41 min read

🎧 #087: Andy Frank on the Independent Data Layer

“You've got control systems in the building and interesting applications up in the cloud. And then there's something in the middle and it's really gnarly and messy and ugly and a pain in the butt to work with. That's what we do."

—Andy Frank

Welcome to Nexus, a newsletter and podcast for smart people applying smart building technology—hosted by James Dice. If you’re new to Nexus, you might want to start here.

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 87 is a conversation with Andy Frank, founder of Novant and formerly of Tridium and SkyFoundry.


We talked about Novant’s founding story and their unique approach to the independent data layer of the smart building stack.

For context, I shared my not-so-great experience with implementing tech before products like Novant were available, which should entertain some of you.

Then we zoomed out a bit and talked about the challenges Andy sees in implementing smart building use cases and where the market is headed in the future.

Without further ado, please enjoy the Nexus podcast with Andy Frank.

  1. Novant (2:47)
  2. 🎧 #034: Brian Turner on the Ontology Wars and the role of the MSI (14:50)
  3. 🎧 #048: Andrew Rodgers on the IDL... (14:50)
  4. Healthy Buildings by Joe Allen (38:31)
  5. James' "Healthy Buildings" LinkedIn conversation (41:21)

You can find Andy on LinkedIn.



  • Novant's founding story (2:47)
  • The challenges Andy sees in the smart building market today (14:36)
  • Where Andy sees the smart building market going (37:44)

Music credit: Dream Big by Audiobinger—licensed under an Attribution-NonCommercial-ShareAlike License.

Full transcript

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: 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.

James Dice: This episode is a conversation with Andy Frank, founder of Novanto and formerly of tritium and sky Foundry. We talked about Navon founding story and their unique approach to the independent data layer of the smart building stack. For context. I shared my not so great experience with implementing technology before products like the avant were available, which should entertain some of you. Then we zoomed out a bit and talked about the challenges Andy sees and implementing smart building use cases and [00:01:00] where the market is headed in the future. Without further ado, please enjoy the nexus podcast with Andy Frank. Hey, Andy, welcome to the show. Glad to have you on, can you start by introducing yourself?

Andy Frank: Sure. I am Andy Frank my mood with no Vince. I don't know if I was one of the, now I don't know if I was one of the first people in the, in the nexus for landscape. I feel like I was early on. So it's been exciting to actually be part of the.

The cookouts finally, on one of these things, maybe, hopefully at least one person finds anything. I have to say. Interesting. Yeah.

James Dice: Well, thank you for being a part of the community since early, early times. I think we're more like the cool nerds, which is like, it puts us in the upper echelon of the nerds, but it doesn't make us cool kids.

I don't think, When I describe what I do at like a dinner party. I don't see like a cool kid. I feel like a weird, weird kid. So, well, cool. What's before starting event, what's your background? What'd you do before that? I mean, I [00:02:00] already know, but can you,

Andy Frank: So I've been in this space, I guess since I got to college very reluctantly you know, speaking of explaining to people what to do if building.

If you don't think the buildings are cool now they definitely were not 20 years ago when no one knew what that building automation was. I'm not sure I even did. So, I went to work Tridium after college. I was there for five, six years. Worked on the core team on Agra X. And then once when Honeywell acquired Tridium my brother and I left and we started to Skype.

Did that for 10 years or so? 12, 13, I guess time keeps ticking up on me. And then maybe a year and a half ago, two years ago, I sort of formally started to know them. That's the British version.

James Dice: Nice. Cool. And what is no van and why did you start?

Andy Frank: Honestly it was just something that I wanted to do that was different. I think I spent 20 years largely doing the same thing. Pretty much. I find it couldn't be more different, but fundamentally it's visualization and tooling around, [00:03:00] you know, building out. And I think I just did all I could handle after awhile.

And so I think, you know, going back. Maybe 2013, 2014, I worked with company to put sky's spark on embedded systems. And I got introduced to, to embedded Lanex hardware, that kind of stuff. And just really interested me. And so in the back of my mind, I was like, how do I come, come back to this?

And so before Nova was really an event, it was, I want to build hardware. And I want to do stuff in building automation, because I don't really know how to anything else, but at this time at 20 years and I'm sort of, am I leaving out? What the hell can I do? that led to he was a base level of technology.

You need to operate in the space you need IO need, you'll talk back net you know, my bus haystack and XL stuff. He's a kind of pluggable the protocol to an ISO system. You need to be embedded at trending. Well, I think you need a better trending and embedded trending on low-cost side was very different than if we get it in the cloud.

There's different. There's connectivity. I think you need to think about security in 20 21, 20 22 you know, [00:04:00] remote connectivity, all that kind of stuff. And data gateway was the easiest product to build while flushing, all that technology out. And then once we have that in case which we do now, it gives us a lot of options for what we might do next.

Okay. So, I wasn't even sure that this thing was a market. And then what, six months in, I think you wrote a blog post on the independent data layer. And I was like, oh, I'm not the only one. I think it's sort of morphed into a real thing. You know, since then,

James Dice: Totally. And I don't mean to gloss over the whole like founding sky Foundry thing and all of that.

We dove pretty deep into that topic with John, John petsy. And we'll link to that in the show notes. So sorry. If, if everyone was like, well, wait, we're skipping this huge story or we'll gloss through that, unless there's anything you want to share about that phase of your life. Uh, As you look back on.

