Watch this episode on YouTube
In this special partner episode, Clifton Harness of TestFit joins the podcast to talk about the founding story of the company, the decision to pursue technology over traditional architecture, the missing elements in the design process that led to the development of TestFit, TestFit as a co-creation tool, the impact of AI on site selection, design optimization, and pro forma valuation, TestFit's unique approach and value proposition, the difference between generative design and TestFit's optimization solver, benefits for players in a development deal, multifamily market trends, the recent investment from Prologis, upcoming developments for the TestFit platform, and an upcoming webinar on AI in real estate and industrial real estate.
- TestFit website
- TestFit on LinkedIn
- TestFit on Twitter
- TestFit on Instagram
- Clifton Harness on LinkedIn
- Clifton Harness on Twitter
- Webinar: It’s Time to Get AI-Ready in Real Estate Development
- About the webinar: Join TestFit and Prologis for a discussion on leveraging artificial intelligence (AI) in the world of real estate development
- Date and time: October 26, 2023 at 12:00pm CST
- Click here to register
- Previous episodes of TRXL featuring Clifton Harness:
- Episodes featuring TestFit people:
Connect with Evan:
132: ‘Breaking All The Rules’, with Clifton Harness
[00:00:00] Welcome back to the TRXL podcast. I'm Evan Troxel. In this episode, I welcome Clifton Harness back to the podcast. He's the co-founder and CEO of TestFit. But of course that's not where his journey in our industry began. Clifton has been obsessed with building things from a very early age. During his education while obtaining his bachelor of architecture from the university of Texas at Austin, he was the first B.Arch to obtain certifications from the McCombs school of business in business and in real estate. And after graduation, Clifton joined streetlights residential, a Dallas-based real estate development company. Prior to starting TestFit, Clifton worked with college roommate, Ryan Greige after work and on weekends and their work would become Residential Engine, which was the precursor to TestFit. Clifton is based in Dallas where he continues [00:01:00] his passion for improving AECO with technology, which we get into quite a bit in today's episode. In this special partner episode, we cover the founding story of the company, the decision to pursue technology over traditional architecture, the missing elements of the design process that led to the development of TestFit, TestFit as a co-creation tool, the impact of AI on site selection, design optimization, and proforma valuation, TestFit's unique approach and value proposition, the difference between generative design and TestFit's optimization solver, benefits for players in a development deal, multi-family market trends, the recent investment from Prologis, upcoming developments for the TestFit platform, and an upcoming webinar on AI in real estate and industrial real estate. The webinar is October 26th at noon central time, which we didn't mention in the episode, so I'm mentioning it here now. Again, that's October 26th [00:02:00] at noon central standard time. There is a link in the show notes to register for the webinar. This was a great, wide- ranging conversation with Clifton, and I want to express my sincere thanks to TestFit for partnering with me to make this episode possible. So without further ado. I bring you my conversation with Clifton Harness.
Evan Troxel: well, welcome back. It's, it's great to see you again. It's been a while since we've talked.
Clifton Harness: Yeah. Really has. I think Ian, maybe it was the last time, wasn't it? Yeah.
Evan Troxel: That's right. There was a combo episode, you and Ian. Uh, yeah, that was probably even a couple of years ago, year and a half ago at this point. I'll have to put a link to that one in the show notes, but yeah, it's been a while. So, and you guys have been on a tear. I've talked to several people, TestFitters. Uh, I've talked to Nat McDonald and Kyle Bernhardt. It's been, you know, [00:03:00] Flurry of development going on over there. Uh, and so I think maybe before we get into the, the topic of, of being AI ready, like what, let's talk about your story and just where you saw a hole in the market. I know you've, you've talked about this before, I remember joining onto one of your YouTube live streams when you were doing, uh, like a, a comparison, a straight across comparison of what you used to do in AutoCAD. Right? Laying out multifamily and, and, and you kind of, we just did like this live commentary session on YouTube where you just, you just let the stream run while you did the thing in AutoCAD and then you did it in TestFit. And I, I don't think I joined during the TestFit part. I think I joined during the AutoCAD part when you were laying out units and mirroring and copying and arraying and, and doing all that kind of stuff and
Clifton Harness: Yeah.
Evan Troxel: counting and making sure you have the right number of them and all this kind of stuff.
And, uh. I, I just remember [00:04:00] that being really fun, but, but talk about what the whole was that you saw, the pain that you saw in the industry to talk about how TestFit came to be.
Clifton Harness: yeah, so I've done the, I've done the founding story like a million times, you
Evan Troxel: I know.
Clifton Harness: it's, like, how do you, what's the right flavor, you know, and, um, I love talking on Troxel because I get to talk to architects, you know, in a way that I, I don't really get to anywhere else. And, um, so I'll go back to some of my first internships in architecture and kind of.
Yeah. You know, what responsibilities do you give an intern? Well, uh, not, not many, not, not deep responsibilities. And oftentimes it's going to be in, you know, schematic design or concept design, even where the stakes aren't that high. you know, I work for a firm for a summer and then I'd, I'd go back and get lunch with a partner or something, you know, during the fall or in the spring and say, you know, how, how did those projects pan out?
and I'm, I'm not a normal architecture [00:05:00] kid, you know, I'm, I'm fighting for internships like in high school, you know, with firms that are like, there's an email of me sending, uh, Ted Flato, uh, an email cold email, like, Hey, Ted, I'd love to work for you. I'm a architect and, you know, junior year of high school, you know, no skills or anything, but, um, You know, I say that to encourage, you know, young people start young, go, go network and, you know, you'll learn about projects and you'll learn the stories that, like how real projects to get born, like, like how they came to be and, um, I always pursued mentors that had, had successfully built, you know, big buildings and stuff and, um, and I'd work for them too.
And just the, the thing that always drove me nuts was like, we would throw away all of this work that we would spend, like, you know, the whole summer working on, like it was [00:06:00] some working for overland partners, you know, they get, they get to do these really fun projects and it was like a golf clubhouse and I spent.
You know, in a godly amount of hours, like laying out furniture and how you put, you know,
Evan Troxel: Great.
