industry is behind from a technology standpoint?
John Petze: [00:21:06] Okay. Yeah. So I know that's your favorite question. Okay. And you want me to ask it? Yeah, no, no, I do. I do, but I I'd reword it. Why is the adoption of technology behind? I won't go to say the technologies behind this industry just invented a lot of first, so I can give you some examples, but, and the first thing I'd say is the word split, split financial incentives, split responsibilities, split buying centers, split domain knowledge, right?
There's so much complexity of different systems in a building. This. Desire to compare them to a car I completely disagree with. Right. Totally. Okay. The only car you might compare with is, you know, a custom built car you see on, you know, motor trend TV, where they go to the junk out and they weld all this stuff together and put, okay.
Cars are mass produced. Tightly coupled buildings are bespoke, loosely coupled problem. The other thing is the long capital cycle, but dominates the majority of the investment in a building. Yeah, right. You put stuff in a building and a short capitalist cycle, but 95% of it as long capital cycle, that's how it's financed.
That's how it's thought of, right. Those are the things that cause slower adoption. It's frustrating. Um, I think there's been belief throughout the years that, well, a technology solution can break this apart. I don't believe that's true. because of those, those reasons are in there, we're all there.
And they're just part of the reality of the built environment. And I think we'd be better off instead of wishing that some magic technology will make it go away to understand where are the pain points? Where are the bottlenecks? How can we. Ease them and improve this. I see this as an incremental something we can incrementally improve, not something we can snap our fingers and we'll go away.
And yesterday was this. And so I firmly believe that, so anyways, yeah. I like the split.
James Dice: [00:23:08] the word it's split. I used the word silos a lot. I also been using the word handovers a lot. There are just so many different handover points where yeah. Especially from a digitization standpoint, just stuff just gets lost.
and you know, that adds up over and over and over again to yeah. Slower adoption. I think, I think maybe I'll change my question. I've been debating on whether to just retire the question or just modify it a little bit, but I still love the answers that I'm getting. So,
John Petze: [00:23:35] yeah. Well, that's why I wouldn't give up on the question because I think it's helped informing the audience to think that this is maybe some simple problem, but you know, we're just all dense and we haven't figured it out.
Right. And that's not helping anybody. We need to understand the reality of the problem if we're going to address different pieces. But your idea of handoffs, if the handoffs are between the split areas of responsibility, I'm responsible to design it, you got to build it, operate it. You've got to optimize it and they need to improve the energy efficiency.
Hey, my job was done. I designed it right. And they are, they're gone. They're not even around typically. So I think that's a really important perspective. We all need to keep in mind as we keep improving, you know, digitizing and hopefully improving the built environment. Yeah, totally.
James Dice: [00:24:21] So let's talk about Tridium.
So, given when I started my career, I think, you know, first heard about Jason's like pretty early on and the, the concept was, you know, get data into the analytics, like that's, and that's what I've been doing for a long time, but haven't really ever had like a history lesson around, you know, the early days of Tridium.
So before Honeywell acquired it, so you hinted at it a little bit, we're, you know, we're going to go sell software to all these controls companies. Can you, can you just give us a history lesson? Yeah.
John Petze: [00:24:52] Yeah. So, you know, they started in the late nineties and in the basement of Jerry Frank's house with, uh, Brian, Frank and later Andy Frank, who was a little younger and, uh, John Soublet and, you know, the, the key concepts were a number one move to.
IP and web technologies. And now this is really early to do that. In fact, the web technologies couldn't do everything you wanted. It couldn't completely Emily it stay there, windows base building automation system, but they had a belief that it would. And the other thing was unifying protocols. Being able to bring them all the different protocols together so that they work together.
So drivers too different protocols in one box, right? With a unified data model. To do control, though. It was primarily about doing control, right. And serving results in a standard browser. So your UI, you know, there wasn't a more fat client, you used a browser and, you know, in the very early days where you couldn't do, you know, uh, sophisticated things, spinning fans and all that as the fancy windows apps, but that caught up, right?
Because look at what the it industry, the investment in the web and web technologies. And so there was a Niagara version framework, version one, hardly anybody ever worked with that. And then there was what's called R two, which was widely deployed. And then there was a ax and now Niagara four.
But the other thing about the business model was. They really believed they were, their intent was to sell this software platform to the major OEMs, right? Not to, we're going to, stand up a company to compete directly with Honeywell or Johnson or seamless or whatever. It was intended to be a software platform sale.
And of course, you know, I made that earlier comment. That was odd. That was different. That was not well understood and accepted, but it did happen. Every major company has a product line based on Niagara, right? Whether it's Honeywell or Siemens or Johnson, You know, Schneider, they all ended up with products based on it.
Maybe not every product they sell, but it became a universal platform for connecting to diverse systems solving that backnet lawn works, mod bus and hundreds of other proprietary protocols, right? I mean, literally hundreds so that people could connect to the reality of what they own. And remember, you know, we may talk about the challenges of the new standard protocols and the new generation of API APIs, but we'll back 20 years.
It was proprietary protocols every way it turned ours 45 networks. And, you know, that means it's a serial bus. That means you've got hardware involved. You know, it's not just clean IP protocol conversion. So you had to have physical hardware to do that. Right. You had to have the bullet level stuff. So, but it became successful because it solved problems for the system integrators and contractors who even if they were dedicated to.
Pick a brand they're going to sell a project. There's other stuff in the building. Even if it isn't a Honeywell compete with Johnson, there's meters and there's this. And there's that the idea that one company could build everything that we're going to every project and it would all be one protocol that's flawed.
Right, right, right. But none of those majors could see it in their own interests to embrace everybody else's system. Uh, that was key. Right. And we used this, you used to say when we were talking in Solon and OEM, so you say your engineering assumes just as smart as, as ours, they could do this except you can't bring yourself to let them right.
Integrate all these. And the other thing is if you invest all the time and a platform where there's this huge amount of engineering to support multiple protocols, You'll only be able to amortize that over your sales, we build in a platform that we get to amortize it over everybody who uses it.