Andy Frank: No, not really. [00:05:00] It was fun and interesting. I'm not sure I have any desire to do it again, necessarily twenties then. And I, don't not sure that I have the energy to do that in the same way again. But it was it was, it was some time back then. Technology was different. Startups were different.

Everything was different. Back then, so it was fun, but. Ah,

James Dice: nice. Okay, cool. Talk to me more so, so it sounds like Nirvana was similar Nirvana, no vans.

Andy Frank: No, I don't correct people. I'm trying to see who says it more and then I'll get with that.

James Dice: I've always been like no avant was in my mind. And now I've heard you say no, a couple of times some Metro.

Andy Frank: So I, I mean, November was, I had to find a domain name that wasn't taken in a trademark. And so it was like the third name I picked with my attorney and I was finally done and said, and had done everything. And someone was like us. He worked for Nevada health and I'm like, who's not on health. They're [00:06:00] a giant health system in Southeast.

You asked them a great, but at this point that ship has sailed. So, and worst case they're there. They helped me with brand recognition.

James Dice: Yeah. Yeah. That happened to me. A real calm this year, there was a huge booth of a company called nexus. And I, I was like, look at this booth that James got nexus lab. So I was doing really well.

I was like, nah, that's not me. But where I was going with that is the. It seems like you're similar. Like I, when I started nexus, I was like, I think I just want to write this newsletter. And then as it has kind of led me into this company that actually. Didn't have fully planned out at the beginning.

Right. Whereas some people start a company and they're like, this is my long-term vision. But talking to you offline, it seems like you do have this longer term vision versus what you guys started with, which is this gateway and independent data layer software. Can you, can you talk about that [00:07:00] long-term vision before we would talk about the short.

Andy Frank: Yeah. So in a lot of that is also, you know, I've done the things certainly 20 years. And part of it was, what if I did it differently? Is this a grand experiment? That's going to fail miserably. I think I'm past that point, but You know, in the past with the way we've approached stuff is we spent, you know, months and years building a very specific tech stack and that works great if you know exactly how you're gonna use it and what the product market fit is, the reality is you almost never get that.

Right. And so you're sort of forcing it to fit in. And so, I tried to keep it a little more loose with Novant where we've got a much lighter technology, it's more composable and set that we can sort of mix and match things more easily to actually fit the real problems that, that, that, that surface.

And that's just the thing that it takes time. You know, you, more customers, you get more exposure, you get more problems you get, and you need a feedback loop to really figure out, you know, what, what you're trying to actually solve. [00:08:00] So that's sort of our philosophy with, with how we approach it from technology perspective.

But you know, long-term. I'm a big believer that technology can automate and solve a lot of problems in any field. We're seeing that all over the place, particularly in buildings. And so that sort of, how can we.

Be a technology partner to the people. I think we're sort of a horizontal platform layers where we want to stick. Mostly just, I can't do another visualization or chart. I just don't want done. I've spent to help solve these problems. Right. And I don't know what those problems are. I just know that you've gotta be nimble and responsive and there's, there's a certain technology and expertise that I think you need to do that.

But a lot of it is going to be a cooperation between You know, emphasized building owners, other technology, also evolving and communities like nexus that brings people together and actually forced them to talk to sort of see, you know, tease out, you know, what, what does actual, you know, solutions and pathways look like?

Cool. So

James Dice: it sounds like what you're up to today, like where you started with is this [00:09:00] concept of that I've written a lot about, and I'll link to in the show notes is the independent data layer. So what, when you described this concept and obviously I've tried to explain it to everyone as well, but when you describe it, How do you describe it?

Like, what is it? And maybe pretend that like you're explaining to a five-year old before, and then we'll build on it from there.

Andy Frank: The way I think I frame it. And I don't know if I can dumb it down that bad. Um, But impossible questioning. You know, when I think about this in early, like course blocks, I think you've got sort of controlled systems and building. Interesting applications up in the cloud and then there's something in the metal and it's really gnarly and messy and ugly and a pain in the butt to work with.

And you know, specifically, I think you can break that middle block up into three things. One is just connectivity, right? How do I connect to the systems? Right. It's the protocol implementations You know, it's, you know, how, how can they do [00:10:00] discovery? How efficiently can they pull stuff? Did they manage to bring down the whole, whole control network when they're doing stuff?

Right. Cause a lot of times it's secondary to the building systems. Um, Good role. And then, you know, the second one is. Yeah, trend data which I think is largely, even today has gone overlooked by a lot of companies that sort of secondary to what their actual purpose in life is. And you know, there's some groups that have, that are like, yeah, this is easy.

We do this all time, but it's likely with the dome, right? If you're an energy consultant or something that you, you probably don't have that expertise on staff and you don't have. Yeah, you've got some data sitting on some instance that you booted up three years ago, that if it went down, you wouldn't be able to bring it back up, who knows if it's secure or backed up or anything.

And so there's a lot of hidden complexity to actually doing that and storing that data. That I think is a great candidate to outsource for almost anybody. And I think third, the third really, really big one is this concept of onboarding data. Right? I've got a bunch of [00:11:00] data in the system. I got an application that needs it.

How do I, how do I immerse this too? So that's a big, gnarly problem. That's really the biggest problem in the space that I see. I'm probably not exactly answering your question here. And it's certainly not simple and quick, but I think the value from that is you outsource a lot of complexity. Yeah, that's the number one bolt to people to, to vendors, right?