Clifton Harness: you know, um You know just didn't proceed and it it felt like such a downer, you know Even at these at these incredible firms we we get to put all the all our heart body mind and soul into design It just gets you know tossed So I felt that in the high design realm.
Um, and then when I worked my way down into, you know, the, the commodity world, um, you know, part of my story is my, my dad's a real estate developer. Um, he's a commodity guy. He's a garden apartment guy. Um, and so I grew up around that and I, you know, this is why I went to high design or in my early career is like, I was just like, I don't want to be a part of that.
That's crazy. Like, you're just slapping stuff up without, you know, Um, [00:07:00] You're not building contextually. You're not, you know, you're breaking all the rules, new urbanism that they're teaching me in school.
Evan Troxel: Yeah. Mm
Clifton Harness: But, you know, you eventually get to a place where it crystallizes in your head. Like, okay, there is waste happening, not just with architects, but like with developers and with the GCs and with the vendors and the manufacturers.
And, you know, if you want to get the right curtain wall done for a big building, the. faster you order that thing, the earlier you get it in the queue, the higher probability it's going to show up on site.
Evan Troxel: hmm. Yeah.
Clifton Harness: Um, so, yeah, you see the ability to reduce risk by making big decisions early and, and like staying committed to those decisions and the best of them, the guys that are doing the 150 million.
You know, towers that I was fortunate, fortunate enough to work for, [00:08:00] um, they make decisions very quickly and they have a very, very robust, uh, system for how they make decisions. And, you know, this is all in an effort to reduce risk and eliminate waste. Um, and they still spend, you know, millions a year on A& E fees on projects that don't move forward.
so yeah, that, that is the problem statement, man, is, is SD doesn't really pay the bills, you know, and if we're going to build an AI today, you know, let's attack something that isn't very profitable in firms.
Evan Troxel: So when you think back about this story, I mean, you're talking kind of about your personal story. Let's talk about the tech side of that story. So, I mean, a lot of interns are hired as CAD operators back in the day, or maybe BIM operators today, or maybe SketchUp. You know, drivers today, or whatever it is, whatever the tool of choice is for that office, and depending on [00:09:00] the phase of the project that they're working in, it might be all, it could be all three of those, right?
It could be one or the other or the other. Talk about what you experienced in that, because I think this also just helps paint the picture of where TestFit was born, right? And, and, and then ultimately why you decided to pursue a technology path in AEC versus, uh, you know, another role in a more traditional sense.
Clifton Harness: all right. So it's two questions, you know, I, I'm going to, I'm going to answer the first question. You're gonna have to remind me the second question. So the first question is really like. You know, the tools that I was using and, and how was I leveraging them, you know, and, and then, you know, maybe I'm going to infer the next bit, which is like, how did that influence what that, what the TestFit app looks like today?
Evan Troxel: Mm hmm.
Clifton Harness: Um, and then there's also the technical aspect was like, you know, I'm not a coder, but how did we get here? Right.
Evan Troxel: Yeah.
Clifton Harness: Um, and then there's the other half of it, which is, you know, [00:10:00] I'm not gonna remember the second half of the question. So let's, let's start with,
Evan Troxel: There we go. Start
Clifton Harness: yeah, let's start with the first bit. Um, so I, I learned that, uh, the more that you could leverage lots of kinds of tools, the more effective your workflows could be, and the more powerful you could get work done.
So, uh, AutoCAD for me was always the most effective 2D drafting tool to get, you know, Roughly the size of the thing that I needed, right? I'm not gonna use any other tool to do that, but I'm not gonna use AutoCAD to create 3d massing So you'd export that and you go into Sketchup and you'd build out your massing tool and then somewhere along the way, you know, someone at Autodesk decided to put like a Tabulation Lisp routine that Kind of works, but doesn't really work, um, with using, you know, polylines.
And so, you know, you [00:11:00] figure out this really archaic way of extracting all the data you need into a, into an Excel table. Uh, and then you're doing the same thing in SketchUp. You're kind of checking what the AutoCAD math was for the SketchUp math. And so you have Excel. So really, I, you know, my 3D modeling tool is SketchUp.
My 2D modeling tool was, was AutoCAD. And then, you know, my database was Excel. Um, and when, when I was communicating some of these problems to Ryan Greig, my co founder, CTO, college roommate, you know, groomsman, we're, we're, uh, almost neighbors now, you know, my buddy. Um. He was like, this is great, you know, okay, you got all these components, you kind of distill it down, you've got like a 2d thing, a 3d thing and a data thing.
Um, and so that's what we endeavored to build was, you know, the, kind of the strengths of those three apps without all of like the extra [00:12:00] garbage that came with it. and that's what we built initially that, you know, had. Sort of 2D interaction that was reminiscent of AutoCAD. And then, then 3D, it was sort of this push pull thing from, from, uh, you know, inspired by SketchUp.
Um, but then, you know, we started getting our, our customer feedback of our own. And, you know, it was like. Thanks for generating a building for me, but I would like to edit this crappy thing that you generated for me, you know, like this thing is 80% of what I want. I just want to do, you know, minor adjustments to this thing.
So, you know, then it was like, all right, what, what, you know, we like the block grip. You know, tool in, uh, in, in, uh, AutoCAD, so we'll put some grips and, you know, the sort of manual editing tool and, you know, now there's Illustrator and Figma, there's all these other tools out there that you can point to like who have really great grips, you know, [00:13:00] so, uh, yeah, we're just, we're not like trying to reinvent the development process.
We're trying to take, you know, who has the best grip out there and we're going to try to do a grip, you know, um, And so we did this sort of manual edit tool and oh my gosh, architects, they loved it. Um, and when we finally got analytics in the app, you know, 40%, I think of, of generated models have edits done to them, you know, so it's, it's obviously got product market fit inside our, our ecosystem.