You know, you write a driver and it gets used a hundred or a thousand times. If you read a driver, you get to use it twice. How do you amortize that? So it became an economic model. So the technology is important. The strategy is important, but the economic model was a reality that this is a horizontal problem to solve.
Yeah. That's what treatment. And that's why even today it continues to have incredible momentum because it solves a horizontal problem. Hmm. Right. Um, that still exists today, even if less and less are all proprietary forty-five networks. There's still, you know, we used to joke. We love protocols. Yeah. All right.
Bring on a new standard. Definitely makes it more reason you need our system. Yeah.
James Dice: [00:29:21] That only makes the flywheel accelerate faster. Talk to me about like, was there a system integrator before this? Because like what tools they use to do it or was it a bunch
John Petze: [00:29:32] of contractors? No, there were no, this, this, you know, they weren't the first people that identified the problem.
It was the first time it got commercially solved in a big way at telephone. You know, I told you, we believed in standards. We built a tool set called the open protocol bus. You system integrator could write drivers to other systems. We didn't have the team that we were going to start writing drivers and selling them.
Whereas Tridium did focus and invest in that more capital to do it, but we had the tool. And if you were a programmer. You could write and people wrote drivers. And one of the big integrations we did to tell a troll was between ASI controls, terminal unit controls, which were a major breakthrough, you know, based on custom Silicon, you know, terminal unit controls, right.
And they had a protocol and they were getting tremendous pickup by major property developers because they dramatically lowered the cost of putting DDC down to the zone. You know what? They didn't have anything beyond the zone level. So tell the troll, we did an integration with them and did some of the biggest skyscrapers in the U S for Heinz, because we had the.
Big controllers and almost no terminal. And they had all the terminal unit and no be controllers. And we did that and it looked like one system. So this was being solved. Okay. But Tridium saw the vision of a horizontal platform to do it in the business model and had the re you know, the resources brought together to make that a product.
So when somebody said, well, I got a protocol. We didn't, you know, they didn't have some, well, go try to write it yourself with this kit, if you know your way around programming, it was like, okay, we'll write it. You know? And they built and built and built this hundreds of protocols supported, you know? So what
James Dice: [00:31:14] was the acquisition like, look obviously they built the business and you joined it to sell software to the OEMs.
And then one of the OEMs comes and buys like,
John Petze: [00:31:24] can you talk about them? We was shocked. Yeah. So I'll put a little, context, just to that, to help. I mean, you know, they industry had some major downturns there, right. This was the first age of the web, you know, and, and so treat himself there and we've got some great projects.
One of our big customers was Enron energy services. They were building but they called it a negawatt utility where they could do demand response. And they had one of the other big trading desks, so they could, okay. All right. And they were big customer B is we gave them technology a Tridium, they could do this, right.
Remote telemetry, command and control interface, your systems, you know, control, demand and all that type of stuff. But remember what happened in 2001, right? Yeah. Nine 11 and Ron the scandal. Right. And then collapse of, you know, the over hype web. Right. So trims was moving forward, but, you know, Now the investment community is like vanish from the face of Europe and we needed, um, additional funding to continue on with the plan.
Anyways, the capital structure and control changed, and it put other people other than the founders, you know, with majority ownership and they, after a number of years, when we became successful and profitable, they decided they wanted to exit manage. Your team, wanted to do a buyout. We tried, we tried that, but in the end, accompany, you know, Came calling and Honeywell.
So now Honeywell's a customer, Honeywell webs was based on Tridium Niagara. It'd was becoming incredibly successful. And, they pursued, uh, the acquisition process and we got acquired by Honeywell. We do have a thought that one of our customers could acquire us because well, Honeywell is still sell the Johnson and Siemens and Schneider, and they did, they left that alone.
But then the other question is, will those companies continue to buy. Right. And you know, when interesting thing is that I learned through that processes, hang on, Johnson, sell stuff to each other all the time. Okay. Now this happens. Nobody ever wants to talk about it. This happened. And that was able to be navigated.
So the company continued on with what it's viewed as the independent, but all of the brands that were selling it all. So the Johnson FX became a incredibly successful product and on and on and on. So, that's a little bit of perspective on that. And I left a while. Uh, you know, as I said, after Honeywell acquired it to do the security company, so, okay.
James Dice: [00:33:50] Hey guys, just another quick note from our sponsor nexus labs. And then we'll get back to the show. This episode is brought to you by nexus foundations, our introductory course on the smart buildings industry. If you're new to the industry, this course is for you. If you're an industry vet, but want to understand how technology is changing things.
This course is also for you. The alumni are raving about the content, which they say pulls it all together, and they also love getting to meet the other students on the weekly zoom calls and in the private chat room, you can find out more about the email@example.com lab. Start online. All right, back to the interview. All right. Last, last thing on tritium. What do you think about it today? 15 years after you've left? Something like that? is it the same place you thought it'd be or is it different?
John Petze: [00:34:34] Well, you know, when big companies acquire small companies, things change, the culture changes and things like that.
So I think, I think you see that in us market overall, Big companies have trouble innovating. That's why startups do it and startups get acquired, right? So there's a lifecycle to rapid innovation and, you know, pioneering innovation that big companies have a hard time, you know, with funding innovation that doesn't have, a really good calculate ROI, Hey, we're going to build this.
We don't know is going to be successful. No, you're not right. So, that's why, so, um, you know, I think it's created opportunity, but it's still the most adopted platform out there for doing control and, the major data source for, uh, companies like ours who are taking the data to create value.
James Dice: [00:35:25] Yep.
John Petze: [00:35:25] Cool.
James Dice: [00:35:27] Let's talk about sky Foundry now. so go back to the, you, you came back from Cisco, you and Brian got together in your basement.
John Petze: [00:35:36] Yeah, we started before that, after the, um, security company, you know, wasn't going to make it and had to be refactored. You know, I told the board one of the first people to leave should be me.