Is they don't have to do this work. No, I haven't talked to a single person that enjoys doing this, or even the wall to do this. So if there's an easy way to do it, they will happily outsource them. Yeah. I think from the end, from the building's perspective, you know, the owner I see that as a celebratory.

Yeah, they have more, the data. They're not asking their, their contractor or their vendor to give them their data back to them. Right. They have it, they control who gets access to it. Having that onboarding in this trends done once I think gives you a better economies to scale, right? You're not duplicating that work for different, you know, I don't think the silos are going away in the next year or two.

[00:12:00] So the short-term fix to that is okay. Well let's slowly work our way up that stack, you know, first one let's do a common. Right. And then you, you know, you're D duplicating that work from the data perspective, but also gives you as a building owner, more portability. Right? If I don't like vendor X. Okay, well, I'll stand up vendor B here's the data, you know, what you got and if it's better, then I'll switch.

Right. It gives you more portability. Um, And

James Dice: I want to underscore that the outsourcing point. So circa 2015, I think it was, I was leading a team. We were sky spark reseller. And because I was the leader of the team, all mechanical engineers. Right? So no software people, no networking people, no cloud experts.

I, myself, basically in my free time, stood up an AWS. Whole environment. Right. And figured all of that out on my own pretty much. And started connecting that [00:13:00] environment to hospital networks. W where else was I connecting it to everywhere that I needed to, to, to be able to set up site, to site VPN connections and get data flowing.

And obviously it worked, but like how much, how many products would I have done? What I have needed to do to pay back that investment time. And I think that is happening all over the world right now, because we were early ish then you know, the number of consultants that were doing what we were doing in 2015, that was early.

I think. Businesses like that nowadays it's gotta be table stakes. You're doing retrocommissioning projects. You're doing commissioning on a new building it's table stakes to use FDD and those companies, can't all be trying to figure out the cloud game and the integration game and in that working game.

Andy Frank: Yeah. And I think the two and the [00:14:00] two things, I try to emphasize that as one. Maybe it's relatively easy to set up, but six months down the line, when you, you don't remember what the hell you did and something breaks and it always happens at the worst possible time. Yeah. And

James Dice: It was not relatively easy to set up.

I just threw a lot of hours at it.

Andy Frank: Yeah. And it's just, you know yeah. And you don't know what you did right. And what you didn't do wrong.

James Dice: Yep. And then there's the risk of doing

Andy Frank: something wrong. Yeah. You've got those possibility. You know, it wasn't security breaches where the data breaches, you know, all that kind of stuff that goes along with it.

That just really isn't necessary.

James Dice: All right. Yeah. So the, you know, year and a half or two years that you've been doing. Initial product thing. I'm curious, sort of what challenges that you see to implementing the IDL. So on this podcast, we've talked about the IDL concept. I've talked about it with Brian and Andrew we'll link to all those conversations, but like, I don't know [00:15:00] that we've really gotten into like, well, what, what are the practical challenges to actually doing it in the real world at this point?

I'd love to hear what your thoughts.

Andy Frank: It's sobering. How complicated actually as a practice you know, I've done this for 20 years, but it's always been in the context of get a good enough to move on to the next person. And that's why I think you see from a lot of people, right as this. Onboarding data better than the next guy doesn't make you better at what you do.

So, it's sort of what you did. I mean, it's what you do, but you've just spent as much time as possible to get to the next thing. But when all you're doing all day long is focused on this. I think you. You appreciate the complexity of everything involved. And it's all, but it's also an interesting challenge because you start to see some of the details that he previously probably overlooked or didn't go deep on enough.

Okay. And you know, he just had a call with the customer. Yesterday actually. And I I'm [00:16:00] I'm I know all of my clients, cause I'm like, tell me everything that sucks about this. How can we get better? You know? Cause some people are just too nice to them and they don't want to tell you the best stuff.

I do my best to be brutal, brutally honest with me, but you know, it's like, well, how are you doing this? And you know, they're like, whoa, you know, we take all the data. Some of the naming conventions Are inconsistent, which I think that's people know that by now. But then they're like, oh, I want you to know which one is which so we'll take, you know, and I don't know if they're doing it on a point basis or spot checking, but you know, it was basically taking the point out of that, that they think it is.

And they would go into the BMS, pull up in the graphic of the BMS and actually match the points one by one, like visually. And I just cringed at that. Cause I was like, I don't know how you solve. Today, but I'm like, that's got to get better. You got, I mean, we just, you know, that's what I think has been most eye opening is how much of a house of cards the sexually is.

Okay. We keep these huge sophisticated, you know, digital twins, analytics, all this stuff, but they're on like the, like a, like [00:17:00] a sand foundation that's already part of any time. And you can make a lot of assumptions. I think that that debt is correct. And I think a lot of cases, I guarantee you, there's quite a lot of people pulling in data that isn't what they think it is.

And that's a big, gnarly problem solve. You know, I think, you know, AML, AI, those things will aid and, and doing sort of identification and modeling, but I don't think they solved the other part which is. What's the other problem. I'm not strata capture that explained that. I think it's the fact that I've got all these sensors in the building affect control systems.

How do I know what's what, in that it's correct. It's almost like, Like unit tests in a different program, maybe development. Right? I can do that. I can run this unit tests. I know that this, this, this, this, this is the way it works. Both that upfront and also as a regression tests, as I advance the system, you almost need, like I said, not like that for buildings.

