This week on The Home Builder Digital Marketing Podcast, Will Zhang of OpenHouse.ai joins Greg and Kevin to discuss how home builders can use artificial intelligence tools to better understand and predict home buyer behavior.
The home building industry is unpredictable, so insights into potential trends can be instrumental in helping home builders determine effective strategies. Will says, “Home building is a very, very, very fascinating industry, but there's a lot of risk-taking the home builder has to take. You have your construction capacity you have to consider. You have to really do a lot of forward planning.”
Artificial intelligence tools can forecast future home buying tendencies based on present home buyer activity. Will says, “... a different approach to do it is that by simply looking at what customers look at right now, then we can actually make a prediction about what's happening in the three to six months. It's a simple idea. Because what people are looking at right now is actually what they will buy in three to six months.”
AI focused on home builders will make home builder digital marketing and sales teams more effective and successful. Will explains, “Very quickly, probably in the next six to 12 months everything will have some kind of AI and everything will be branded with some kind of AI…how is it relevant to my job? I think AI is just a tool and they're going to enhance the tools that I already have. So, it's almost like in the past, I need to use a piece of paper to do the calculation and if I need to take a square root of 3 million so it is 1, 2, 3, 4 going to be difficult, but today I can just pull out the calculator to do it. So, I think AI would be a similar concept. It's just actually going to help you to do your job better.”
Listen to this week’s podcast to learn more about how AI can help home builders foresee potential home buying tendencies.
About the Guest:
As CEO of OpenHouse.ai, Will pioneers the home building industry's transformation through advanced AI solutions, democratizing data-driven decisions for home builders and land developers. Under his leadership, OpenHouse.ai empowers production home builders to adeptly manage demand fluctuations, streamline sales and construction processes, and optimize resource allocation.
With an extensive 17-year technology research and development background, Will is a highly-regarded expert in AI, Operations Research, and Software Engineering. Demonstrating exceptional adaptability and a comprehensive understanding of the industry, he held various executive positions in public and private organizations throughout Europe and North America. Will received his Master of Engineering from Université Grenoble Alpes in France and his Master's degree in Business from NYU Stern School of Business.
Greg Bray: [00:00:00] Hello everybody, and welcome to today's episode of The Home Builder Digital Marketing Podcast. I'm Greg Bray with Blue Tangerine.
Kevin Weitzel: And I'm Kevin Weitzel with Zonda Livabl.
Greg Bray: And we are excited to have joining us on the show today, Will Zhang. Will is the CEO of OpenHouse.ai. Welcome Will, thanks for being with us.
Will Zhang: Thank you for having me.
Greg Bray: Well, Will, let's start out and just help people get to know you a little bit and tell us that brief introduction about yourself.
Will Zhang: Yeah, for sure. My name is Will. I'm the CEO of OpenHouse.ai. So, we are a company for home builders. My background is in technology [00:01:00] development across many public and private company. For my 17 years career, I have developed and deployed various AI system for a large property company in Europe and North America, across finance, telecom, utility before I come to home building industry. We've been very proud of helping some of the four local home builder to help create value for them using AI for the last five years.
Kevin Weitzel: All right, before we get into all this nerdy stuff, that Greg's just gonna be smiling ear to ear because Kevin, the Neanderthal is gonna be going duh about. Anyway, could you tell us, Will, something about you personally that has nothing to do with work that people can learn about you on our podcast?
Will Zhang: Yeah, for sure. So, most people actually don't know that I'm actually a ski instructor and I've been teaching for 10 years. My oldest student was 60 years old and my youngest student was my son. He is three years old. I personally learned a lot from that experience. One can apply the similar [00:02:00] first principle analysis, like a typical engineer to help people understand their own challenges. Then, I can help them to overcome those challenges regardless the level of skill and that was very fulfilling and exciting.
Kevin Weitzel: Number one. I did just learn that about you personally. I've known you for several years now and I did not know that. We're talking snow, of course, right? Cuz you're up in Canada. Not water.
