Moving from Static Data to Decision Velocity in Commercial Real Estate

Introduction to Data Challenges in Real Estate

Drew: Welcome back to the Peak Property Performance Podcast. I am your co-host, Drew Hall, as always. And real quickly before we introduce everybody, today's theme is going to be the move from static data to decision velocity and the real power that comes behind that. So before we make introductions, I just want to remind listeners to like, subscribe, share the podcast. This is, again, how we're spreading the word on the Peak Property Performance movement and changing the industry. So also if you think you'd provide value as a guest here with us, we welcome commercial real estate thought leaders from all stages of ownership and operation. And of course, please reach out to us via the PPP website, that's peakpropertyperformance.com or find us on LinkedIn as well. And all those links are shared amongst all these different platforms as well. So feel free to look us up there. Look forward to the potential of working with you in this endeavor.

Drew: So without further ado, Bill Douglas, co-host, welcome.

Bill: Thanks, Drew. I'm getting to where I really, really, really look forward to these shows. They're so much fun and we get to have great conversations with cool people. Like today's guest is Dr. Wayne Geary. Wayne is a PhD in data with a background spanning CoStar and now Numark, where he's working at the intersection of brokerage, analytics, and what's actually changing on the ground with AI and logical, I mean, excuse me, location intelligence. What we're going to dig in today is pretty simple, but it's a problem most of the industry still hasn't solved. And that's kind of why we wrote the book too. So we have more data than ever, but very little of it is actually usable at the point of decision. So the reason I'm excited about Wayne is he is in a unique position where he's seen the limitations of market data, he's seen how fast AI is changing analysis, and he's also seeing where it breaks when operators don't actually control their own data. So this should be a practical conversation, less about tools, more about how decisions actually get made, where they go wrong. So we look forward to you sharing some stories, Wayne. Welcome to the show. Happy to have you.

Dr. Wayne Geary: Oh, happy to be here. Yeah, welcome.

Drew: All right. So Wayne, let's start off with that notion of lagging data and why it happens, why it's a problem. So where have you seen market data fail to reflect the real conditions in time to make good decisions?

Dr. Wayne Geary: Well, because we rely on other companies to collect that data. As a person who works in a real estate company, that can be problematic on its own, given that they each have their own lag time. And so, you know, let's talk about rent rates there, 30 to 90 days already out of date when you get that information. And other kinds of companies as well that serve up even demographic data, they each have updates at different times and different ways. And so as we try to serve our clients on getting to the right location, for example, we're using a lot of times other people's data, but I think that's changing. I think we're moving to wanting to really collect our own benchmark data, which will sort of fix the problem.

Drew: Well, and I think you started to allude to it, but, you know, as a result of this data lag that we're talking about, you know, maybe the asset gets mispriced or the leasing strategy is off one way or the other, or just the timing is not right for buy, sell, whatever. What types of decisions are most exposed that you see where that lagging data changes the outcome?

Dr. Wayne Geary: Well, we could get into more of the real estate side or the business or operational side on the lagging data, because let me just give an example of the operational side. If we're putting a firm needs labor, putting them in a location and we don't have the most up-to-date competitive information and for that matter, facility location data, we're really in trouble in advising our clients on where to go. If there's already, if it's a warehouse distribution center, for example, I'm looking for workers, we may not have the most up-to-date information on who the competitors are in the facilities that are there. And again, I think we'll talk about how that is changing with technology, but certainly they could go into a facility and all of a sudden there's a competition for labor and they're in trouble. They can't operate without the people they need to do their business.

Leveraging Geo AI for Rapid Decision Making

Bill: Wayne, you and I were talking a few weeks ago and you mentioned geo AI. Yeah. Elaborate a little bit on what you're able to do in hours now that used to take days.