Right. So if I do something, I can go back into like a switch and say like, yeah, you know, these 300 sensors, air quality sensors in my building are actually one working, but two, they are [00:18:00] aware they are, they are what they are. They are. Where they are and what you think they are. Yeah. And that I think is that's not something that we've solved on our own.

I think that's probably for community effort that need technology and inside buildings to continue to evolve. If anything, I think that's been.

It's mainly a huge proponent of NSIs. I think, you know, what I do with my customers going into buildings is it's also a huge people problem. Like one wants to help coordinate anything. And it's like not having that point person on top of everything just slows everything down, too loud to crawl. And so I think that sort of a combination of all of these people sort of slowly and iteratively getting better over time to help solve.

But I think that is I just, the engineer in me thinks that sees that it's like, that needs to get done and be super solid before we put too much effort and building on top of that and they can happen in parallel, but I think they've got to catch up at some point, I think.

James Dice: Yeah. And in that catching up to me, it seems like [00:19:00] there's so much opportunity for whoever is setting up and maintaining the software in the building for them to do a better job, documented this driving, what what's actually happening.

So you're talking about what sensors do I have? Where are they at? What are they measuring? Right. What meters do I have? Where are they at? What are they measuring? Which ones are upstream and downstream of each other. I think there's so much room for. In existing controller and existing system to self-describe better than they're currently doing.

Andy Frank: Yeah. And I think there's a lot of out of band contexts. They're like, yeah, they probably have a big spreadsheet and Excel somewhere that describes the structure, but it's not. Yeah. I posted on last around a couple days ago. You know, I always find interesting technology, you know, the fields that can be applied to buildings, one of them was IDB, right.

I can plug the little reader in any car and get all of the diagnostics information and it tells me interesting comments, but I think one of the takeaways was that that, yeah, I can get sort of [00:20:00] raw data, but it's sort of that, that super structural. Information that describes it, all that I think we're missing today and that's not readable.

You know, we, I can pull a hundred thousand points out of the building, but it's just a flat list of points and controllers. Right. There's not that sort of, and I think there's, I think there's, there's steps between the full digital twin semantic haystack. Just sort of some structural things in between that I think we're missing that need to be machine machine-readable that that's not there.

And I don't know exactly what that looks like yet, but I think that's a gap that needs to solved. Yeah.

James Dice: Yeah, those are huge challenges with the independent data layer. One of the ones that I have that I've had a couple of different projects related to this is just being able to make the business case that there needs to be a dedicated infrastructure layer, because so when people are so used to making business cases that are based on energy savings or based on tenant experience or whatever, right.

And the independent data [00:21:00] layer. Infrastructure. It's like, it's, it's a couple degrees removed from the actual use cases that you're trying to enable. So how do you think about business case? It sounds like when you said my client early. You're going directly to service providers and application providers that need the data to enable what they're trying to do.

You're not necessarily going toward, going into a building owner to make the business case.

Andy Frank: I think there's a couple of different at least from our experience so far. And I think to add onto the fact that you're removed from the energy stuff is also that technically you can use. So it's, why am I paying extra for something I can already do?

I think people that have done this long enough know that, that you get what you pay for it. But I think the thing, the three markets, I think that we are seeing you know, one is I think the long-term yeah, I think it's from MSIs and building owners. It's going to be a spec item or even required in a building.

So they have this IDEO just because they see the value in it. And I think there's people that are taking the tires in that today, but I think that's a long-term. Yeah, [00:22:00] well, two ones I see today are, you know, people like energy consultants or optimization consultants who. Or not it experts or, or just don't want to, it's not worth it to them to manage this stuff and do this, this kind of thing.

So if there's a good cost effective, easy use alternative, you know, they're happy to use it. And that's sort of get the data out of the control system and then they don't have to have access to the actual, the amassing, that kind of stuff. So it just works economics and then process works a lot easier.

The third one is. Which I'm not sure that, that I anticipated, I guess I knew, but I was surprised that I got the feedback that I've had is that people that it is part of their product, but they don't want it. Yeah, right. You know, all these products and platforms need the data, but I hear the same story and every, like, it's, it's a pain, you know, we don't, we don't make any money.

We lose money on this stuff. So if there's someone else will solve this problem for us, we will happily outsource it at the right price. So that's sort of where I see, I see three things emerging right now. [00:23:00] But it's a B2B business, I think. Right. I don't see at least the way our technology works.

We're not going directly to building owners because you're right. I don't, I don't have a compelling argument of why they needed to do this today. Yeah. Yeah.

James Dice: Okay. And so when I think about my past life as a energy consultant or commissioning agent, right. I would have a lump sum project or in some, sometimes an annual maintenance.

Contract that I would be trying to fit this effort underneath. Right. So I'm trying to budget out how many hours it's going to take me, how many, maybe skies from our points do I need to buy what's my aggregated cloud costs for this project. Can you describe what it would be like hiring you instead?

Like, I think it's just a, it's a, a center point or something like that. Like how, how does your.

Andy Frank: Yeah. Yeah. Yeah. And so I like pray, wait over [00:24:00] simplify what we actually simplify, but that's me. I just, I moved to, like, my mind is like a hundred miles per hour all the time. You know, when I say make it easy, there's a lot of context stash of that one.