Will Zhang: Yes. Yeah, Western Canada.
Kevin Weitzel: So, are you a Rossignol or Elan? What's your ski brand of choice?
Will Zhang: I'm a Rossignol.
Kevin Weitzel: That's the only way to go.
Greg Bray: And Will, where's your favorite place to go ski?
Will Zhang: Sunshine and Lake Louis is the two ski hill that's just one hour and a half away from Calgary. So yeah, it's actually, it's beautiful town here. We have five world-class ski hill just in our backyard. So, you guys come by and let me know. I'll take you guys up there.
Greg Bray: My skiing experience did not go well and I've been traumatized, so I'm not sure I'll ever do it again, but that's a story for another day.
Will Zhang: That's actually exactly what I'm passionate to fix. So Greg, I [00:03:00] definitely would love to be able to go on the hill with you.
Greg Bray: All right. we'll have to look into that and see if you can fix me. So, there we go. Well, Will tell us a little bit more about OpenHouse.ai and the kinds of services and products that you guys are offering builders today.
Will Zhang: OpenHouse.ai is a data science company for home builders. We've been around for a few years. So, actually, the best way to describe is home builder, our customer told us that, we help them to go from hoping, wishing, and praying to actually knowing. So, what does that mean is we help home builder figure out the market, what the market want and need, where to build, what to build when to build them. To do this, you need to actually deeply understanding your customer. We achieve that by harnessing the existing marketing data and we also deploy a digital sales tool to help them to collect those insights.
Greg Bray: So, tell us a little bit more about how you decided you wanted to be, first of all, in software engineering, and then how you decided to go and apply it to home building versus [00:04:00] some of the other places you could have gone.
Will Zhang: So, how I got into the technology is, I actually did my degree in software engineering. Right after the internet bubble, it was actually quite challenging time at the time. I'm just very fascinated and interested in the extracting and knowing what is unknown, extracting pattern from data. So, that's how I got into career of AI today. But I actually been working on that for 17 years. I started to actually realize that the potential for AI and technology when actually working for like Electricity of France, telecom, and finance. Actually, there's a lot of intrinsic pattern that you actually can extract from it.
Roughly five years ago, I had a chance to decide where my career can go. And I realized I'm actually more interested in solving problem that are related to the fundamental need of people, such as shelter, food, and clothing, mobility. And I think making positive impact to provide those fundamental need is really what got me [00:05:00] interested in the home building industry. It's an industry that's very important, 5% GDP to the economy, and really really solve the fundamental need of shelter. So, I think that's very passionate about that.
Kevin Weitzel: Can we take one step back to just kind of help our audience understand exactly what we're talking about and how OpenHouse.ai helps home builders? So, I know for a fact that you have an overlay for like OutHouse interactive floor plans. I know you have an overlay for the ANewgo platform. You have an overlay for home builders’ websites, but is there any other specificity that a home builder could utilize you on?
Will Zhang: Yeah. So, what we basically deploy is a abstract layer of actually basically some algorithm to understand what the consumer looking for. So, the reason why we overlay all these technologies is actually we recognize that the value of this existing technology investment for the home builder is actually a tremendous value, but something not really known to them. Like, even just [00:06:00] take Google Analytic for a simple take is that there's a tremendous value that can one can extract from those data.
So, if I may use an analogy, I would love to use Amazon when they deliver as an analogy. So, if you think about how Amazon, when you order a mouse or USB cable and they're gonna ship to you. Same-day delivery depends on where you are or next-day delivery. But when was that mouse was made? It was probably produced and manufactured three to six months ago, shipped from somewhere even on the other side of the continent.
Now, how can Amazon actually ship that to you? It's actually the way that they do it, they understand your buying pattern for each individual. Then they actually make prediction what people will buy. So, by the time a customer actually make the decision to order it, most likely there's actually already somewhere in nearby warehouse, already pre-stocked it.