Um, so yeah, that, that's, you know, the architecture component, you know, that's like, thanks for generating this thing, but I really would like to manipulate, you know, at a lower level and craft, you know, what this thing is going to be, um, yeah. And there was a second half of that question, Evan. I don't quite remember what it
Evan Troxel: Well, I mean, before we get that, I think one of the things that you've talked about in the past, and I don't remember if it was on this show or if it was elsewhere, but I know you've brought it up a lot is that the idea of TestFit being [00:14:00] a co creation tool, which is exactly what you're talking about right now with editability and.
And you've talked about technology, technology's role. It's important that it stay a co creation tool and not just do all of the things for, you know, just, it's not just a push button solution, right?
Clifton Harness: Yeah. I mean, that, that's the, you know, in, in some of the worst possible cases, like. Some customers want a single click button, you know, that does everything. And you kind of ask them why,
Evan Troxel: Mm hmm. Mm
Clifton Harness: well, you know, I get really frustrated when the developer says this, and then I don't know about this.
And the money comes in and the pricing, and I just want everything to, you know, I just want to manage the design process or, or something. Um. And the way that I read that is like the design process is literally so overwhelming to be placed [00:15:00] on the shoulders of one person and that's the architect these days, like, the level of constraints that you have being poured onto you.
It's, you know, it's not just building code. It's not zoning code. There's other things that, that you have to consider on top of that, that are, are not objective, you know, like there, there's design, like the basic component to design is that you can't really, you know, objectively describe the subjective of the building, you know?
Um, and so, uh, do I think that, you know, Using AI or co creation or any sort of automation tools to assist in that process absolutely is valuable, but what is different, I think, from how TestFit's approaching AI is that, like, we're never just gonna let the AI tell us what the building needs to be, like, you can click a button, you can generate a [00:16:00] building from a preset, but That's to get the user started.
It's not to say this is, you know, what TestFit has found to be the highest and best use. Like, an AI cannot do that. Humans are the only judge of what highest and best use really is.
Evan Troxel: there, there's a component to architecture that you're talking about here, which is like, what is space like? And so we're not talking about the envelope, we're not talking about the the program. We're not talking about the efficiency of the layout and, you know, the circulation and the corridor ratios to the, to this, this, and the that.
But we're talking about spatial quality and AI does not know spatial quality at all. Right? And, and that's where the experience and the wisdom of spending a long time in the profession of architecture and in school and, you know, that progression that humans have through that timeline starts to pay off at some point, and some are better at it than others.
Right? But that's the kind of thing that [00:17:00] you, it's this abstraction. It's, it's the architecture often is the thing that is not right. a component of the design. It is the space. It is, it is not the walls. It is not the floors. It is not the ceilings. It's not the roofs. It's not the structure. It's the thing in between all that.
And it is interesting to me to hear people say, I want a push button solution. It's like, well, the, you're not going to get architecture. necessarily out of a push button solution.
Clifton Harness: No,
Evan Troxel: comes from that editing process. It comes from that wisdom and that knowledge of what it's actually like to progress through space and the emotion and the mood and that, you know, solving those kind of, you know, those, those wicked design problems that are not something that is, that can be broken down into an
Clifton Harness: Yeah, I, you know, I'm gonna, uh, Overland Partners, again, is this firm in San Antonio, they're so instrumental in kind of my thinking about design, because That started working there just so early in my [00:18:00] life and like, they, they have these values for buildings and, uh, buildings should be strategic, systemic, sustainable, beautiful, meaningful, contextual, experiential, like buildings should be all of those things.
They should be. Right? Like, no one would fight, no one in the architecture profession would fight me when, you know, when I say, like, buildings should be that, right?
Evan Troxel: Right.
Clifton Harness: tools that we use to get there are, you know, trace paper? Really? Like, you know, we, we are so good at specializing in, in typology and morphology specific buildings, but we don't build on the history of what those, you know, what the designers before us have done with those buildings. And when I think about neural net AI, you know, that, that has just totally exploded since the last time we talked, you know, on, on this podcast, but, [00:19:00] you know, how, how is neural net AI and AEC? I, I don't know, like test fits. Procedural, we write with human intention, every single algorithm that goes into this thing, um, procedural, you know, human controlled, um, and we, we can, we can kind of guide our AIs to try to deliver the best possible building if we could, right?
But what I know about neural nets today is they train on existing material, right? They train on what is out in the world today. Okay, so if we train a neural net to design a building from what exists out in the world today, it's just going to be a thing that displays all of the compromises from all of the other buildings that were designed previously.
So, like, if a
Evan Troxel: designed incest, man.
Clifton Harness: Yeah.
Evan Troxel: You [00:20:00] just continually water it down, water it down, water it down.
Clifton Harness: Yeah. Yeah. So, so that, that, that kind of, you know, I'm, I'm not too, you know, I'm not too worried about the architect losing their, their position in this world with, with AI coming. Cause it, you know, like the strength of a designer is to bring clarity to a design problem with other humans in the room, you know, you know, there was a guy posting on LinkedIn today about how terrible that it is that test fit generates terrible looking buildings.
And I'm like, okay. Okay. Yeah, man, isn't it terrible that architects design and create terrible buildings every day too? Like, at least these are ones done virtually and 95% of them are thrown away. Like, you know, like we're saving the architect that time so they can focus on the 5%. They're gonna proceed. Um, and so that's the role of AI today.
Evan Troxel: I was just going to say, back to your point of attacking the problem where the, where the most [00:21:00] waste happens, right? Like that, you want to get to a, a spot where you can push it forward as fast as possible so that the chances go way up of it actually happening.
Clifton Harness: Yeah. Firm leadership today is mired in project feasibility and didn't really have time to care about process and stuff downstream. They're trying to win work.
Evan Troxel: Yeah.
Clifton Harness: Um, it's incredibly important to improve this part of the process. Like actually let me rephrase that. It's incredibly important to improve any part of the AEC design process if you can.
There's so much like meat on that bone that, you know, test fit for all intents and purposes. Like we've basically automated the drawing of parking. Like that's, that's our contribution, you know, to, to some of these firms. And it's a huge business. We've raised a lot of money, you know, like guys, like there's [00:22:00] other low hanging fruit out there that.
You know, other firm, you know, other startups can go attack.