It's you gotta be able to put in the building and just work. Right. And, and I, people that have never stepped foot in a building, they just shipped the box to a building, someone plugs into the network and they do what. Too, is that, you know, as a we're a volume business play, so it has to be minimal.

I can't hold everyone's hand. So the tooling and everything has to be super, super simple, easy to use self-explanatory. Which means it won't be the most complicated product. But I would argue that most people don't need what they think they need. They'll happily get by with something. And the other part of that is pricing where you. You need something that's easy to figure out, you know, go back even even a sky Foundry. I think there was not as much anymore, but I think originally early on, I was, well, how much, how much this space we need for how many points and let me know, we can do the back of the napkin math, [00:25:00] that's it.

But it's, it's, it's relatively straight forward, but it's like just enough of a mental hurdle that maybe it slows people down or just that you want to remove some impediments. So. Technically as a data provider, I should store by gigabyte, right? Because that's where my costs are and I can pass on to you much more, but no one understands what a point per gigabyte means.

No, I think you've got to capture this cost in ways that really people really easy for people to, to, to, to, to, to think about which is, you know, 0.001, 5 cents per point per month. Right. And it's all utility based usage based. So they know that if I'm going to do this job, then I need 5,000 points. You know, I know that what I'm paying for month and if I only need the job three months, I pay for that.

And then I'm off and never have to pet on that. I don't continue to pay. So, you know, I think it's, it's, there's a bunch of, a bunch of simplicity involved in simplifying around that.

James Dice: Cool. I, I just, I just wish you were around. You would have been doing this in 2015. So keep, keep going. Hey guys, just another quick note from our sponsor Nexus labs. And then we'll get back [00:26:00] 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 course@courses.nexus lab. Start online. All right, back to the interview

how, how your, I want to talk about how your approach differs.

So we've, we've talked about the IDL, you know, you're obviously seeing the vendor landscape, a bunch of people are attacking this problem have, have identified this as something they want to build a product for. How are you thinking about this differently than others that are, that are approaching that.

Andy Frank: Well, I will say in the context of the companies that I know we're doing this, so which is probably not exhaustive. And so I think, you know, [00:27:00] probably in a couple of different ways, one I don't think there's necessarily one way to do this. I'm not even sure we know how to do it yet. Did you have this baby steps?

You know, I think we've solved the connectivity. We've got a great product for connectivity, remote access. We've got a great story for trending, but you know, that last step, that talking point onboarding is, is it's, it's a hard problem to solve. And I think it's going to be in a roof. And so I think there's a journey there over the next couple of years, maybe longer we continue to attack that and that's just sort of based on experience as we get there.

And there's lots of tools in the tool belt that you'll use to attack that. Um, I don't think we have the data sets and the training data to necessarily solve that a hundred percent with ML. I think we get there. But I also, you know, this is big data summit that happens with Richmond every year. And it's, it's actually more geared towards financial companies.

So it's very, very different. But you actually get, you get a a intentioning perspective on how actually early that eScience is in its sort of [00:28:00] career. And we take advantage that it can do more than it can. Sometimes. And so I think there's evolution that all needs to happen to get there.

So, but one of the things is like, you know, if you're automating something, you need to know what you're automating first. And so our approach is let's build better tools to let humans do that. And then let's evolves slowly to automating that w you know, with machines as we get out you know, I don't know if we're different.

Cause I think we're one of few companies that just sit down puts are all out a pricing online. Um, Yeah, but I think, I think economics is an important aspect to this as well, because you know, like I said before, you can already do this today. Why would I pay extra to do this again? And so I think that you've got to be conscious of the fact that especially in consultants, you know, they make X percentage on a job.

You know, they can't probably pay $5,000 a month or a year or whatever it is. To do data, you know, they might, there might as well as it's sweat equity, they might want to just do the groundwork themselves to [00:29:00] justify outsourcing this kind of stuff. You've got to be, there's a price point you need to be at.

We will do things over the past year and a half. I'm not sure where exactly where the market should be, but I think we're close. I'm probably too cheap in some cases. But.

James Dice: Yeah. And, and one of the things that I've thought about in this, this sort of technology category, right, is a company like you starts out selling to the energy consultant at what point, and they set it up on their client's building.

Right. At what point does it now become an asset that the building owner cares about and. Wants to pay for and wants to add other systems, you know, energy consultant, my dad, BMS and meters building owner has a lot more concerns that could make better use of a data layer. Right. At what point does it traverse over to this bigger picture?


Andy Frank: I think it's two driving things there. And one of them is, is I'm seeing a [00:30:00] little bit today, is that little it's on the consultant or, or whoever is buying from us, right? Is they may put the box there temporarily to get down. But obviously they're looking for service revenue, recurring revenue.

So they're doing some degree I'm at Liberty to how they approach that. But I think the other one which only really talked about yet is really focused on the data, but I'm not sure that that's actually really our mission statement. What we really want to do is sort of democratize how you access buildings and that.

Okay. And so, you know, long-term, I talked to this a little bit at the beginning, you know, long, you know, and I think we're tried trending to this way is that I don't care what excellent. Great. I don't even care if it's an open, I just want to rest API right. To give to my developer. And I think you're seeing that with prop tech companies coming in more traditional Silicon valley companies entering the space that you're, you're seeing that transition.