The same technology actually have been tested and tried and true and fulfill many, many years, many other industry. And that data asset is actually sitting right in your Google Analytics, your OutHouse, and direct floor plan. All [00:07:00] those data assets, it just sitting there. So, what we actually realize is that we have the potential to apply those technology to help home builder to deeply understand what the consumer needs just by leveraging the first step from the existing dataset they have. So, that's how we do it.
Kevin Weitzel: So, in all reality, what you're doing is you're setting it up to where I could go to my front door and there'll already be an Amazon box there, and then I'll go, I was gonna order this in two or three weeks.
Will Zhang: That's exactly the magic we wanna create. Yeah. So, if you think about that, right? Home building is a very, very, very fascinating industry, but there's a lot of risk-taking the home builder has to take. You have your construction capacity you have to consider. You have to really do a lot of forward planning. If you have very accurate demand prediction, 90 days at a time, you actually can do way better planning. Your tray actually have way more certainty. But traditionally the home building operational team actually was not aware the tremendous data asset [00:08:00] that the marketing and sales team has to offer to them. So, we are here to basically bridge that gap. So, that's in a nutshell what we do.
Greg Bray: So Will, just as I understand it, we've had some conversations prior to this as well. You guys are taking the data from the website and from traffic into the sales model, and you're bringing all that together, and putting together models that say, Hey, sales are gonna go up in 90 days. So, you need to start planning this. Or, Hey, their sales may go down in three or four months, so let's ease off on how many inventory maybe we're putting out and some of those things, and being able to predict it.
And I'm fascinated by the idea that you're able to predict it just based on what people are doing on the builder's website and how many of people are doing certain activities and looking at different things and how valuable that analytics data becomes when you can use it in this way. It seems a little kind of alright, really? Can you really do that? Tell us how you proved it, that you [00:09:00] can really do it. If there's an easy way to do that without charts and graphs since we're just talking?
Will Zhang: Well, absolutely. And I'll do my best to do it. So, traditionally most forecast model use historical data to say, My sales are going well for the last three months, so probably I can consider is the trend gonna continue or not. So, that's a typical what we call analytical solution to make forecasting based on historical data, which is very valuable.
However, actually a different approach to do it is that by simply looking at what customer looking at right now, then we can actually make a prediction about what's happening in the three to six months. It's a simple idea. Because what people looking at right now is actually what they will buy in three to six months.
Traditionally, actually business owner and the CEO and operation team, they did not appreciate the value of web traffic. Marketing professional themselves, they all understand it. But the broader audience beyond the sales and marketing team in a home builder, they typically don't consider the value of it. But they do appreciate one thing is foot traffic and appointment, [00:10:00] right? So, oh, if I have how many foot traffic, how many appointment is more tangible for the construction and operation team? Web traffic become a bit more abstract.
But one thing that they did not recognize is that, well today, 99% of people start from the website, so that actually make web traffic a very good lead indicator to what actually happened in the foot traffic and the sales, and hence the construction. It's not black magic, it's fundamentally science. You actually understand what people are looking at and, well, if I look at certain home, then I stop moving forward. But most likely that product was not great or not good enough or mismatched to the need. If I look in the right product and I stopped progressing into the funnel, and then we can understand what really caused them to actually move forward. What problem are they actually facing in their living condition? And helping them to tie the loop to actually, that's how we make the prediction.
Kevin Weitzel: Wait a minute, Greg. I just wanna point out that Will did discredit [00:11:00] black magic. He said it's not black magic, but not anywhere in his statement did he discredit voodoo, that there isn't some level of voodoo in AI.
Will Zhang: I did not intend to describe Black Magic. Maybe it has his own place. I'm just saying that we're not using it.
Kevin Weitzel: Totally get it.
Greg Bray: Well, Will, all great scientists have been considered black magicians by the masses before they were suddenly understood. So, I think that's a historical something there, but. Help us understand then you're now taking this website traffic data but what about the quality of the website itself? Because, you know, garbage in, garbage out is one of those famous computer science phrases we get hammered in to these processes. If a builder's website isn't structured well, if it doesn't have the right kind of content, does it mess up your data? I mean, are there certain things that have to be there to make this data work, right?