Right. You've got an asset here that you can work with, but in a different, totally different way. And I left. And so I'm independent, Brian, I get together, you know, we'd worked obviously closely through the years of a Tridium and the success there, although in totally different roles, right? I mean, Brian's running the brain trust up on the fourth floor and I'm running the business, not on the third floor and people would be surprised how little we actually talked during those years.
Um, but anyways, We got together and started talking about this stuff. And, you know, I've mentioned that, our original business plan, we look carefully at what's happening with the cloud, Amazon web services and all that stuff. And you know, this whole idea of maybe, maybe all the BAS applications will move to the cloud.
Right? So we, we had thoughts around that. We also started looking at this data, and in any, it just kind of emerged in an interesting word, you know, because that's what we say. Patterns emerge from data, with analytics. Well, these ideas around, Hey, there's something that could be done here with the data, but isn't trying to displace a controls company.
Nobody's doing it. Nobody's truly working with the data other than putting it up on a screen and making fans spin. Right. and, uh, that's another point we can get into a personal note of, you know, I, I believe color graphics are the tail that's been wagging the dog, a BAS sincerely nineties.
And there's been one of the greatest misuses of resources in mankind history to come back to me, come back. but anyways, so we saw this opportunity and, you know, Brian more often, you know, with the deeper thought about how to work with data while I was paying my bills consulting and stuff like that.
But it was, he started building stuff around that while I'm off at Cisco. And then we come back and he says, you know, we've got something. And I already got a couple of people and you know, you're going to say, I don't want to be answering the questions on the phone and getting on the plane and going out to meet the potential customers come on down.
And we started, when I left Cisco, I had not yet called Brian. I made a decision to leave. For my own reasons. And then I call and say, Hey, by the way, and he's like, get down here, you know, because I didn't know all he'd been doing, but he, he was of course doing stealth mode building. And so we get going and with version one, I mean, we, uh, secured Kohl's department store, you know, nationwide, huge, huge project still running to this day in one of our leading system integrators, ESI, who is now a division of a CBRA.
Right. And, um, you know, it quickly showed people how they can get more value out of their systems by working with the data, right. To identify issues, problems, patterns, you know, the whole series of benefits you get by using data to generate value. And, uh, it went well from there. And then we had, um, we've been incredibly fortunate with the company, a great technology.
And, um, we. Got exposed to the GSA fast 50 project at the time and had the opportunity to present to consultants involved in that. And, we got approved to be part of that. And, our technology ended up being selected for the GSA fast 50, which we came GSA link, which based on any public, information we have available is the world's single largest analytics project for the built environment.
Now over a hundred of the largest GSA facilities over 55 million square feet. Um, and that, you know, that was critically important for us, both as a sale, but also as validation and, uh, proof points. you know, the GSA is wasn't is very serious about, running their buildings, um, efficiently.
James Dice: [00:39:25] And I, when I was at unreal, I had a chance to work on a couple of those buildings a little bit. Okay. Yeah. before we dive into that, what were the other options you mentioned. Jim at Semetrics. like? What was analytics at the time?
John Petze: [00:39:38] I only know that Jim Lee at symmetrics was doing it.
I'm not aware anybody else was. Um, the story I tell about that is the first year I went to, um, you know, real calm even before they had the IB con went to real calm and we got a 10 foot booth and you know, I'm known because I've been around with three, Hey, John, what are you doing now? Well, we're doing data and I'd explained to people and they'd go, what are you doing?
Tell me, explain that again seriously. Right. Because nobody was doing it. you know, I think the point here is about society in general is learning how to use data. This is a trend across every industry. Right, right, right. Learning how to use data to generate value. And in our industry, one of the statements I make is that, you know, the data.
The device produces is more valuable than the device and people go, Whoa, wait a minute. No. Okay. I got a VAV box. What did it cost you? 500 bucks. Run it wrong for a year. Okay. Well, certainly can apply to a chiller. What did that cost you? 30,000. Okay. We're not wrong for you, right? Right, right. you know, actually say it to make people, you know, basically react because this is the point is the data is critical to everything.
Right. you know, I like to make the analogies. I mean, you know, how do you think they know when they optimal time to put bathing suits on the racks are in a department store? You think somebody is gone? Maybe this one? I don't think so. Right. They using data. Right. And what type in what area for what demographic.
See analytics. Well, why don't we have the same tools in buildings, right. Um, but I don't believe there were a lot of options. Somebody may come forward and say they were doing it too. You know, from what we saw that, uh, and our big difference was we weren't doing it as a service in line with, uh, you know, our past and our mindset and Brian Frank's development, capabilities.
We're going to build a platform to do it that anybody could use. We weren't gonna we'll offer it as a service. You get to come to us. It was we're going to build the platform. So the entire world of system integrators and engineering firms, they can implement analytics for their customers, a software platform, and that's a different mindset, right?
You're going to make shrink wrap commercial off the shelf software. That's different than saying I'm going to build some software. I get it in the back room that I can use and glue together to do project work. Right. It's a whole different mindset development requirements, et cetera. W where did that mindset
James Dice: [00:42:12] come in?
You're saying that's those developed at
John Petze: [00:42:15] tritium we're we're product guys. I'm a product guy, except the stuff I went to school to build was made out of metal. Right? Yeah. And Brian, that's his mindset on software is he builds products. Totally. Right. Okay.
James Dice: [00:42:29] And so what was the, like the decision at the time like, cause sky spark is a platform for people like me, the nerds and the people that want to set things up for their clients usually.
Right, right. What was the decision at the time to go that way? Like, why did you do that versus a more sort of off the shelf? Like you said, SAS
John Petze: [00:42:48] product. Oh, well, instead of saying, yeah, well, so there were two reasons for that one, you know, we were bootstrapped. And, um, you need a whole different financial backing.
If you're going to, you know, go direct to end customers nationwide, worldwide, Salesforce has and project delivery capability. And I had mentioned, I ran the sizable system integration group. There were three of them and ended up at controls before I took over product development. and being a system integrator and doing project work, as you know, is a different business than building products.
You have different accounting systems, different requirements. I mean, they are night and day. Right. And I didn't enjoy it. Okay. I realized, no, I'm not a contractor. Okay. I'm not, I'm a product guy. That's what I went to school for. It. It's funny, my best friend in life. He became a contract he's unbelievably successful, but I would never do what he did for the last 30 years of his life.