Dr. Wayne Geary: Yeah. I mean, so AI, we've all used geographical information systems. We like the maps. The maps really help us with putting those in front of clients and telling that story about the location. But it's not always an accurate story because again, we're taking data that has a lag factor, we're mapping it and presenting it to the client. Well, that's changing quite a bit with the introduction of AI and geo or GIS. We can use remote sensing and we can look back historically in the remote sensing or satellite data and we can start to measure change over time. And change over time is going to tell us construction, for example, and it's going to focus specifically on not only construction in, let's say, a DFW market. It's going to focus on specific types of assets being constructed in those markets. And we're able to get a much more accurate assessment of what's available in the market, at least from a facility point of view.

Bill: Well, you were excited about the speed of this process changing the status quo. So where does the speed actually change the decision versus just making the process faster?

Dr. Wayne Geary: The speed comes in using AI bots to look over. We can train them to look at anything, any change over time. And once that model is trained, it does it in seconds. It tells us where construction is. And if we look temporally, let's say last two years, we can see, or last few months, we're able to get the imagery from any of the satellite companies. You can get it up to the day, change from one day to another, or you can get change from a week. You're going to pay a little bit of money from one day to another, but you can get every week or every month change.

The Role of Speed in Real Estate Decisions

Bill: So the speed you're talking about is not just the speed to process, it's the speed to current. Like there's no lag in it.

Dr. Wayne Geary: No lag at all. No. So the speed is actually two levers. I was only thinking of how fast you could get your decision versus taking four days before. Now it takes an hour, not decision, but to get the data.

Bill: Yeah. No, there's no lag now. Now for that data. Do you have a story you can share, maybe a moment where speed either won a deal or avoided a bad one for a client?

Dr. Wayne Geary: Well, yeah. I mean, I've been using this data for a long time, right? We can, environmental data, for example, yes. Okay. Putting another manufacturing facility in for, in the trading card business. So the manufacturer of trading card. The client wants to know, you know, if I put, if I stay in the South where there's an abundant amount of labor and I have a facility in Dallas, for example, and I need a new facility somewhere else, how, what is the impact on risk in this area? And this, this speaks to AI. We can look back at the weather. We can start to project what weather risks there is. But people think when you see, like you did this morning, storms strolling through Dallas at this time of the year and going eastward and in a straight line almost, right? We see that happen all the time. They're going, well, that's a bad place because it's, you know, if we stay in the South, we're going to get hit by those storms. Well, no, when we have a GOAI, we can look back at the data for all the storms that have come across in the last 10 years during that time. And once, and I calculated the answer for the, for the client and it was one in 200 million chance that two facilities would get hit. It's just not going to happen. They have two manufacturing centers and it's going to roll straight across in a straight line and, you know, include risk to all those facilities in that straight line.

Bill: Follow on question. Did that reduce their insurance rates? Any, if they, if you shared that with the underwriter, would that impact premiums?

Dr. Wayne Geary: Oh, that gets, that's not my expertise area, but I would, would, you know,

Bill: You think it would?

Dr. Wayne Geary: I stopped mine either, but we do try to, to use data to reduce insurance rates. Yeah, of course. I've worked for insurance clients where, you know, building a hail model, for example, where, where is hail going to occur? Where's there a high propensity for hail to occur in a, in a market? And that was for them to set up locations to predict, be predictive about where we know in the future, we're going to have to provide people to the community for hail damage.

Bill: Oh, that's more like a business continuity or a community assistance program.

Dr. Wayne Geary: Yeah. I like those.

Bill: Okay.

Drew: Yeah, that's interesting too, because I mean, it's obvious about the safety of the facility itself, whether the facility can operate, but I would think too, there's that factor of getting people to and from, you know, I mean, that's, we, you mentioned Dallas, Texas as an example, I spent some time there myself. So not just the safety of the building itself, but the safety of getting all the people in and out in order to do the jobs.