Will Zhang: Yeah, well, absolutely. Actually, if it is a really poor quality website, it is actually fundamentally not only affecting the AI system, [00:12:00] it's actually affecting the fundamental business. So, we're not website developers, so I would encourage them to talk to the agency like you guys to be able to fix it. However, we are still able to make a difference for them no matter where you are. So, meaning if a builder have a great business, even the website was not great, that means fixing the website will unleash you how much more potential they can have. So, that's number one.
The second thing is that even with that, consumers still go into a website first or eventually go to website. So, whatever we can do, we actually are able to identify what are the good quality information from it. In a nutshell, this is actually what we do is our system, after a few years operation, we already have 8 million how people actually make a decision. So, we can actually take this traffic that you have and we can match the good quality data to our algorithm database. Then we basically can understand, okay, these are the strong signal of buy. We already passed the phase that [00:13:00] the small number of the full quality data we interviewed algorithm at this one.
Greg Bray: Hey everybody, this is Greg from Blue Tangerine and I just wanted to take a quick break to make sure you know about the upcoming Home Builder Digital Marketing Summit that Blue Tangerine is hosting together with OutHouse, October 18th and 19th in Denver, Colorado.
This is gonna be an amazing event full of digital marketing insights, knowledge, best practices, and most importantly, some fun. So, be sure that you get registered today and come hang out with us, an amazing team of speakers and presenters that are gonna be together. Again, that's October 18th and 19th in Denver, and you can learn more and get registered at buildermarketingsummit.com. We'll see you there.
So, let's talk about the AI piece, right? Which implies learning. How do you define AI, Will? When's a computer system AI versus just a nice tool?
Will Zhang: Very, very high level. I think AI, there's two major category. One is actually causing a lot of people to concern about, they call it artificial general intelligence, meaning a single AI system can solve many, many different problems, just like human brain. And the other class is what people call it weak AI, meaning the AI gonna solve a specific type of issue.
A lot of people are actually scary of is actually artificial general intelligence. There's a lot of debate in my field, in the AI practitioner field, is it really near or not? So, actually if you ask me when is it gonna be AGI, I think maybe in five years. So, potentially ask me in five years later, I'll give you the same answer. So, so I don't know. There's no theoretical breakthrough on understanding of the consciousness at this point.
But weak [00:14:00] AI, so if you think about ChatGPT has this remarkable capability. It is fascinating. I think the world gonna move very fast from here. However, I think it's still a class of weak AI, meaning it's actually doing some simple thing, try to predict the worst that actually gonna match your expectation. So, that's actually what it's doing. So, what it does compared to a traditional analytical model is it actually learn from your engagement. So, it actually produce some text and then learn from your response, and you have human feedback to make it better and better. The same concept actually had been applied for over two decade in many, many significant industry, just people just not knowing it. It was working behind the scene.
So, if you think about Google search, as simple as you punch a few keywords. It actually try to search the content from the website that actually then see if it's matching expectation. Every single click on the link is actually giving some feedback signal to Google. This, actually, I'm talking about [00:15:00] 1990s now, right? So, and then you use the same technique to apply to the ads deployment. So, that's Google AdWords. And the AB testing, the marketing field, right? So, which image is gonna be more effective? The difference is that it actually look at that outcome and then go back to see what was working or was not. The same concept had been applied also Amazon distributions network was done. Even Google Map is the same, right, when they actually optimizing route.
In my past experience involving all this industry, actually, I know how it worked, but technology was very inaccessible beyond this big tech. In the past you have to have your private data center and whatnot, but today, technology is increasingly accessible and until it finally surfaced to the public that you actually now can use ChatGPT interface to access those technology.