And I think that that would be important to, to, to make [00:31:00] that type of software coming down the line, work with, you know, existing building. Obviously you need this sort of intermediary layer. At least in the short term. And so I think when you've got this really, really easy to use API that becomes ubiquitous.

I think you started enabling a lot more use cases. So when you say, why don't you start a new blog where use cases, but also you lower the barrier for future innovation. Right? So someone's so that, so, so to some degree, this is a long-term play that like, okay, well, you know, build a point where they come.

If it's there, people will be like, you know, some 20 year old kid in Palo Alto is like, I'm going to build the next Whatever, you know, Hey, there's just easiest API. I can talk to any building, you know, I can get a start. If a context for this is straight and I've talked to us all the time, like Stripe is a big influence on what I'm doing.

When I, when I first started using them for this sort of side hustle, I did years ago for credit card payments, I was like, I had, you know, in an hour or two stood up predator processing, which no one, if anyone was doing that even now 10 years ago knows that it was incredibly painful. It was a huge moment in transition the whole industry.

And I think that you're seeing that now [00:32:00] with this space where we're getting the guys in the building, because that, I don't think people realize how much innovation will happen because of that. Right. So, so it's an, it's a, it's a technology enabler in a lot of cases as well which is sort of hoping to prayer that it happens.

So I think you've got to balance that out with what are the short term needs. And I think there's there. I mean, everyone will tell you that as hard, they want it to be easier. So I think there are short term needs, but I think, you know, I, to me personally, for Novant to stay relevant over the next 30, 40 years, it's are we on the same trajectory of where technology needs are actually.

And not just sort of have our blinders on that, you know, we solve the problem and rate it, you know, we're done.

James Dice: sticking on to your sort of approach to the IDL. I I'm wondering, you've talked about maybe your skepticism on how much machine learning actually provides to the process. Talk to me about your approach to data modeling itself.

So that's, that's, I think that's a key debate in the nerds [00:33:00] fear around the IDL is like, should the IDL have a fully descriptive data model or is there going to be, need to be modeling done at the application layer as well? Talk to me

Andy Frank: about your approach. I pull this out probably overemphasize it. There was this this is probably early two thousands.

So this is where I think when law enforcement was in development for anyone old enough to have the Longhorn. windows and stuff at the time, like all these Microsoft engineers were blogging for anyone doesn't know lone star. They were French assessor to social media, Twitter. Um, But um, they were blogging about all the stuff they're doing.

There's one guy that's blocked, but this is a stuck in my mind for 20 years, 15 years. Was that at the time I hadn't realized this, but all the controls and Microsoft office. We're custom rendered and built because the standard toolkit and windows wasn't flexible enough. So part of it was working in Longhorn was and, and I'm probably getting this wrong, [00:34:00] but in dynamic and XAML some stuff they were doing at the time was that they were building this technology stack where you could tie in a different.

To where it made it perfect. So for Microsoft office, they could, you know, baby, they could import all the inventing for instance, and they could customize the look and feel of the actual button using this word. And. I don't know if that's why, or if it just contributed to, but that's sort of been, so my philosophy to technology today, right, is that it's not about taking, you know, this big mountain of technology stack and forcing you to come in at the top.

It's like building these layers, you solve complexity and layers that sort of mind all technology approach. And so, and you make each of those layers accessible so that where it's appropriate, you plug in. Yeah. And so when you look at modeling, I think that's important that I make the distinction because sometimes you don't care.

Sometimes you just want the data out and you want to pull it into Python or are you just want to play with the data and you don't want that sort of impediment in a way. But there's a lot of applications where it absolutely is critical. And so, you know, where I look at our roadmap for that is one, [00:35:00] the applications that need it mostly, already have great tooling to do that.

So I think now you look at plugging in an additional tool in this stack. Like no event you want to minimize. Our goal is that if we forget that we're there, we're doing a good job. Okay. We want to minimize the time that you actually have to interact with our system. And at some point when the entire thing is actually enabled through the API, you'll never ever log in to Novant dot IAO ever.

You won't even notice there because it will be native tooling and all the extra downstream tools. That's sort of the Nirvana from the API perspective. And so. Part of that is let's get the data somewhere else who already has the technology and the expertise. And I use guys work a lot with that context because, you know, I don't have the concept of sites and floors and stuff at the data layer.

Right. But I, and I cases I need that prescribers. So then building up that structure, it makes no sense to duplicate. Twice, right. I think that's true for live applications. So you sort of let the, the application handle however they want [00:36:00] to. But you know, having, you know, a year of experience in real buildings, you know, it's very obvious that it would be relatively trivial to tag a lot of the data coming in.

We know that this is a you know, a district Navy controller or Siemens XYZ controller. And with very high confidence, we scan. Pre tag these things or suggest what these things should be tagged. That's the only part of the model, but I think it aids that it does those further steps, but I think that's sort of an optional thing that E that you add.

but also as much as I would like the whole full, fully tagged digital twin to happen. I just think it's still really murky how you get there a little bit. Um, So as, as with everything I do, it's like baby steps, right? Until it's really, really clear, you know, what the problem space is and then attack it at full speed

James Dice: and how you can provide the tooling for it.

Then maybe how to automate it later on. It makes sense to me.

Andy Frank: Yep. So, um, everything, yeah, everything we do is really context driven. You know, and the [00:37:00] feedback loop is so critical. It's finally went out. It's why we released so early because I needed that feedback loop to see what real people are doing.