Evan Troxel: Yeah. So my second question to you that I'm reminding you about now was just about choosing the path of going technology versus more traditional path in architecture. And you've already spoken to some of those, I think, pieces, but what really solidified it for you that this was how you were going to move forward in the AEC industry was actually developing software, raising money, building a team, selling it to people versus, you know, becoming an architect of buildings.
Clifton Harness: Yeah. Um, so I had it in my mind in early undergrad that I was going to be like, uh, Bjarke Ingles or, um, yeah, basically him. That was, that was my, you know, you know, just
Evan Troxel: aspiration.
Clifton Harness: starchitects, you know, uh, model. Um, but then when I got into practice, I started to realize that the. [00:23:00] The times where I was expending my intrinsic motivation, the, like the stuff that I wasn't either being carroted or sticked towards was building tools of my own, um, little calculators that helped, like in my first job, I was doing code review, uh, calculating, you know, the areas of rooms and, you know, getting the life safety yield and, just built a little calculator in Excel that made my job like 30 times faster.
And, you know, Basically went home early that day cause they didn't have enough work, you know, cause I get billed by the hour. So I made a tool that made myself obsolete. It's great. yeah, so I, moving into, into like actual practice, you know, I was working in development. I was writing list routines to generate unit layouts based off of like, uh.
bounding box. Um, you know, I was only allowed to have AutoCAD. So everybody's like, [00:24:00] Hey, why do you like, I was like, well, they didn't want me to use Revit. So I was a 2d workflow. Um, dynamic blocks. It's like this parametric system, in AutoCAD, that's rather powerful. If you're into 2d modeling and you don't need to use BIM.
so I was always. Really excited about the tools and making tools and, um, you know, I had kind of an 80, 20 rule that I could spend, you know, 20% of the time automating. And, you know, If I had spent all of Monday automating, then by Tuesday I had to start actually working on the project because it was due Friday.
Um, and so you, you eventually get to a place where your tools are so powerful that you're done with like all of your work by Tuesday. Um, and I, I got there at, at Streetlights. I mean, it was, it took a while, but every tool in the book I was using, every list [00:25:00] routine I could update. you know, 7, 000 blocks with the Lisp routine overnight.
you know, when I left, I think the, the guy I worked for was like, yeah, we had to hire four or five people to come in and figure out what you did. I was like, yeah, I told you, I was like, I warned you. yeah. So the reason to go to software is like, my passion was the tools. My passion wasn't to create.
You know, the architecture, like I think eventually I might find architecture to be my passion again, as a, as a person that is, you know, understands that it takes probably 30 years to really be a competent architect. Like this is my first, you know, 10 years is trying to automate parking. It's a great contribution to the industry.
Evan Troxel: It is. And so, let's just take a, like a, let's pause here and talk about, because, believe it or not, there are people who are listening to [00:26:00] this episode right now, and I'm also talking to certain audience members, who have never heard of TestFit before. And so, most people are going. Everybody's heard of TestFit, but I'm here to tell you they haven't.
So tell us exactly what TestFit, like at the current snapshot is, because I'm not even sure everybody knows what TestFit currently does. We all have kind of an idea of maybe what it's like. You've talked about parking. It's doing a lot of things now. So maybe you can just kind of break down at a high level overview of what the, what the tool's doing.
Clifton Harness: Yeah. Uh, so, uh, I'll, I'll do the practical, uh, impact statement and then I'll go into the sort of theory pitch maybe. Um, so we are a real estate feasibility tool for multifamily and industrial buildings, mostly in the United States or IBC compliant markets. Um, Our software generates site plans for those building [00:27:00] types, um, for those, uh, morphologies, you know, industrials, you know, got 15 different subtypes of buildings that like warehouses are actually fairly robust and the kinds of things that they do.
Um, and then in multifamily, uh, we can generate. You know, probably from, I'd say probably four units to the acres, probably the lowest sort of density, uh, up to, you know, 500 units, the acre towers. Um, and we might be the, the first to get to sort of product market fit and housing in the United States, uh, meaning that we've got, you know, enough AI that we can do all the different segments, uh, of, of density and multifamily, uh, that.
You know, housing firms can get a lot of use out of it at every scale, um, theory sort of pitch, uh, you know, our vision is that we'd like to, to reinvent the, the [00:28:00] real estate development process, um, for the, you know, for the betterment of our future communities, uh, we think that land isn't particularly valued as highly as it should be, um, and I think through, uh, years of effort, you know, and sort of.
improving the AI behind highest and best use, uh, we'll be able to give humans a better idea about what they can do with land earlier and with more clarity.
Evan Troxel: So is that the driving force behind why you want to reinvent this part of the real estate market? I mean, because to me, it seems like, like architecture, like a lot of AEC, it hasn't really, truly been disrupted. I think now with AI, people are actually questioning because there's, there's fear that we're going to be disrupted in a major way.
And so, I mean, is that, does that make sense to you?
Clifton Harness: I mean, you know, you want to bring disruption to AEC, change the OAC contract, you know, like,
Evan Troxel: Mm [00:29:00] hmm. Mm
Clifton Harness: you know, there's, there's so little margin in AEC that software vendors aren't, aren't going to try to go attack, you know, how you're making profit. Um, you know, it, the real, the real play here is to get the capital, to get the, the debt and equity excited about BIM, excited about, you know, The process part of AEC and how we can reduce risk through technology, how we can actually start to string together supply chains and more, you know, prefabrication and lower carbon, like, you know, if anything, what TESFIT can do is to try to bring a window, um, to all of our problems in AEC to other stakeholders.
And that's how we've been able to monetize by, you know, working with real estate and working with their architects. To find, you know, a common ground where they both would get a [00:30:00] win. Um, I would be worried, like, you know, if I'm an architect in today's world, like I would be worried about AI, but I also would be worried about AI if I was in any other industry, um, you know, TestFit's got a lot of funding, but I'm worried about AI too. Like there's people that are getting funding for ways to generate buildings that aren't the way that, that we are particularly effective at. So, um. You know, so that's interesting to see, you know, what's happening, um, with AI, even in, even in the startup space.