Right? Five o'clock in the, in am, in the truck. No, I know mine don't want to do that. You know, contracting is 24 hours a day, really tough. Right. It's a very tough, but it's, you know, it fits a certain type of people, you know, there's certain types of people. So we knew we weren't going to do that. And the whole SAS thing, that's a separate question.
Um, when I joined Brian went with the original business, or in fact, before I went to market back in the consulting stage, we took different business plans out to people. We knew, you know, mutual respect and talk to them. And as much as this cloud stuff was really interesting. We came away with the belief that if we are cloud only that's key.
We can say goodbye to 50% of the market. you can say what you want. There are companies who can't move their data to the cloud, be dependent on a cloud, whatever, but that's not to say cloud is bad. And in fact, our platform is extensively used by service providers who use it as a cloud platform, but we built a platform for that.
That could be used the cloud platform, or it could be deployed on premise. Again, it's a standard product. The third party, the question is how we decided to go to market. So, you know, early days of Andover or, you know, I got towed around by the sales guys as they were building a distribution network. I didn't even know what that meant.
All I knew was John, you come home with me, we're going, uh, you know, Oshkosh. Right? Okay. We're going to meet a system integrator. In fact, they weren't even called it. We're going to meet a dealer, a dealer of what they have. Okay. Okay. Yeah. No, a dealer they sell all. Okay. And you can drag around and learn. So I learned the ropes of distribution.
Well, I was there and then it tele troll. I built one, right? Yeah. Led building on one of my own. And then I went back to Andover and I worked with them and it prevented us. We went through a distribution channel. So I've had 30 plus years of understanding. What's the value chain that puts technology in buildings, building owners.
Don't buy a lot and install a lot of stuff on their own. Yeah. So we just, it, to us, it was obvious. We're going to go through distribution one. We're not going to hire worldwide Salesforce. We will work with companies who are the worldwide sales force and number two, the implementation. I mean, that's the biggest part of the job.
Those companies specialize in that, right? So, you know, we work with reseller partners. Some of them are what you'd call a control system integrator, but others are especially engineering firms. There are a lot of their, what you would consider consulting engineering firms, who you use our software to deliver their services, their analysis, their commissioning, their continuous monitoring base commissioning, and then we have the, Service providers who are monitoring buildings as a service, right.
Using our software to monitor and analyze and dispatch or consult or recommend or all of those different types of things.
James Dice: [00:46:42] Okay. That makes, that makes perfect sense. Looking back,
John Petze: [00:46:46] um,
James Dice: [00:46:47] how big is sky Foundry today
John Petze: [00:46:49] now? So the metrics we, we focus on are the, external market metrics. So, you know, we have back in 2017, we validated that our software was used in over a billion square feet of facilities worldwide.
Um, over 15,000 facilities, it's now, you know, over 20,000 facilities best we can get. It's actually harder to get the day to be, as we have so many service providers, well monitoring buildings, and they are not, required to tell us, you know, you're right. They don't. So, um, so we have that and we have our 150 plus partners worldwide.
And with the software's installed in six continents around the world. So it's, you know, it's become very well adopted around the world. Our team is small. If that's your question, um, we're in a eight person company. Wow. So that's awesome. Yeah, it is, it is but, uh, invest a lot in building tools and backend and processes and automation everywhere, right?
Like kind of eat our own dog food on how you should address, you know, these data challenges and stuff. So,
James Dice: [00:47:54] all right. So where's the, where's the product going in the future? So, I mean, I'm going to assume that people understand the concept of collecting data into the platform, fall detection, you know, that kind of thing.
And What's the future of, fault detection and analytics and the future
John Petze: [00:48:09] of sky spark? Yeah, sure. So first of all, I want to break those out though in our mind, those are two different things. Okay. So we make a software platform for data analytics for the built environment, the types of devices, equipment sensors.
Okay. Fault detection is one use. I consider it a sub and important line of analytics, but there's other. Things you do with data, right. you know, it, isn't all about finding something broken. It's, you know, predictive, it's finding things degrading, right? That's not a fault. Right. It's knowing it's tracking KPIs.
How many service calls do I have on this brand of equipment versus this brand? I know what I'm buying for the next 20 stores, right? Yeah. How to different building styles perform. I've got three different designs in my portfolio, which one should I build next? Right. you know, it's working with data and the purposes and understanding of what we can do with data.
That's a journey. We're all learning more and more. So, you know, that example about tracking service calls against vendor of atria stapler that comes from one of our customers. We didn't not think of that off. Right. One of the other one is true story orientation of the sun. Okay. Okay. Is that a quote from a customer, right.
You know, you're supposed to do the consultative selling, right. So what would be helpful to you, James, to track? So you can compare your portfolio, you know, and energy. Can you, uh, do analysis around orientation to the sun? And I'm like why he goes, yeah. Tired of management complaining to me that all of these stores are using too much energy.
John, what is wrong with you in those stores? Well, you see the convenience stores. They have a glass front and all of those stores are facing due South stop torturing every month to tell you a hundred times time's over. Wow. Okay. Okay. I didn't come up with that insight. What is that? Data and analytics and a customer use case of how data could be meaningful to them.
You know, pretty well. So stories, he's fine. They're going to be able to filter them out. So you stopped getting, you know, the C-suite looking at them and going, ah, you're running those stores. They're officially, no they're cited wrong or cited inefficiently or whatever. You know? So
James Dice: [00:50:31] things I love about sky spark is like, I felt as someone who is serving a building owner, anything, they asked me.
I could then go back to my programmer and say, create this KPI or create this
John Petze: [00:50:43] yeah. The analysis correlation. Right. And that, and that's, again, that was our goal fault detection happens to be a key thing that people can get their head around. Right. One of the concerns we have is that that narrows people's view of why data is important, what you can do with it.