So, we actually take the same idea. We basically say, well, we provide something to the marketing people folks to say, well, your marketing attributable sales. If we actually can say, well, [00:16:00] that ads that being clicked, drive to that web traffic session, which eventually turn into sales. And then now you can go back to feed to the algorithm what, what did not work? Then algorithms start to optimizing it for you. So, that is actually what creating a amazing and hyper-personalized experience for consumer. And I think that will be the game for the home builder.
Greg Bray: So, I love the way that you broke that into kind of two groups, right? Because I think right now AI's all over the news. It's becoming dinner table conversation for people who had never heard of it a year ago, or at least never thought of it, outside of maybe a sci-fi movie type of scenario. And so now everybody's kind of, Oh my gosh. You know, when are the robots gonna take over? And that's, I think that one conscious level thing you were talking about.
Those are the robots that are gonna ruin the humanity as a whole and destroy the planet and all those kinds of things. And you just said that's five years away. Is that what I heard you say? No, I'm just kidding. So, um, but then what you called the weaker [00:17:00] AI is more of very targeted at solving a specific problem, and learning as it goes though by continuing to improve as more data is available, as more interactions happen and be able to improve the solution as it processes more information. So, Kevin, the robots are not coming yet. I just wanna make sure you're okay.
Kevin Weitzel: There's aspects in life where I want the robots to come. If AI and just all the interconnected cell phones can give you the red zones on a map, when you're going from point A to point B, why can't the phone movement, the GPS locators, figure out traffic and adjust traffic lights to accommodate better flow automatically. Because AI could do that. AI absolutely could, but the systems that run traffic lights are archaic, yada, yada, yada. A whole bunch of different things. Anyway, long story short, there's certain things like that that could help society.
Will Zhang: Absolutely. I used to work for the biggest, largest telecom in New York. And [00:18:00] we know we can do exactly that. And I was involved in project those things too. Lemme put it simple. The simple reason is because the traffic light system is owned by government, is not connected to the infrastructure of telecom.
Kevin Weitzel: Air traffic control. I know that has nothing to do with home building, but air traffic control could literally be run by computers.
Will Zhang: Actually, they are. They are. In fact, your internet traffic is actually being run by algorithm. When I was working with telecom, we actually use prediction model to understand where the traffic jam on the fiber optic network is coming, and then we make prediction around that.
Greg Bray: So, Will, let's bring it back to home building cuz Kevin likes to go in other places sometimes.
Kevin Weitzel: What?
Greg Bray: So, for the people who are trying to understand what does AI mean today for my job? You guys are trying to solve a specific problem, but let's go a little more general, just some of your thoughts on how marketers should be approaching AI. How do they decide if a tool really is AI? Just because it has AI in the name, does that really mean that it is? What do you think about some of [00:19:00] those questions?
Will Zhang: Very quickly, probably in the next six to 12 months everything will have some kind of AI and everything will be branded with some kind of AI. From the bigger scheme, I think if I unpack your question to two part is one, how is it relevant to my job? I think AI is just a tool and they're gonna enhance the tool that I already have. So, it's almost like in the past, I need to use a piece of paper to do the calculation and if I need to take a square root of 3 million so it 1, 2, 3, 4 gonna be difficult, but today I can just pull out the calculator to do it. So, I think AI would be the similar concept. It's just actually gonna help you to do your job better.
So, I don't think that will, particularly for a marketing professional if you're really good at creativity, your ideas and your intention of understanding who is your customer, get to know about them, have a more effective market segmentation so that you actually can have your messaging more resonate to the audiences. Ultimately, who have the best empathy to people [00:20:00] is actually human, right? So, I think that actually would not change, but I do see the potential impact to beyond home building and other industry, particularly actually tech industry, right?
That software developer, for example, that fewer people would be so much more effective to do things. If we consider the same kind of job needs to be done, it probably can be done by a lot less people. So, it is a concern about how the job market can be impacted. I do think that's actually a huge topic in the industry. But specific home building, I actually don't think it will probably make that much change because home building is still your construction. The need of construction is actually very physical. It's not that digital yet.