Dr. Wayne Geary: That's true. And, and what people, COVID changed drive times too. And we need, we use drive time modeling now for people going to and from work. And, you know, a city like New York building could be two blocks away from each other, but they, when we run the model about how much time it takes to get to work, it's very different for the two locations, depending on where the train stop is, how long to walk from the train stop to work in the winter. How does that factor? What distance do you need to offer a hybrid policy for work for your workers? So with AI now, all of that gets cut. It's not just a drive time map anymore. It's an actual model that tells us about what the best location will be to operate their business.

Data Ownership and Ethical AI Use

Drew: Yeah. Well, let's, let's think about that. Like the, the data ownership, like the gap in the ownership of the data itself. We talked about that, I think a little bit at the top and how that's changing. What kinds of things break when an owner tries to connect their internal property data with external analytics?

Dr. Wayne Geary: Well, I, it's almost the other way around. You, so with the introduction.

Dr. Wayne Geary: Of AI, you've got a ton of people just taking it, taking, you know, they're using it. Are they using it responsibly? And are they dropping in the right data from the right sources? That can be a problem. So I think that the, you know, the spin around AI will solve all your problems. I mean, I could get into why that's not true, but no matter what you do, the broker is still going to be extraordinarily important in the process. Whether AI is there or not, you can't eliminate their, you know, their expertise, and I can get into why and what studies were done of that, starting with Kasparov, the chess player.

Drew: So yeah, go for it. Give us a little bit of that. Go for it.

Dr. Wayne Geary: Okay. So Kasparov goes and plays this machine in 1997, the blue machine, I think it was called, and Big Blue. Yeah. He loses. And so from then on, he gets really interested in machines and people and how they possibly could work together. And they come up, he comes up with this theory called Centaur. It has to be half machine, half person. You can't, even MIT later does studies on this, and it shows that in the end, the best model is that expert-based system using the machine to automate those things for time to get to a quicker decision, you know, because we all know the hustle and bustle of corporate real estate. It's like, you know, I needed that decision yesterday.

Breaking Down Data Silos with AI

Bill: Do you see data silos inside buildings limiting real operational decision-making today? Like, you know, the property management systems, building automation management systems, so many vendor control platforms out there in any particular building.

Dr. Wayne Geary: Yeah. I don't think that companies are, you know, real estate companies are learning from that right now. I mean, I know we're changing the way we look at that issue. We want data, we want a data lake, for example, that's good for all assets. Instead of 10 data lakes or 10 access points to sources and data, it's the only way also that AI will work properly is that you have that lake to pull the data from. Even if we just think about labor and access to labor analytics, you know, every one of the assets needs to understand what the labor market is for a location that they'll be in.

Drew: Yeah. Yeah. And in the spirit of examples, does an example come to mind about, especially on the positive side of that, where, you know, maybe there is an effective data lake or something like that, at least at the beginning phases of it, where acting on that data was a positive experience?

Dr. Wayne Geary: Yeah. I mean, we, you know, we built a data lake and A, we're able to respond more quickly. And with a extremely advanced, could be a site selection model, we can do drive time to work modeling that, you know, right now doesn't just say where, you know, this building has less attrition than that building would if you, if we had our employees coming into one location versus another. And we're able to, you know, quickly, you know, run a model, talks about attrition, talks about policy wise for impact to HR. So we're able to get much more out of it when we're using this one source and it's consistent. So when we go from one asset in office to the same, same client as an office requirement, the next requirement is a manufacturing facility. We're using that same lake, the same, and very similar methodology and process within the AI, between the AI plus the broker expert to determine what the best solution is for the client.

Bill: That's cool. Well, Wayne, you brought something up a minute ago, and I know you and I have talked about it before. So if we can bring it back up and elaborate a little more on how AI is not the expert and why human context still wins, right?

Human Expertise in AI-Driven Real Estate

Drew: Okay. I know you're passionate about that. So where have you seen AI produce something that looked right, but wasn't actually usable in the real world?