It isn't all just about, and in fact, that, and this is public knowledge. One of the things that came out of the GSA link project and studies they've done to prove it. They've had Carnegie Mellon to analyze the project and the funding and verify the financial returns they saved. I believe approximately 50% of their savings came from maintenance, repair, operations savings.
Okay. It wasn't just energy. Right. And the other side of that, and so you'd say, well, that means faults. No, it means knowing what a building needs before a truck goes there and gets there and said, well, I haven't got the stuff to fix it. I'll come back tomorrow with another truck room. And that's, that's another thing, but your, your question is you're where it's going.
And, um, I'll answer that with one of our most recent additions that you didn't notice and mentioned, which we appreciate, which was, you know, we built a complete workflow system in the product. Okay. Because you know, analytics finds patterns in the data that have value and meaning.
Right. And I remember being on stage with, uh, one of our, um, you know, one of the most notable guys in our industry is now retired. Paul law's wall that we were speaking at a real comp and him and I both on the panel. Right. And, uh, as you can tell, we, we like to joke a bit and stuff like that. And he says, hang on to your seat, John, because I'm going to, I'm going to really get the audience here.
And we're talking about analytics and he gets up and he says, well, I want to let everybody in the audit. It's no analytics doesn't save any money.
James Dice: [00:52:23] I can
John Petze: [00:52:23] picture that in the real copy. Yeah, it was great. Right. And if it goes dead silent, you know, so what did John and I doing here? Well, the point is it tells you what you need to address.
If you get analytical results that say, James, this is broken. It's not operating right. It's inefficient. And you do nothing about it. You will save no money. So the analytics starts a workflow process. Now we had lots of input over the years from our partners, big and small who are dealing with this, right?
Because we think, you know, I mean the early stages, Hey, I told you something to broken, go fix it. Right. But, you know, it's more complicated than that. And we took input for, um, this is a clear case of a product feature, taking more time to design and define than to code. We took input from multiple years on what people needed to do.
Right. We had people interfacing, you know, our software with our open APIs, with standard CMMS, you know, the maximum of the Archibus all of that stuff. And that can solve it for certain people, right. Automatically generate a work order. Right. And then it gets assigned and all that. But you know what? Those.
Systems are pretty heavyweight, right? And oftentimes not every operator can, you know, they, don't not, everybody gets to touch those things. Right. you're a technician. You may get your work assignment from it, but you never touch it. That's you know, with, with analytics, you're in there looking at something you.
The guy who knows the children and needs to create something. So anyways, we implemented a very comprehensive workflow tool set. We call arc meaning the arc of the life of an issue. And that's where the name came from, because that's what it has. It has a arc of both of the story and arc of the love. And we implemented that in the product and it's lightweight.
It's fast, it's easy. And while you're looking at a problem, you can create an arc. You could create a work order you could assign, or you could just flag it saying. What's going on here. Right? You don't have to know I'm going to assign it to this person and do dated, and this has these, they can just say, Whoa, capture this.
Right. So someone else will see it. Um, and that was really, I mean, it could virtually be a product on its own. It was really huge, uh, addition that we understood the problem, but we were working on how do you solve this again? Because people are using our software for so many different things, right? So we created the idea of a note, which is just a flag, something and writes on them.
We also created the idea of a ticket because one of the things you were working with Brian, Fixing software defects are typically all tickets and they actually progress differently than a work order where I got to get a wrench and a filter and a valve, and I got to replace it. Okay. So there there's a distinction.
And the other thing we did, because you know, everything's a program on Scottish where you can completely define your own work line full of documents, your own archive. So you're not fixed with what we did. We built like the faults, build your own with as many stages and many tags and made descriptors as many fields take your workflow and you can adapt our to exactly meet it.
Right. And that was always, a thought process in our, in our software. You know, people would say, well, once your data model that people have to employ data model, not data modeling technique, we've of course have found our haystack and we'd say, no, no. The purpose of sky sparks database technology is you should be able to model anything you have.
I don't know the model of your equipment systems. You do, you better be able to model it the way you want and our software and haystack tags, a lot of people to do that. And that became our fundamental methodology for doing semantic modeling. so anyways, that shows where we're going, it's really addressing more and more of the.
Process of using data to generate value. Right? And so we have a number of other things in the works that, uh, you know, the development schedule is always full with things. We'll be introducing another major app here. probably in the next 60 days, we'll make some noise about, because it brings in yet another use case that people want to do with their data and they want to do it in an easier, simpler way.
So we've kind of amplified it, right when the other things we have amplified, if you will, was a tariff engine. Right. Yep. So the idea that, you want to calculate energy costs, you want to relate energy savings. You need to make a calculation. And what are most people don't matter. It's 12 cents a KW you can put it a metric on things.
Hey, that's a big problem or a small one, but you know, energy rates don't work that way. Right? You could use more energy and pay less depending on your rate. Or you could use the same and pay less if you get on the right rate and all that. So this very, uh, kind of like non-linear factor on energy is the tower freight and their complexity.
So years ago we built the tariff engine so that you could define all the charges associated with your chair, freight distribution, transmission, peaks, and time of day, time of year taxes, service charges, block, and range, and on and on and on. We had people from Europe Oceana us contribute. So we can hopefully build a model that will work for everybody.
And then more recently we've built a gooey. So it's easy to enter the charges or perceptively, easier to enter the charge fees. We always told people, don't worry about the coding of the, rate. Your biggest problem is can you read the document you get from the utility? That's where that's where people most likely had to hire a consultant, right?
And then you had to enter it and then, you know, so now it's a point and click to enter it. But now people realize, yeah, it's really the reading the document. Right? So anyways, but that was another use case. We built into the software, the workflow, we've got another one coming in and we'll continue because this one thinks we say analytics is a journey for us, for our product, for customers, for society, you know, it's not going away.
James Dice: [00:58:22] uh, Let's circle back to your graphics. I want to hear you about your graphics
John Petze: [00:58:27] being the tail, wagging the dog. Well, ever worried about crows and how they build their nests? Nope. Okay.