And so you build sort of more specific workflow tooling to sort of handle all this things set of sort of saying here's an event you make, redo everything you do in your work. To work with us. I think that that's the wrong mentality. It's got to be flipped around to make people's lives easier for how they already work.

James Dice: Yeah. And that's the thing with the IDL is that the concept is new. And so it's coming into this world of full stack solutions and it's being inserted into the middle. And I like that philosophy as far as it's like, well, how do we make. Value happened today, given that full stack world. Let's talk about speaking of full-stack worlds.

Let's talk about the future. Let's kind of zoom out a little bit on the future of the industry as we kind of wind down here. The main thing I'm wondering about here is like, You talked about house of cards, sitting on a [00:38:00] foundation of sand, which I feel like is like a Coldplay lyric. Is that right?

I'll have to look that up, but um, where do you see like that foundation of sand going? Like, what's the future of that? As far as that technology progressing and becoming more structurally sound, if you will, to support all these applications that are sitting on top.

Andy Frank: Yeah, I think where my head is right now and it's been there, but then I think it was.

Influenced a lot by um, I read um, the healthy buildings um, Joe, Joe um, Thank you. And, and I think I put this on LinkedIn, but if you haven't read that book and you're listening to podcasts, you need to go stop and give, read that book. I think it was fantastic, great, and really influential.

And it, and it changed my perspective on buildings because. Yeah. Prior to that, I was, so it was always like, well, justifying technology is about cost reduction in energy production, and it's actually not true at all. It should [00:39:00] be about the occupants and the comfort and the productivity and their health and those things are implicitly more expensive.

And so it just changes your whole perspective of how I think you go to market. And so I think that that's where. When I look for the future it's that like, what do we want buildings to be? Right. And how do we get there? And I think healthy spaces, you know, lighting, thermal air, all that kind of stuff, contributes to that.

We need to make it simple. The example I use, it's not a great example, I guess, analogy. Well, I guess it works more than you think it does is cars, right? Cars are today are very, very complicated and they're almost an, almost a similar path as buildings, right? Because I've got per passing. Thermal customization, right?

Again, multi zone HPAC. I've got heated skis, cooling seats. I've got continuing more stringent admission standards I need to meet. And you know, you want the overall terms of the document to get better. You, haven't got a lot of driving. I've got, you know, nicer materials and more enjoyable, but lower noise floor, all that kind of stuff.

[00:40:00] And, but when you get into. there's a key, well, not a key. If you want, you press the button. Now it's a steering wheel throttle and the brake. I can get any car and I can just drive. I know how it works. Like why do we not have the technical pace and that level of simplicity in buildings?

And it's not a technology problem. I mean, we're, we've got people living in space right now on with reasonable rockets, but we can't make the west side of the building stay to say the temperature and the afternoons. I know we can do it. So I think part of mine is like, well, how do we. How did we get there, but how do we, how do we create the pathways to where we're actually enabling smart people to do smart things, to actually make the offices we spend in?

I think Alan's comment was that we spend 90% of our time indoors. We're indoor species. Why do we not pay any more attention to making that space more comfortable and more enjoyable? And. Those are the macro ideas that I'm thinking about. And then in the background, it's almost secondary, which sounds bad as technology company, secondary of [00:41:00] that is how do we buy technology that enables that, right?

And how do we, how do we build the ecosystem for the right people to come together? You know, how do we break down all of these stakeholders and these competing incentives? So the key split and symptoms as the term, Joe uses to get everyone motivated in the same direction to sort of accomplish anything.

typical. Long-winded all of it.

James Dice: I think a piece of that and yeah, I love the book too. I also posted on LinkedIn about the book a little bit. I was a little bit harsh on the book and it got, I'll put that in the show notes as well. Well, I, I feel like that I don't want to turn this into a roasting session, but it's a great book.

Everyone should read it. I just feel like it looks at indoor air quality and its own silo. And I'll link to the post in the book. Yes. Indoor air quality is extremely important, but I think we've had this kind of rightfully so because of the pandemic, We started the view into air quality as like the outcome that we're trying to optimize [00:42:00] for.

Whereas really in any given building, there are many outcomes to optimize for. And it's just one, one thing. Right? Okay.

Andy Frank: Yeah. There, there is. I mean, technically I think he's got some came in and was denying step framework. There is a whole bunch of other stuff contributes to that, but yes, I agree. The emphasis is on air quality, which is probably.

More weight than it needs to. But I think the concept of healthy buildings is something it's reframed how I think about a lot of stuff. Oh, me

James Dice: too. Yeah. Yeah. It was, it was the, like the main basis. So you took our foundations course. We have an entire module. Well, you took the first cohort it's

Andy Frank: evolved since then either.


James Dice: no, we're not sticklers about homework in the course it's but, but we built a lot of our business case, like the foundation of our business case module off of that, the arguments made in that book. So it is really foundational. [00:43:00] It really changes how you think about the. So, let me ask you that question again, though, from a technology perspective in the building, and I'm thinking about the silos, the traditional silos, how do you see those?

How do they need to evolve to, like you said, stay in line with where things are going. Technology-wise like, what's going to happen to the, the BMS, the access control system moving forward.

