How We Use Pattern Recognition in Real Estate
Listen
Watch
Episode Transcript
So today we wanted to talk a little bit about one of the key aspects of our strategy, which is pattern recognition and identifying patterns, how we identify them and, and what we do with them. We always use pattern recognition in our day-to-day life. You know, this is being a human is having experiences and letting those experiences develop our, our intuition over time.
And so all we do regularly is apply that to the investing process because investing we believe ultimately is a process, not an art. I think that's important. I think that's an important differentiation. Well, what we're doing is sampling. We're getting sample after sample after sample, and we're looking for which ones stand out as things we should invest in.
So what are the criteria that create a, a sample or a process? Well, it's, it's the various inputs about the current state of the asset. It's the inputs on what we think we can accomplish, and it's the timeline over which that happens. And that's gonna lead to two things, an internal rate of return and an equity multiple.
And so we're looking to be in this sweet spot of having actual inputs that we think were, are actually achiev. Not too low. We're not, we're not sandbagging and we're, we're, we're being too aggressive. We're not being too aggressive. And then having results that that are acceptable that mm-hmm.
In our case, high teens or low twenties, internal rates of return and equity multiples of 1.8, 1.92, 2.1. Yeah. Ideally over two. Yep. So the, so the question is how do you find those, those things? And we talked in an earlier podcast about starting with submarkets, where where demand is, is strong, usually vacancy is a few percentage points.
Mm-hmm. Where prices are affordable on Rente, where we think there's still some room to run, and where our, in our investment cost, our price per door, our price per square. Is well below replacement cost 50, 60, 70% of replacement costs, which would suggest that there's still room to run against new supply.
So we, we are looking for patterns where all those things intersect. Mm-hmm. And one key to this is you have to sample often. Yeah. So I think we underwrite generally about three properties a week. Mm-hmm. It takes our analyst you know, a day or two to underwrite a property, and he usually has two or three going at any given time.
Often we'll get an, we'll get a deal presented to us at one price point we'll underwrite. We'll be at a different price point and we'll, but we put out offers. We put out offers because we're sampling too. Cuz each time we send down an offer and we get feedback, we're sampling how motivated that seller is.
We're sampling how that market's changing. And sometimes we get a new piece of surprising intel that we can then factor into our models to have more expert level sampling and pattern recognition on the next few models. Well, and it, it, it makes it much more of a science than an art when we have a regular filter that we're filtering everything through.
And obviously that model changes, tweaks over time. There are certain, you know, variables in that model that are related to that particular market, you know, but it gives us this standardization so that as we're testing opportunities in a given. We can build conviction around what is the right buy and what is not.
And I, I mean, I think of the Salt Lake City one recently where you know, seemed like $150 a square foot made sense just looking at the market. And and it seemed like that's what, and that's what the seller's expectation were. So it seemed like a deal. And then we put that through our filter and it's $130 a square foot.
And, and The, the, it allows us to take the emotion and the art out of it and use that as a filtering. Obviously then as we get new market information or we get additional intel that can change, but then that just goes on the shelf. That's, that's, that's an offer that sits there. And, you know, the broker gets back to us and says, well, you know, it may take a little while for the seller to get there.
And and that just goes on our, you know, figurative Ferris. An interesting part on that Salt Lake deal was it's a new submarket for us. We have one property there and we're working on a second third. But this sh this seemed to fit within the mental model we had of what a property should sell for based on what, what the rents were achieving.
On paper. This looked good. We modeled it and the analyst sent back a an email and said, we're just not there. You know, If you come at this price, we're there. And he, he wrote a detailed reason for why we're getting these returns. And in this case, we just forwarded the email to the broker who brought this off market deal and said, look, we love this deal.
We're just not there. You know, maybe, maybe we're missing something. Let's get out. Let's do a call and let's figure out what we're missing. And after puzzling through with the broker for a little bit, he said, I think, you know, I think it'll get to this. Let's put in the offer and let them go through a few months of their process.
And I think there's a decent chance it comes back to us. So that is what we did. Time. Time will tell. Yeah. But the main point is we, we saw something, we reacted to it. Yeah. And we responded. And you know, may maybe the tennis ball comes back over than that, or maybe it stays over there. Who knows? We did our part.