Marketer, the job to understand the consumer, what's your need, and salespeople be a trusted advisor to close the deal. Those are still very much needed. I think that's why we build AI around center. We basically say human-centered AI, so really about [00:21:00] how do we actually amplify the profession to make them more productive by leveraging the tool?
Greg Bray: I love the calculator analogy. I think it really is a good one because I remember back in school there was a time where on the math test, nobody was allowed to use a calculator, right? Because you had to prove that you could write it all out. And now, in school today, my kids have graphing calculators and all kinds of things that it's almost as much power as a laptop on some of these calculators that they take into tests and are allowed to use, right? Because the evolution of where is the value is in understanding when I need to take the square root and why I need to take the square root, not how to actually write out the square root on the piece of paper, right, and solve that piece.
And so I think for the marketers, it's a similar thing, right? It's like, alright, I can use this ChatGPT to help me write something faster or better, but why am I writing about that topic in the first place? Why does it matter? What are the important things that somebody needs to learn about? [00:22:00] That's still is where the value creation is not in just slapping all the words down and putting it all together. At least, I think that's where it's headed. Right. Is that, would you agree?
Will Zhang: A hundred percent agree. Yeah. I take AI in home building industry is more like allowing sales and marketing professionals to really focusing on the customer, the human, and then let AI to take on the mundane, difficult job or repetitive job. And that's actually how I see it.
Greg Bray: Awesome. Well, Will, if people want to learn just more about AI, what are some places that you look for like, what's the new tools and what are some of the different opportunities for them to find some of these tools to use in their job?
Will Zhang: The easiest way to get started is just start using a system. Then you can actually get a sense of how powerful it is. I know a lot of people are already starting using ChatGPT, but there's many other language models as well. Even Google has its own version in Microsoft, which is actually behind ChatGPT again, but just get starting use it.
It's already all [00:23:00] over the news. There's a lot of news about AI. I follow a lot of economic research and all that. I think it's a bit overwhelming where all the information's coming. It's changing so fast. For homebuilding industry, my recommendation is to listen to podcasts like you guys. That's my simple recommendation.
Greg Bray: Well, I don't know how much AI stuff they're gonna get from us but yeah. Kevin and I are all natural. No artificial intelligence here, right Kevin? So, that's.
Kevin Weitzel: There's hardly any organic intelligence sometimes. Trust me. I know. I live in my brain.
Greg Bray: Well, Will we appreciate learning from you today. Do you have any uh, last thoughts or words of advice you'd like to leave with our listeners?
Will Zhang: AI can be overwhelming with all this news. But that's actually a simple way to get started. Actually, is first of get yourself a bit understanding and get a feel about how powerful it is. For home, builders is really actually really focus on the business. What are the fundamental question that you would like to have answered and then in the past you might not be able to answer, but now today you can? Like I mentioned, what are the best product for this market? Who are [00:24:00] my buyer? Where do they come from? What are they looking for? What motivated them to move? These are the questions that in the past it's quite difficult to understand quite difficult to get to. And today by leveraging technology, you actually can have way better clarity on that.
Now if one of these answer can actually tie to business objective, that is actually a great starting point. Then just start treating your data and data asset that have been collected by the great marketing folks and salespeople to actually use them and then answer those question, leveraging those existing data. If you have a great business objective, you have a strategic alignment of your technology to the business, and you are able to control and own your data, and then deployment of AI becomes very simple.
Greg Bray: Well Will, thank you so much for being with us today. If somebody wants to reach out and connect and learn more about OpenHouse.ai, what's the best way for them to get in touch with you?
Will Zhang: You can just search our website, openhouse.ai, or yeah, you can reach me via [00:25:00] email@example.com.
Greg Bray: Awesome. Well, thank you everybody for listening today to the Home Builder Digital Marketing Podcast. I'm Greg Bray with Blue Tangerine.
Kevin Weitzel: And I'm Kevin Weitzel with Zonda Livabl. Thank you.