Andy Frank: I find it super annoying that, well, first of all, anyone knows me. Nothing's ever good enough. Everything's crap. Impossible to please. Um, I guess that's the product designer in me. Like, I've got a brand new space here, I guess this coworking space in Richmond where I am that I put up. And again, the HVAC Sox has always in the wrong temperatures.

It's freezing cold in the winter and it's bakes me in the summer, but it was in 2021. But like why is there not just a theory that it gives to the sees all of us? I mean, not me as a tenant necessarily, [00:44:00] but it's all like 10 different systems and the vendors aren't going to like get together and be like, well, let's all cooperate and do this.

So that's why I think I've really wrap my head around the true role. I think, of NMSI that there needs to be a sort of the single point of contact they ever see is all the technology and buildings. And they're really wanting to drive some of this forward. I think when they get leveraged for vendors to cooperate to some degree but it also, I think.

the big opportunity in buildings as the sort of build to know us. And there's a couple of companies doing this today. And I think that's what I see that, you know, I just, you know, look at these, even in a car or app or Macalester on a computer, right. There's just something that goes in there.

That's why I interact with, I don't care all the subsystems underneath that. And there's just consistency there. My, and I it's something I even toyed about doing, you know, two years ago it was like, well, do I want to try to tackle that problem? The problem that I, and the reason I didn't was, cause I don't know how you justify it to building owners.

You get, I think that problem will get solved. Maybe it wasn't smart enough to figure it out. But I think then I think it's wrapped up in that [00:45:00] whole healthy building thing. Right? Well, we just need, we have to have higher expectations to the spaces that we live in is what drives some of that stuff.

simplicity, simplicity, simplicity, simplicity, everything. And that's nothing simple. So

James Dice: you're picturing like an overlay software layer that is now replacing like the supervisory layer and today's building systems and their, their actual controllers are much simpler and actually do less.

And a lot more of the intelligence actually comes from the overlay. Is that what you're saying?

Andy Frank: I think it's probably one new ones that in practice. So I don't think it puts it necessarily goes away, but I think we, you know, we hide it progressively with layers, but I do think from you look at Zillow control systems and buildings.

Absolutely. There's no reason. And I think this is ridiculous. And people will beat me up after this. I know I'm gonna get kind of like hate mail that I go and custom commission this project as a snowflake. And I'm like, they'll work the same. I've been here, your handler, a chiller plan or whatever, or, you know, got an E unitary control or [00:46:00] something, but it's not the same thing I got, I got to work heater, heating coils, cooling coils.

I got spaces. Why the hell am I redoing that every single time, like that should be automated and smashed. Right. You know, maybe you don't think it's a gap created or good enough, but there's guidance there from the front of the math to make it work. Like that should just work. It's just put something in books.

It's ridiculous that we're actually custom programming these things because, and that's why probably buildings don't look that great because people get around. Right. Or the building changes, the environment changes around it. And then as things sort of drift and then they're not working as well as I do anymore.

And so I think part of that is also that we get away from this sort of on-premise custom permissioning. We do something, I call safe mode, fallback mode. I don't know if it's a better term for it, but there's basically not. There's an operating mode when I don't have connectivity, the internet that the buildings can operate a certain.

Yeah, it's best to keep the temperature. 72 degrees, light management, occupancy, all kinds of crap. But I think the logic for how we do those things is exactly what it meant. Matt Schwartz. And you [00:47:00] guys talked about last year with the advancing rights control. I mean, I think that is where things start to go.

We've got a cloud platform. It manages all that. And then they do an hourly, honestly, if I do it daily, I do a daily model downloaded to the building for how I wish would run the next day. Right. Weather patterns and all kinds of stuff that would be too difficult to probably do across a portfolio with the house buildings.

And one you've got that locals relation. There's one place where everything's stored, modeled updated, but also you start to have all of the data sets in one place to. I think that's where stuff gets really cool, because now you can look at across like different temperature zones and regions and operating performance.

It's like, that's how you tease out the algorithms and the optimizations to start applying to these things across. It can be more predictive in some cases. I think that's a really cool area that's yet to happen, but I think it's going to happen. So

James Dice: cool. I agree. All right, let's close this. Close this nerd session down.

Thanks for coming on the show. What are you looking forward to this year?

Andy Frank: Progress. I, you know, I think I say this to every [00:48:00] single the next bro, like, And he'd been saying, stuff's going to get better for 20, 30 years. You probably before I was in the industry. And I think there's a melting pot currently happening from, I think one, the age of people entering the space and the importance they're putting on, on buildings and healthy spaces, all that kind of stuff.

And how people have grown up also. I can't remember who told me this, but it's stuck in my mind ever since might've been Brian Turner the kids have grown up with iPhones and technology and they have a different expectations of how they interact with everything and that input, that energy entering the space.

You know, the. The, the influx of prop tech and more traditional software companies. And then, you know, groups like nexus where we're also pulling altogether. I think this is probably the most opportune times since I've been around that something is going to be meaningfully better for technology in the next five years.

So I'm optimistic that that happens. And that's sort of what my goal is, is that no, then to do our part, to help [00:49:00] shepherd. In some shape or form,

James Dice: I love that. Obviously that's what I'm trying to be a part of as well. And so I think that's a great place to, to end off. Thanks so much, Andy, for coming on the show.

Andy Frank: Yeah. Appreciate it. Thanks man.

James Dice: 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 subscribe at nexuslabs.online. You can find the show notes for this conversation there as well. Have a great day.