Dr. Wayne Geary: So if I, if I were not going to put anything in, and I just said, show me the best location in Dallas to put a warehouse in that we'll have access to all the, all the packers that I need for my distribution center. And it picks some different locations, but when I test it later on with a more accurate data from a data source, there's two, really only two data sources for a good labor data in the country. And one is Jobs EQ by Chamura and the other was Emsi. And we use Jobs EQ, for example. When we use that data with, you know, other data layers, drop those into AI, and then we're able to just get a more accurate model out of it on top of what the broker has written for the client, what the needs are. So the broker's statement and their experience and what they're looking for and what they think the client needs is, creates a highly accurate output. And it's because of the broker knowledge, their experience, they know they have relationships, they've spent years asking the right questions. It's just, you can't model that. It's impossible.

Drew: I completely agree. So you gave a perfect example of when human override changed the outcome, right?

Bill: Yeah. How do you train staff at either your client or your brokerage to not distrust AI, but only trust it 60 to 80%? Like it'll speed it up, it'll weed out some noise, but you, I mean, we do this at our company, you still have to do the thinking. The AI is the thought partner, not the thought leader. So how do you do that inside of your world at Newmark?

Dr. Wayne Geary: Well, first of all, we've been working with AI for the last three or four years. And so we've had a lot of time to look at all the options, all the companies that, and they do different things. Some do things better than other. And when you, when you think about chat GPT, it's not quite as accurate on certain things as cloud might be. So training our employees to understand that, also having some rules about access, what data we can put into that AI, you've got to be very careful. We don't, we're still testing out what that means. And also you can tell instantly almost if somebody went through and they should edit anything that comes out of AI, right? You want to, you should read it, edit it. And you know, they haven't, when you see that em dash in there and no one has taken out the em dash and, or replaced, you know, cause an em dash will be put in where there should be a comma or maybe a dash, but a lot of times it's just the comma.

Drew: Or those dreaded lines across the page.

Dr. Wayne Geary: Yeah. We go out of our way to share examples of, I did this and this is what it told me. And this is what I really wanted. Like it's, hallucinations are real. We can't, we can program machines to reply really quickly, but we can't program to think like humans do. We can't mimic the conscious brain yet. That's why I'm not fearful that AI is going to rule the world, at least not in our lifetime. But we use it as a tool and a thought partner is awesome that the people that are using it as the truth scare me.

Bill: Yeah, no, it's, it should all, you know, as an academic, if I were to write a paper, for example, it would happen very quickly and it would actually go and do the, look up all the sources. It would create the, you know, stick them into the document, but they need to be checked. I mean, there's been, you know, they, they do need to be checked. I have checked a couple of sources and those sources didn't exist. So yeah.

Drew: We got a pattern in the back a little bit in our book. Go ahead, Drew. I'm sorry.

Bill: No, I was going to say it's confirming and humbling at the same time, because I'm sure most, if not all of us have done this at this point where you feed something back in and ask, say, chat GPT, what's the likelihood that this came from chat GPT and the thoroughness of that answer and, you know, the certainty like, probably did. Here's why. Here's the top six reasons why it probably did. So that on top of the fact that like, especially if you know who that's coming from, you can say, hmm, interesting, but not really accurate in the overall sense.

Dr. Wayne Geary: Yeah. It's revealing. It was quite the process and we published the book, Fast Company Press published it and they're adamant this can't be an AI book, can't be written by AI and they validated every piece of data in the book. I mean, like we made up the two fictional characters so that clients didn't have to be shared, like the exact client wasn't shared, but all the data in there is true. And they were adamant, like, did you write this? Like, let me see the transcript and, you know, you would talk it and then write it and then, you know, transcribe it and they would not allow an AI book. And that was beautiful. It was painful, but the result was beautiful. It would have been easy to say, write this book and you get 150 pages in five minutes, but that did not happen.