Um, I don't think there's going to be a large language model for buildings, honestly, like how many petabytes of images were, were used to train mid journey compared petabytes you know, the pro core construction cloud. I, I bet there is. An order [00:31:00] of magnitude, maybe two orders of magnitude more data, uh, about images, you know, just clip art images than there is, you know, construction data.
So, we, we simply don't have the quantity, um, but perhaps test fits AI, you know, procedural AI could start to generate. test data that could be used, um, to fill in the gaps between, you know, building A and building B. Um, but it's a, it's a big hairy up, up, upward, you know, battle to, to even get an AI. You know, like I've spent six years on high density and we're still finding out ways to make, you know, make the building more, more efficient, more compact and squeeze out, you know, gross area.
And, um, you know, and it's taken hundreds of customers and thousands of feedback calls. And, um, you know, do I think that [00:32:00] AI can pose a threat to that easily? Absolutely not.
Evan Troxel: was gonna ask you about that thing that you just talked about, which was, you know, you're looking for more efficiencies. You're looking for ways to squeeze things down. And my question is going to be how, and then you started to answer that right now with talking with customers, getting their feedback.
And so talk about that a little bit, just because it's not like you just come with the answer. Like TestFit is not just the answer and it stays the answer and it's never going to change. You're constantly updating and manipulating the model. So that it's addressing the feedback that you're getting. So maybe you could talk a little bit about how that's working.
Clifton Harness: Yeah. So, I mean, every, every new line of code that we add changes the app ever so slightly. Um, so if you think 2017 TestFit was, uh, a very, it basically only thought that Texas was the only place that you could make buildings, you know, like that was what we had was a [00:33:00] thing that made donuts. Right. Um, and so you add capability.
Okay. You know, we got it. Customer in New Jersey, you know how many podiums they build in New Jersey? Almost everything is a podium in New Jersey. So, you know, we learn about. You know, water retention problems and, you know, needing space for pumps and, you know, there there's, you know, customers in the Midwest, like, still, they have snow load problems and how they want to keep their garages, like, you know, TestFit still doesn't fully address part of that problem for them yet.
So it, it's, you know, we're fortunate to be in the U S because in the U S you have 50 states that all have these subcultures of sub problems. And, you know, we're fortunate to have, I think, customers in 46 states now, um, you know, across every, you know, all architecture, all development, we got everybody, um, big, small, medium, and their [00:34:00] feedback is pouring into this, you know, one AI that does.
Multifamily. Um, yeah, so that's the feedback is customers, man. Like, like we, we didn't know that much, you know, in 2017, but, um, when you have six years of feedback and the, you know, Trello cards that make your eyes want to bleed because you haven't been able to do that specific customer feedback in over five years, um, you have a pretty good idea of what your, what your destination looks like from a product roadmap.
Evan Troxel: Yeah. Right. Yeah, roadmap. So, I remember when we chatted at AU, and you were talking to architects, and now you've, I think you've seriously pivoted to real estate, not just design, right, from an architect's kind of... Yeah, so talk about adoption, because for TestFit, you just said you're in 46 states, you're in a lot of firms, you're in a lot of development offices, I assume, but [00:35:00] big picture, would you say that the, uh, adoption curve is doing well, or, or is there just a huge opportunity there still, because I'm wondering if they're seeing tech serving them well, at this point or not.
And, and I would assume not because that's kind of what you're aiming for.
Clifton Harness: Yeah. So, so, okay. Um, it's kind of funny cause, cause last time we were talking, you were, you had a stable job and I was the startup guy. So now we're, now we're totally, we're totally flipped, I think.
Evan Troxel: Right.
Clifton Harness: product market fit maybe is like the right way to answer this. It's like, if you focus solely on a repeatable sale. Like this is, this is not going to be like a fun startup thing to hear for people that want to go do a startup, right? Like this is the [00:36:00] dirty part, right?
Evan Troxel: Real talk. Yeah.
Clifton Harness: yeah, real talk. Um, you know, if you want to be successful, you have to figure out a repeatable sales process.
Like that is, that's it. You know, like this is why inferior technology wins all the time. It's like, cause some idiot in
Evan Troxel: part out.
Clifton Harness: had an inferior technology, but figured out a repeatable sales process for how to. You know, how to bring a firm in and, you know, respect their, their problems and show, okay, this is the automation that we can provide.
This is our value. This is how we've been selling it. And this is, you know, and we've been able to scale that. Right. Um, I think test fits four and a half times as expensive as, uh, as, uh, Forma and we're, you know, probably have more customers in the U S I don't, you know, they're, I don't, you know, there's.
What other, what other competitor, you know there, there's not much out there right now. Um, so when I say things like, you know, [00:37:00] any person could just arrive in AEC with something meaningful and put their head down for four years and grind on it and keep their focus. Like, we never moved away from multifamily until industrial.
And that was mid 2020 when industrial was exploding. Like that market went from 2% to 8% in a year. Um, So just saying, staying laser focused on providing like that specific value that you can get that repeatable sales process. You can create that income stream and then you can go talk to VCs and say, Hey, I've mitigated the risk of this company.
You know, can, can you help me grow faster?
and it boils down to that repeatable sales process in every case. I mean, like what's the biggest exit we've had in AEC in the last 10 years? Was it a, Oh, it was the iPad.
What is the name of that company I would ask bought for 800 million? building
Evan Troxel: I know what [00:38:00] you're talking about. Yeah. It was for, for doing , punch lists. Right. The name is escaping me as well, but,
Clifton Harness: Tracy young. All right. So Tracy, Tracy founded this company and, you know, you go, you go on LinkedIn, you follow Tracy, all she talks about now is how to build sales teams and how to make repeatable sales process. And she built this company, you know, it's like to view PDFs of CDs on the iPad and sold it to Autodesk for, It's You know, 800 million.
It's like, okay, value prop is simple. We're doing this thing and we figured out a way to sell it to people. And in their case, they talked to foreman people on the job site that actually needed to look at construction documents and they sold it to those people.