They're very, um, they're very much like shiny objects. Okay. So the pick them up and embed them in their nests, certain birds like shiny objects. And so I call it the shiny object problem. Right? Okay. That, Oh, look what we can do. And yeah, it is cool. But one of the things that got me to have this opinion is when I saw being a system integrator and being involved in implementing controls, just pre analytics, more money was being spent in the bid to draw pictures of the equipment than to do good control sequences.
That's not right. That's not right. No. You know why? Because it's easier to do easier mentally to do that and to think about complex, sophisticated controls that are really optimized sophisticated systems. But no, we had spinning fans that was young Bitmap was for us with data that would change 74.2, then the fan spins.
Now it's 3d. Now it shaded from a 45 degree perspective, right? What value have we increased? Right? We've increased the sales value, Wiz bang sizzle. But I would argue, in some cases we've diverted our tension intellectual focus away from what matters, which is good, sophisticated control and control sequences.
Right? And so that was the point I was making about, you know, the graphics being the tail that wagged the dog, and then development companies who are building these products would end up putting more engineers on the next version of a graphics package. Then on doing more control algorithms that would run better.
James Dice: [01:00:16] Yeah. And I think that same thing's happening with SAS products today with digital twins, things like that, that are out there and. Yeah.
John Petze: [01:00:24] And it's not that they don't have value in our presentation. I'll say I need to look at pictures of my equipment from time to time. But I have 5,000 pieces of equipment that can't be how I glean information about what's happening right.
And wrong in my building today. It can't be it's impossible. Right. And that's the value where analytics comes in. So that was the point I was alluding to there. with my, uh, sarcastic continent, I guess. So love it.
James Dice: [01:00:50] Uh, I, I love those anecdotes commentary on the industry is just quite funny
John Petze: [01:00:56] sometimes.
Yeah. Well, it's human. There's a lot of human nature going on here. So the shiny object problem. That's, that's what I call it.
James Dice: [01:01:04] All right. Let's talk about haystack. So, this could be a topic that we take more time to unpack on it on a future episode, but let's talk about the founding story, at least, today
John Petze: [01:01:16] So you're talking about my evening job. Yes.
James Dice: [01:01:21] Yeah. tell me about your,
John Petze: [01:01:22] let me tell you about my other moonlighting for a while here. All right. So Brian and I get going, we're doing analytics and he comes to me, right? I'm the industry guy, the domain guy, Brian's the software guy, software architecture.
He comes to me. Hey, John, you know, one of the things evolving in analytics is data preparation, data categorization is stuff. So go find the industry standard naming convention so that we'll be able to quickly categorize the data coming from these systems. And I said, ah,
James Dice: [01:01:55] that question
John Petze: [01:01:56] is, Oh man, because you know, Brian's a software guy and this is, you know, this is, this is 2010 now, you know?
And, and, uh, Okay. And you know, I go, you know, M and Ms. Maybe ashtray publish something. I do research. So I come back. No, there's nothing. There has to be John you can't. Yeah. And industry with Navarro, this is enough. Wow. So he goes out and he looks at, he looks at what's happening with the early days of the semantic web.
you know, think about Gmail. Right. Remember, I, you know, for years I had my email was on a, you know, with a big company, Microsoft outlook, right? If you want to organize stuff, you had one tool folders. And if you put any email in one folder, it is not in any other folder. But if I put tags on, on an email, it can be, I can find it on this topic.
There's so the whole tap, anyways, he settled down to the right way to build a. Data modeling approach is with tags. Okay. What we started doing, he's working on and working on working on, we get together and he goes, you know, Hey, this, the state of tagging thing, he goes, this is critical. And this is not a competitive advantage.
This is a barrier like for us and for everybody, this should be solved. And so we had that conversation and Brian had the vision, and then we started talking to some other people. We knew, we respect, we talked to people at some of the energy labs, by the way. Okay. And I tell the story this way. They said, yes, the industry and solve this problem.
Great. Let's start an open source project and do it together. And they say, great. If we apply for a grant, we might get it in two years. Yeah. Just go do it. And so in March, we're celebrating the 10th year of haystack and March of 2011, we stood it up about a dozen. People are involved, stood up a server. And we turned over everything we'd done to make a functional product around tagging concepts.
Right. And which is still around. And we turned it over to open source under the academic free license. And we encourage come help solve this problem together. It's open source, let's work together and people come and say, wow, you guys don't know anything about electric distribution systems. I didn't say I did.
Right. Do you just volunteered? In fact, there was Steven Frank at N row. He came in and he built out using the methodology. We defined, built out a haystack model for electrical distribution, metering systems and all the tags. You need to describe that stuff. And that's just a key, really key example of something that's really important, but the community game together.
Right. And that is key to what we did with haystack. It's a true open source initiative. And I make a distinction. I think we did on one of your other calls, you know, There are companies who do development and then make it available as open source. This was developed as open source through an open source collaborative initiative, which w it was very different for our industry.
In fact, I think it's been one of the challenges is. The BAS industry is used to having standards handed to them. Okay. Right. This is the standard, right. And that doesn't mean you couldn't have participated, but the amount of people who participate in getting there is small, In general, and the methods and processes are rigid and complex and mind numbing.
Right. And this was, how's the software world solved source. Meritocracy is a better solution to that problem. Totally. And that's what we did that came together and it developed, you know, generation one, which was primarily the vocabulary. Let's all use the same words to markup data.
And then, you know, in between generation two or three, which enhanced that with the rest API for a standard way to query, right. And then of course, haystack for, you know, which will be finished its public, review at. Uh, the upcoming haystack connect in may, which is all virtual event, which implements taxonomies and ontologies, and the ability for inference based queries and RDF compatibility and all of those things that have come from openly collaborating, bringing in critique complaint requirements to enhance the standard.
So I'm the executive director. I kind of. Run the business side of it, even though I have some command of the technology I'll, let's this, stuff's getting pretty sophisticated around and Brian's the lead curator still. but there's lots of people contributing. We have a lot of work in groups where people run with it, they run with the methodology, they bring back the consensus.