AI's Limitations and Real-World Applications

Drew: No, that, I mean, that's a good point. And it's interesting what's going on in universities too. I mean, I think they should be creating a foundation for how to use AI, but they're not exactly focused on that. And so, you know, all they're saying is you can't use AI. I mean, you know, we have kids in universities and, you know, it's, you know, it's pretty tempting to go and use AI to write a paper for political science, but they should be, I think they should be allowed to use it to a certain degree, but then they have to sort of demonstrate how to add your own thought and build your own argument into the final piece of work.

Bill: Yeah. I think it depends on what they're studying. Yeah. My older, my younger son was a aerospace engineer, so AI didn't really do him any good other than to help him learn. Julie's son's taking engineering and he's like, we're not only, we're not allowed to use it, it doesn't help me take a test. I still have to know how to do it.

Drew: Right. And the middle one is an art student and he's like, it doesn't help me a bit. Like AI is not going to make, you know, unless he's going to go into graphic arts and you can talk to it and change it on the screen, but he still has to make art.

Dr. Wayne Geary: Yeah. So there are a lot of things that are immune to that, to AI and AI will be a great tool. Maybe it'll speed them up now, but it's not going to replace them. I'm not fearful that it's going to replace everybody.

Bill: Oh, I'm not. I'm not at all fearful that it'll replace anybody, but I think that there has to be some sort of foundational course that today.

Dr. Wayne Geary: I think it will impact in the long run, the way we use geographical information systems in real estate. And from the point of view, AI can actually do things like point and polygon, these geocomputational processes that you buy software for. It can actually do that now. And so it's going to change the way we create maps and apps, and it's going to combine those two things together very quickly.

Personal Insights and Career Advice

Drew: Progress. Right. Well, Wayne. Yeah, for sure. For sure. So Wayne, the final thing that we do in wrap up here is called the extra floor, which is just a short series of questions here that are, it's not conversational, just, you know, gut level responses, short little answers, just so that our listeners can get to know you a little bit. So here we go. Question one, what is the best piece of career or life advice that you've ever received?

Dr. Wayne Geary: First job I ever had, I worked for Roger Stavok and his company. And the best career advice he ever gave me was at that time, you know, a young person starting off my career in real estate was make your boss look good. And so that was what I did and learned about not going. It's me that created that it's, you know, it's our team and it's the direction of our, you know, the person in charge. So it was great advice.

Drew: What's one habit or practice that consistently makes you more effective?

Dr. Wayne Geary: Not thinking that I know everything as a PhD and learning from others.

Drew: Wow. That is humble. I love it. That's funny. Yeah. Yeah. Without outing anyone, this is a not quite immediate family member. That's all I'll say. Was saying to her spouse, who was a PhD after he'd gone a long time about something, she said, I love how you pontificate from a position of ignorance. And I never forgot that. I thought that was so funny because I thought what he said was great. These were spouses of a very long time. Oh my gosh. That's probably been 20 years and it stuck with me. I thought it was so funny.

Dr. Wayne Geary: Yeah. I think the reason why he told me that is because you can't just lead. You need to be aware, you know, from Roger's point of view, you need to be aware of who your team is and how they can help you succeed. So yeah.

Drew: All right. Well, third and final question here on the extra floor, you consider yourself an early bird or a night owl?

Dr. Wayne Geary: Night owl.

Drew: That was solid. And has that changed over time? Curious.

Dr. Wayne Geary: It has changed over time. It used to be early bird and now it's a night owl.

Drew: Yeah. Wayne, we put your contact information in the show notes, but for the people driving or running or listening to this while they're moving, how can our listeners contact you if they want to reach out?

Dr. Wayne Geary: And through LinkedIn, certainly they can connect with me there or at wgeary at gmail.com. And that's g-e-a-r-e-y.

Drew: Yes. Okay. Well, thank you, Wayne. And thank you to our listeners. And as Drew said at the beginning of the show, be sure to follow, like, subscribe, share, if you think somebody would be a great guest, have them reach out to us, we're just trying to spread the conversation about how data and digital are transforming commercial real estate. So thanks everyone. And we'll see you on the next episode.