Evan Troxel: Yeah. Architects doing construction administration as well, right? I
mean, it was, it's That phase of the project. It's like, those are the people at the table in those construction trailer meetings every Tuesday morning at 9, right? Going through all the issues and walking the site together.
Clifton Harness: just amazing. Right? Like.
Evan Troxel: [00:39:00] So, talk about the way that. TestFit is using AI because I, I don't remember when you started, I mean, you never used the term AI. You talked about algorithms, right? You talked about algorithms all the time. And, and that was kind of TestFit's secret sauce was because one thing that you was always had in your favor is how fast TestFit solves, right?
And, and I mean, that probably speaks to Ryan specifically and, and his skillset and his talent. And so maybe you can talk about kind of your I'm going to use a, a startup word, pivot to AI, uh, but, but just talk about that because those are, those are, you know, I, I, they're, they're a little bit synonymous and they're sometimes they're at odds with each other and I think the, these words leave room for the imagination of how to interpret these.
Clifton Harness: Uh, yeah, we, we pivoted and I'm using air quotes here for, for those that are, you know, uh, to, to talking about AI when, uh, [00:40:00] everybody else started talking about AI. So this is our, our hope is to just not get forgotten. Right? Like. And I don't think we will be, like, it's going to be a while for any of these large language models to do anything like what we're capable of doing today.
And, yeah, so that's why we're talking about AI is like, we're, we are the cutting edge for AEC right now. And if you're going to pay attention to what Midjourney can do for you, well, how about a thing that draws parking spaces, you know, like, you know, one's, one is, uh, one is a bad replacement for Enscape and the other is something you might actually need.
Evan Troxel: Yeah. Yeah, totally. so let's talk about your product overview and how you're solving the pain of your customers. Because I think when people don't have TestFit, they do things a certain way and then they get TestFit and it's a different way of working, right? But, but I, I would imagine you get a lot of [00:41:00] aha moments when that happens.
Can you talk about, that?
Clifton Harness: yeah, the aha moments. Um, you know, there's, you know, people stand up in the meeting room and shout and walk out and yeah, I don't know. there's
Evan Troxel: High fives?
Clifton Harness: Yeah, there's, there's been all, all kinds, you know, sending us memes, you know, inappropriate memes about wanting more of that and where that came from.
Um, we've seen a lot of reactions to, to seeing the thing in action. Um, let's talk about 20, about 2018, there, there was this. Idea that you had to do generative design, you had to generate millions of options
Evan Troxel: Mm hmm,
Clifton Harness: um, like you had to saturate the solution space with all possible solutions.
And, um, it took, you know, if you want a hundred percent of the solutions, it would take, you know, a thousand times longer to create a hundred percent of the solutions. Right. Well, like to, to Ryan's credit, like he. He was like, that's stupid, but we can show [00:42:00] 80% of the options and like 2% of the amount of time, like we really can, and you can pre compute it and you can do all this kind of stuff to, to get to a place where you can do that.
And that was interesting because I was like, all right, so you're saying like, maybe we can actually show like the algorithm in real time in action, you know, like if I drag a setback. You know, the building's going to move along with it. Right. Um, he's like, well, I don't know, I wouldn't go that far. Right.
Like, um, but we were able to, uh, to actually achieve that. And you know, the, the thing is like, if you can do that, you can show in real time what the algorithm's doing in a way that if you generated it frame by frame and they had to wait for it, you couldn't, so. You know, I kind of equate it to like, if you're able to do it in real time, you can get trust in the tool because the user kind of understands what the [00:43:00] AI is trying to do.
Um, and so that leads to, I think the title of our, of our first podcast was like "BIM is a T-Rex", you know, it's like, like, in this case, like TestFit's kind of a T Rex, you gotta like climb on it and you get. a couple of parameters you can flip around and that thing's going to, you know, ram its head as hard as it can into the problem and see if it can crack it.
Right. But, uh, you still might have to flex a little bit of the parameters yourself, uh, to get to that right product market fit. Um, you know, product market fit from a real estate development standpoint, like.
Evan Troxel: Yeah, right.
Clifton Harness: a, do a, does a tower make sense here compared to, you know, podium deal or whatever.
Evan Troxel: Mm hmm. Mm
Clifton Harness: So yeah, riding the algorithm, you know, instead of riding the T Rex, you got to like figure out what's the most effective way to get as far as you can.
Evan Troxel: that aspect of real time, it's come up here with, you know, people from Enscape who have been on the podcast [00:44:00] with you guys, where the element of real time has introduced a way to help make decisions along the way, and it encourages by, by encouraging the user to play with it. I think at that point, cause it's like, Ooh, it's stretchy.
What, what does that mean? And you're immediately seeing the results. It's not like you stretch it and then walk away and wait like we used to do. And we hit the render button, right? Which was okay. Now we're going to wait hours for the rendering to occur. Now it's real time and we can actually make decisions using that.
What used to be a final output tool as a decision making design tool along the way. And I think that is a paradigm shift. And so when you're going into, you know, a real estate development company, or you're going into an architecture firm, is that one of the big aha moments? Are people getting that right away?
Clifton Harness: Yeah, I mean, you know, it doesn't, it doesn't take the, the development guys more than, you know, [00:45:00] 30 seconds to understand that they can get design information out, you know, out of a polygon in 10 seconds versus two weeks like that. That's the latency issue for them. It's like they, they want to test really crappy sites that they're not going to send to their architects.
Evan Troxel: right?
Clifton Harness: Because they're really low, like, you know, low probability of success, but they still want to test it because it's part of their risk reduction strategy, right? Um, you know, the architecture side, it takes a little bit longer because they want to understand that, that we've taken time to care about the, the nuts and bolts of what goes into a multifamily building.
And in some cases we, we really have done a good job of doing that. High density, for example, it's. You know, the gold standard solver, like every other solver out there should compare itself to TestFit high density and say, are we doing what this thing does? And even at TestFit, I do that. I'm like, Garden is not [00:46:00] as good as that one yet.
Like, what's, where's the disparity, you know, or why is there a discrepancy? Um, and there's good reasons, you know, like we have six years of development on one and the other one we have three, you know, so there's, there's just a, it takes a while to get these things to be absolutely incredible. But, you know, it's.