It gets reviewed by the curator or curators to make sure that the methodology, the standard was implemented. and you know, it's, uh, clearly the, the most deployed tagging. Symantec methodology out there. It's not the only one. There's a number of them out there. As you know, you've held, some sessions on the different standards out there.
So that's a little bit about where it's at. We're hosting, what will be our haystack connect 10 year anniversary meeting. Won't be in person, but it'll be the fourth, a biennial event. We've had lots of speakers from around the world, so I can about advancing the standard, how they using it. Yeah. So that's a little bit about it.
And your next question.
James Dice: [01:07:16] Well, I was going to tell you, so in 2015, so there are every two years,
John Petze: [01:07:19] right? The history almost was in 2013 first event and 1517, 1921,
James Dice: [01:07:25] 2015. I was just getting started, at a new job. And so I was there for a month or so, and haystack connects coming up and I was like, I have to go.
And I can remember explaining to a bunch of sort of non analytics people. They're commissioning energy management people. Trying to explain what haystack was first and then trying to get the money to go. And, uh, I, I just, I failed miserably. but I can only imagine what it's been like for you for this entire time to explain what it is and why people should care.
And like it's such a difficult topic
John Petze: [01:07:57] to yeah. Yeah. And the way I deal with it without going crazy is until you, I relate to my Island until you encounter the problem, you don't understand the need for the solution. And that's the thing. If you are trying to get data from different systems and work with it, you have to have that data have a semantic model, so it can be interpreted.
Right. And one of the big things is people said, well, I have standard protocols. You have the protocol. Doesn't tell you what it is. It just says, here's the name? Here's the value? What is it? And one of the best examples I have a soft from this was a presence actually, when Tridium came out in support of it.
And John sub-let did a talk, he created a great analogy, which is, first of all, one of the key things we have helped people. Haystack is not a standard naming convention for BAS points. We don't tell people what's in it. And the reason is you couldn't solve it. That your name is James stuff.
What, what kind of car do you drive? How old are you? What's your job? Those are all attributes about you. Oh, we're going to build them into your name. No, you're not going to build it. What do you do? It's a data record and it has some structure and your credit card company has it in one way in a bank as another, whatever.
Right? We need a way to capture those attributes in a normalized format that can be easily exchanged between software programs and be both machine readable and human readable. That's key because one of the other things we did with that was come from the point of view of who's going to be at the frontline of implementing this classification.
That's the technicians who are doing a control system, right, right. The guys out there program or whatever they're coming from domain knowledge, they need to be able to tag things and understand what they're doing. Right. This is a supply air temp, sensor degrees, Fahrenheit associated with . They can understand that you give them the vocabulary methodology, which again, has been enhanced around and they can do that.
so we approached it from that end. In fact, when we talk to, some researchers who are doing the semantic web, they said, Oh, you're building a folk song. Okay. Because they were at this, you know, deep computer science, RDF layer, you know, resource description, which. You know, the new haystack for supports RDF expression of the model.
That's great, but you can't teach that to people doing buildings. This is on, Oh, this is on PhD level, you know, software engineering stuff. We approached it of what do people need to do? They need to classify what it is beyond the name. And so that, that has a lot to do with the philosophy. And when we started, so, you know, kind of ground up.
Right. And of course we encountered the need to enhance it and grow it the standard and you know, the, the acceptance. But yeah, if people don't understand the problem, they don't understand why we're doing this, you know? Um, but once people do encounter it, and of course now that more and more of the world is using data and there's numerous analytics companies out there.
You're either doing this around a standard or you're building yet another proprietary bespoke solution. That's only going to be useful for you. Why invest all that effort and maintenance of that let's work together to solve this together. So that was all right. Yeah. That's
James Dice: [01:11:20] awesome. Yeah. I ended up making it to the 2017 haystack connects because I did what every good project manager does.
And I built it into one of my project budgets. I funded
John Petze: [01:11:30] my own way. I didn't know. . Yes. Yes. Outside Tampa. Yeah. It was a great event. That's where I met
James Dice: [01:11:39] a lot of the people that are great members of nexus pro today. Tyson not sports people that I have great relationships with today started
John Petze: [01:11:47] there. It's uh, yeah, this is just part of the future.
It has to be done. So, uh, let's all work together and have one standard, you know, I'm sure you are others. Who've heard my analogy, but I'll throw it out because I think it's so important. Right. you know, HTML without HTML, we wouldn't be able to work with each other's websites. I couldn't read what you posted.
It's a defined way of marking up your publication, your website, the text, the graphics, the images, right? You can't have five of those out there. You can't have three of those out there. There's gotta be one. And you know, what, guess what? Getting there from HTML one to HTML five that we used to debate, you know, argument, building things out.
And extending feature sets. Look at what browsers can do today that they couldn't do with Aisha. Great. That's part of the process, right? So, yeah. Well, haystack doesn't have this great come on in and let's extend it. And we, our ethos has been very open and, uh, you know, we, we were the ones who reached out to brick and ASHRAE that led to two 23 P.
We've got to unify this. This is not a place for fragmentation because it's holding the industry back. Right. And yeah, you can find stuff that isn't in this standard of that. That's not the point. The point is we need to work together to solve this with one. So the only way we're going to make data easily interchangeable and break down this unnecessary cost barrier to improving the performance of the entire built environment.
James Dice: [01:13:19] Yup. And, as I told all of our members in January, I am very frustrated by the fact that there are a bunch of different, uh, competing, I guess, efforts at this point. And I totally agree that, that we need to converge into one. So and not everyone thinks that, but. I'm still not convinced that it doesn't need to happen.
So where did we go from here with all the different, and for people that don't know this, but just like, listen to John talk up to this point, there are there's project haystack, but then there are three or four or five, depending on how far you go in the world, others at this point.
So, John, where do we go from here?
John Petze: [01:13:53] Where should we go from here? Yeah. So, you know, for us, it's just continuing to evangelize both what was accomplished, how it solves the problem and the openness of the community. Right? So I'm in work, you know, through the collaboration with ASHRAE and brick, we came out with the more defined ontology and a fixed way of expressing taxonomy and RDF expressions that was important to, you know, people working at that level.