Yeah. And I, in the markets where we. It's often the, the other market participants o often are operating much more by art than by pattern recognition. Because being in these inefficient markets, it's often families, it's private developers, it's, it's people that have been used to. I drive down the street and I see a property and I'm like, oh, I think I could do this with that.
And well, if we just did a little bit of this or that, then, you know, we could get to where we want to go. And for, for us to come into those markets and operate with a very consistent, predictable process, it, it makes it much easier to, to identify those true gems versus you know, rough ore. So the part of the process to guard closely.
Is the returns. Mm-hmm. And what we're willing to buy that, that part's pretty much unchangeable. Yeah. Yep. Like that exists. The part to guard loosely I is what you're gonna look through the lens at. Yep. So I, in a submarket, we are really interested in if a broker or a seller or somebody says, Hey, you should check over there cuz you know, or, Hey, over there.
When we look at it from the, from the data, we see a thing, but there is a lot of little undercurrents happening there in that market. You, you could see at one level. Mm-hmm. But to get down lower, there's more detail happening. Mm-hmm. So the sampling's important. Cause sometimes you go places that it feels like.
Oh, that market's too high. Overpriced. Or the, sometimes the thing that's not is, and you know, vice versa. So what we like doing is taking our lens and trying out different things and by sampling in places that you wouldn't expect to sample, you learn new data, that that might lead you back to where you started or might lead you to a different spot.
Yeah, Absolut. Let's talk a little bit about pattern recognition on execution. Mm-hmm. Cuz I think that's a key part of of our investing. You know, we're vertically integrated, meaning we have people on the ground at all of our properties. We have the property management, asset management, fund accounting, property accounting, construction management.
And a big part of that is control of the process so that we are able to execute more swift. But it also creates this amazing feedback loop because we have people throughout that process that are our team, understand our business plan, are aligned toward our outcome, and providing that feedback loop constantly to in the market.
Yeah. Good example is you know, when we own one multifamily asset or whatever product type, I'll just use multifamily when we own one in the. And then we're going to buy the second, the third, the fourth. Now all of a sudden we have all of that real time data from our properties to be able to help us, you know, execute our underwriting.
Yeah. Tighten our underwriting and execute more swiftly on the next properties as well. Yeah. Especially it gives us more confidence and conviction around staffing and how we're going to execute the business plan, how we're going to achieve Yeah. Outsized results often better than our competitors, so, yeah.
Almost always. Yeah. So, yeah. And pattern recognition also extends to knowing the right asset to buy. Yeah. Because we, we speak to this a lot, but I think our, our primary demographic of seller is a private family that has owned a property for 10, 20, 30 years. Yeah. That we've just learned over time.
That, that is almost always an opportunity for buying if the price can get to the price that we need. Yeah. So you know, when we hear that a property's been owned for 10, 20, 30 years, we're probably a buyer. We are a buyer at the right price, right? So normally say pattern recognition looks like a lead comes in in some.
We're, and now we're seeing it through the data. We're seeing it through CoStar and Yardi Matrix. We're looking at Google Maps. We're we're reading news articles. We're, we're developing an idea around the property that way. Mm-hmm. After that, you go to the property and when you drive up to the property, All the stuff you couldn't, you, you, you thought you could see from the data you see in real time.
You see the condition of the property, you see the roofs, you see the, the you, you get a vibe for how the property is. Mm-hmm. Is it run down or well executed? Are people happy? Mm-hmm. Are they sad? Is there graffiti? Is there not? Is it well maintained? Is it not? And now after that, we then generally go to modeling it.
Yeah. Because it has to make it through those screens before it gets. The, the modeling screen. And once you've done all of that, now you just need to get to a seller who's willing to sell the price that makes things work. But it's, it's interesting in that it's a it's a long process and you have to, you have to sample frequently.
Mm-hmm. From lots of properties, 10 or 20 or 30 to get down to one acquisition. And it's a beautiful thing. It's a thing we do well. It's a thing we've done for a very long time. Yep.
Summary
Joe and Ryan discuss how they find patterns and trends in markets targeted for acquisition.