Just like with any masterpiece or work of art, like, you're not going to build it overnight. You know, it takes a while.
Evan Troxel: the decade long overnight success, right?
Clifton Harness: Yeah. Yeah, exactly. Yeah.
Evan Troxel: It is interesting to think about the, the real time aspect of it and the decision making part of it. And I think one thing that I don't think you've brought it up here, maybe I spaced out for a minute, but like the, how you marry the proforma side of things to the visual side of things is a key piece of helping people make decisions because it's not just what it looks like.
It's not just how it, environmentally sits on the site. Like you talked about, you can, you bring in a site, it makes a site plan. You can bring in the [00:47:00] context of whatever area you're building in and see it in real time, but you're also marrying that to the data output side. I mean, just to kind of round out the product offering here, can you talk a little bit about that and how that enables people to make decisions?
Clifton Harness: Yeah. So, so, uh, think about a building. My mental model is it's like a three person tug of war, you know, there's architects on one end that's going to advocate for, for high, you know, better design. It's got to be a good building. Um, oftentimes that means it's going to be more expensive, right? So they're going to pull in the direction of, you know, expense.
GC, You know, uh, there's another component to the architect in a moment and I'll, I'll, I'll talk about that. The GC is going to just talk about the hard cost. Um, you know, it's going to cost you this much and then the developer is going to talk about the cost, but he's also going to talk about what kind of net income it can bring.
And when you work in some of like the [00:48:00] best architectural teams, they'll figure out how to equate like a incredible amenity space to, you know, positive on the rent dollars or something, you know, they're putting it in terms that a developer can understand. So, you know, in some case, you'll have an architect.
That's really talking about how this thing's going to make money and make the deal work and stuff. Uh, so. My mental model, you got, you know, three kind of competing groups fighting for, you know, this kind of income, expend this kind of thing here, um, in the way that, you know, the, the contract structured is, you know, the developers, the owner, most of the time.
And so he's, he's kind of kind of win the argument, but, um, the good ones respect the architects, in my opinion, I've seen it happen in real time, like where an architect will speak up and the developer will be like, Oh yeah, you're right. That's, you know, they put their ego, they keep their ego in check.
Evan Troxel: kind of goes back to that idea that we were talking about earlier, too, with the understanding the spatial [00:49:00] quality and the user experience side of things from an architectural point of view, and then weaving that into the, the future value that the proforma can or can't say, but, but you're like, you're the architect is augmenting that narrative at that point.
Clifton Harness: Yeah. I mean, practically speaking in TestFit we're calculating the size of the building as we go. So we know what that NOI, the net operating income is going to be, because we know that the developer is going to assume this will be, you know, 2 and 50 cents or, or whatever. Um, the developers already plugged that information in.
And so. The goal on the design side is to achieve enough density that they yield on cost, you know, that it's actually worth it to do, to do a project. and right now, what we have in, in multifamily in the commodity space, you know, it's interest rates are crazy high. Um, you're going to see the multifamily industry slow, if not stop, you're going to see a lot of these [00:50:00] commodity markets kind of.
You know, grind to a slow, I'm not going to say it's going to stop. It's going to grind really slowly. And the kind of deals that will get done are the ones that are really, really big, uh, because they only want to focus their effort on, you know, a certain size deal that's going to be really big. Uh, and they only have so much money to play with.
And, you know, so TestFit's product market fit. Like we focused in the early days on big, big, big buildings, you know. Like stuff that capital will continue to build regardless of a recession. Um, so yeah, I mean, it's, you're going to have all of that stuff modeled in the app to, to sort of show what the trade offs are.
Um, but the, the way that you read it is it's a yield on cost. It's, you know, this is what this building can make in a year. Uh, which isn't really a metric that, that architects are, you know, are going to follow. [00:51:00] Um, what I will say though is, is the most effective metric that you can use to improve yield on cost is like a building efficiency metric where you're getting more rentable area than, than you had previously with the same building.
Evan Troxel: so that goes back to the algorithm and, and the, the feedback that you're incorporating into the model to constantly make that better, which I would think then pushes that, that idea up higher to, you know, people's awareness at that point.
Clifton Harness: Yeah, it's, it's interesting to sit in the room where we've got everybody at the table and we do like a charrette with TestFit and, you know, the GC is like, that's going to be way more expensive. And then, you know. Developers like add another, you know, everybody's fighting at the, the kind of operator of the model to, to flex a thing to, you know, like, we're going to have terrible views, flip it the other way.
And then, you know, then the GC is like, the crane isn't going to fit on that side. We got to flip the tower the other way. [00:52:00] Um, and just to see that level of decision making happen. In 15 minutes is absolutely absurd. Cause I've sat through, you know, email chains or it's
Evan Troxel: Oh yeah.
Clifton Harness: weeks long. And
Evan Troxel: Yeah. Yeah, totally. That's, that's crazy.
Clifton Harness: so yes, real time, we are aligned with Enscape. Real time is the future. You respect the user and the ecosystem's time.
Evan Troxel: And you have a, an integration with Enscape. So
Clifton Harness: We do.
Evan Troxel: Yeah. So it's not just shaded models. It's, you can do rendered real time walkthroughs as well of, of TestFit massing and kind of cool stuff.
Clifton Harness: Yes.
Evan Troxel: So you talked a little bit about funding earlier on that you're well funded, but I've been talking to people on your team and, and they've hinted that there's been a major customer investment that's happened in TestFit.
Clifton Harness: Yeah. Yeah. The, uh, industrial, the industrial world. So, [00:53:00] um, yeah, that market that went from 2% to 8% was industrial. And these are, you know, warehouses, like, you know, it's, it's so crazy to me, Evan, that I spent. six years, like fighting for this, you know, product to be built for high density multifamily. And, um, you know, it's, it's a market where nobody has more than like two or 3% market share and it's totally regional.