You know, we've tried to be inclusive, take that in. And, implemented and extend the standard continuously. Right. And, uh, add models for equipment that hasn't been, built into the library. And I've seen a couple of challenges. One challenge I see is that open source community based open source development, hasn't been part of the fabric of our industry.
Right. It's been the handed down standards. Right. Okay. and the other point is there's this, well, John, it's not in there. When are you going to do it? When are you going to come out? Right. And that's not that isn't yeah. It's completely volunteer by jobs. Yeah. there are no paid employees, right.
In haystack, a hundred percent open source. Right. Um, the other point I'd make is that, I see an unfortunate hangover from the protocol Wars, right back Matt and lawn and all that type of stuff. And the manufacturer through him with, uh, with one, and then they lost customer and whatever. And you know, then they kind of did both back then emerge.
I see a mental hang over there because I know that there are companies utilizing haystack in their products who won't admit it, who won't publicize it. Why not get, well, we don't want to throw in with haystack. What if, what if one of the other ones ends up winning? We don't want to throw it on a marker.
James Dice: [01:15:42] Uh, so you're saying people wouldn't say invested in lawn for instance, and ended up wasting that investment because backnet won out
John Petze: [01:15:51] and there's yeah, well, not only wasting the investment, but hurting their market opportunity. So I'll put it this way. This is a Tridium story. There's two stories, right?
Okay. Tridium had. Network management tools for lawn independently developed. Right. And so we showed people that people go, Oh yeah, you're in the lawn camp. Okay. You're against that BACnet stuff then. And we're like, well, wait a minute. I didn't say that. Right. And we show people our backs and it tells us like, Oh, you're good.
You at the back now you're against that lawn. Whoa. Yeah. Well, why it's there's this psychological thing? So I personally believe that has extended to this, the numerous companies we know who are using this, don't want to publicly support it, admit it, whatever, because maybe they're afraid. Well, if the other one, because I don't want to show that I was biased or whatever, even though it's, you know, it's the most developed market deployed solution in this area out there and it continues to expand.
And if anybody's tried to come in and. contribute and extend the standard. they would see that, you know, it's open, um, at Saddlebrook. I don't know if you remember there was, um, I think it was some speakers from Canada and their point was we want a model flows of energy and resources through systems and the haystack model doesn't do it yet.
I had to remember that, remember that I believe it was a lady from Canada who actually presented them. Well, guess what? Haystack four includes. You can model relationship flows of those resources. Those phenomena need the input. Right. Come work on it, come help us solve it. So, you know, I don't know what will happen.
I know haystack has a worldwide base of people use it. Right. Um, and it solves so many problems that we. Yeah, we're obviously biased. It can't help that we're human beings, but we can see other standards and we can say, okay, throw that problem that they haven't even got to solving that problem yet.
That's not an IQ thing. It means. Yeah. It took, you know, we didn't get to it for the first couple of years either, but you're going to have to solve that. And that, and that, and
James Dice: [01:17:48] is that every time someone starts a new standard, all of this stuff needs to get solved again, in a different silo, talks about,
John Petze: [01:17:57] you know, or not solved.
And people may believe it because that's another thing is the. Depth of understanding of people who are looking to adopt a standard. those are often decision makers who don't have deep understanding of this. So can they discern that they're looking at one that actually won't give them the end result they need, but may come from a company with a big, big label.
Right. And so, I mean, you know, we've been fortunate to have Siemens and until it was on the board for a number of years and others, but I don't know why people don't see the similarity between HTML and this. Right. I just, I don't, I fundamentally don't understand that. So. Absolutely. All right.
James Dice: [01:18:40] So haystack four is going to
John Petze: [01:18:41] come out and may, you know, for almost two years in public review and lots of commentary extension, but you know, we're going to close the public review. Brian Frank will be making, you know, a number of presentations around it at closing the public review. It doesn't mean it won't continue to advance, but there's a lot of people looking for that, you know, stamp it's released.
Right. I mean, people have been building products on it for two plus years, so.
James Dice: [01:19:04] Okay. All right. Well, what else in 2021, are you excited about besides that?
John Petze: [01:19:09] Well, I think it's important that, we see the physical, World come back together. You know, there's a lot we can do virtually, you know, you and I probably wouldn't have been able to get together in person today, aside from COVID.
Right. So this has helped, but I believe that it's a detriment that we're not having the shows and the conferences and the collaboration and the networking, in our industry, you know, um, you don't have that, you know, the serendipity that comes out of, meeting people and talking about things that you didn't plan everything's planned.
Well, we, I have a meeting with James, we're going to talk about this stuff, right. It wasn't like we happened to bump into each other at Bartle of nine o'clock at night in Nashville. Right. So we need that back. So I am cautiously optimistic that we're seeing the corner turn in, that we'll come back to some degree.
I think it's going to take, you know, it's not going to be back as strong as it was for a variety of reasons, but, that, and, you know, excited that more people understand. The data is critical and they need to work with it in different ways, whether you want to make digital twins or do analytics, or do all kinds of other things, the foundation is data, right?
so I think we're on this journey in our industry to, you know, learn how to use data. Just like every other industry is all at different paces and different levels of sophistication know.
James Dice: [01:20:26] I love that. I love that answer It does feel like at least in my, the projects I'm still working on from a consulting standpoint, the value perceived value of data is only going up, uh, the same with analytics at this point.
So, all right, John, this has been so much fun. Thanks so much for doing this. I'm glad we finally, finally got
John Petze: [01:20:45] together. Yeah, I'm glad we did too. And again, appreciate what you're doing out there. Helping to educate the industry. It really matters. So doing my best.
James Dice: [01:20:54] All right.
John Petze: [01:20:55] Take care. Bye.
James Dice: [01:20:56] all right, friends. Thanks for listening to this episode of the nexus podcast for more episodes like this, and to get the weekly nexus newsletter, which by the way, readers have said is the best way to stay up to date on the future of the smart building industry. Please firstname.lastname@example.org. You can find the show notes for this conversation there as well. Have a great day.