And, you know, that's why the app is powerful in the way it is. Um, and then. When we decided to go look at industrial, it's a market that's, that's kind of new and growing. Um, these are, these are warehouses, like these are, this is not architect, you know, I mean, some of it's architecture, but like, Honestly, like it's, you know, it, they're, they're located further away from like civilization because, you know, eight, like the visitors that go to these buildings are like 18 wheelers you know?
Evan Troxel: Yeah.
Clifton Harness: Um, [00:54:00] and okay. So we were kind of learning about this problem and we have this, you know, supply chain problem in the U S too. So, okay. There's kind of a story here. Uh, and then we met, Uh, a venture kind of group that works at Prologis, which is, uh, the largest real estate investment trust on, on the planet.
Uh, so this is basically a publicly traded, um, real estate fund. So like you can go buy shares of Prologis if you want. Um, and they're, they're, they're deploying, you know, billions of dollars of. You know, money a year to buy these warehouses and build them and develop them. And, you know, so it, it's really kind of turned into a.
Foundational business for us, like, in the same way that our multifamily business is, you know, being, being foundational. and we've been able to, uh, kind of, they've been an investor. Uh, they're now a customer. Um, [00:55:00] and I think they're, they're, uh. You know, their use of the app is, is absolutely kind of, kind of absurd.
Um, if the hard, you know, they've got the design team in house, design manager and all this stuff that it's actually kind of incredible when you look at it, you're like, yeah, we've, I wish we had that and, you know, multifamily, but, um, it's been a real class and like, here's how a publicly traded company functions and it's been a, it's been a challenge for us to rise to the occasion.
I think we've done quite well. Um, and what you'll see this fall is, I mean, I'm sure we're going to talk about it, um, more publicly, but, um, a fair amount of tools that we've built specifically for our enterprise customer Prologis, we are going to release to everybody. Um, and so, yeah, I mean, I think we have a webinar in October or something. Um,
Evan Troxel: You have a webinar titled [00:56:00] it's time to get AI ready in real estate development. So we can make the call to action for this episode about it. seeing that. So maybe you can talk a little bit about it.
Clifton Harness: yes. Uh, so we're going to, we're going to talk a little bit more openly, talk a little bit about roadmap. Uh, we don't really like to talk about roadmaps. So it's going to be a rare opportunity to kind of peek behind some of the curtain at least. Um, but TestFit is, uh, going to be cloud enabled is the one thing that I will, um, let y'all know.
Evan Troxel: Nice. and what does that mean? I mean, are, what are you willing to say now that, that that might mean to somebody? Because Well, when Kyle was on the episode, we talked and teased a little bit about Rex at that point, and you guys had talked, have talked about, at that point, and since then, pieces of TestFit.
working on the cloud, like the, the collaboration parts that make sense, especially if people are kind of tying into a model remotely. And, somebody wants to fiddle with the numbers while somebody else is [00:57:00] fiddling with the model and, and all of those kinds of things. So, so what does this mean in, in terms of changes in how TestFit works?
Clifton Harness: Yeah. So we, we built everything into a monolithic thing. Right. So, uh, Ryan and I were 25 when we started the company and we started building this monolithic thing. Right. We are 32 now. We've hired people that are way better at all this. They know the right way to architect the stuff, make it scalable for enterprise and, um, ultimately to make Testament a much more powerful, open, you know, more dynamic platform in the long run.
Um, so. What we're doing now is we're kind of unpackaging components of the app that don't really belong in the configurator and those will be, you know, a component on the web and, um, one example of that is like, uh, perform integrations. You know, we have a architect running test fit and he also has parameters that he can flip to for Proforma.
And we're just going to create a, [00:58:00] like a web page for a developer to access that deal. So you can add his, uh, assumptions in real time while they're working on a deal together. So stuff like that, where the sort of real estate execution model that, you know, talks about how we collaborate. You know, as architect developer and GC in these, these early phases, uh, we'll, we'll start to come to fruition,
Evan Troxel: So at this webinar for getting AI ready in real estate, who's the target audience for this webinar?
Clifton Harness: uh, architects, uh, real estate development and, uh, GCs for sure.
Evan Troxel: All right. So we'll have a link to that. I'm sure you'll have a landing page on your website where people can sign up for that. And they can follow you, Clifton, on LinkedIn. They can go to testfit. io. We'll have links to all that in the show notes. Is there anything else that you want to mention during this, uh, Troxel therapy session that [00:59:00] you've just attended?
Clifton Harness: Uh, not really. I mean, look, I, I think that, um, I've been incredibly blessed to like be in this industry for, for this long and been building TestFit for this long. Right. I just, like, I want to encourage people, like, you can do this too. We need more startups in AEC. We need more opportunities for VCs to invest.
Um, we need, you know, more spin outs from firms that have really cool technology that could be democratized elsewhere. Um, you know, it's not feasible for, you know, TestFit and HYPAR to fight for all of the innovation. You know, like, we need a lot more. And I'm not, uh, discounting the literally hundreds of other startups out there.
I'm just saying we need like, we need more, like we need like a horde, like we need like a, like a blue wave of startups or something if to use a political term.[01:00:00]
Evan Troxel: I was thinking like Lord of the Rings when you said Horde. Like I just see this massive crowd simulation, like charging the
Clifton Harness: yeah. If every BIM manager rage quitted and got together and made like the ultimate. You know, here's how we do it, you know, that kind of thing. I'd be all about that. That, that, that would be the kind of insurrection that would get the, the blood pumping in the firm leadership, you know?
Evan Troxel: That's a good call to action right there, man. I love it. I love it. Well, Clifton, this has been a fun conversation and I, I'm so excited about all of the development that I've seen watching from the sidelines over the last few years since we first talked at AU 2018, you know, over a cup of coffee and, uh, just to watch your guys' success.
Obviously I'm a huge, uh, I'm rooting for you. I'm, I'm, I'm a big time supporter and, uh, I'm, I'm excited about where things are going, so thanks for taking the time to hang out and chat today.
Clifton Harness: Yeah. Thanks Evan. I've been really encouraged by our relationship
Evan Troxel: thanks, man. Until next time.
Clifton Harness: Till next time.[01:01:00]