What linguistics has in common with real estate data

When looking at data and data architecture, especially in real estate, it’s important to not simply come at it with a singular lens. You have to have a full understanding of the data and its meaning – to ask just the right questions of the data, depending on who needs it and what they want to use it for.

CohnReznick Chief Innovation Officer Tama Huang, a linguist by training, shared her perspectives at the intersection of linguistics, data analysis, and commercial real estate for a recent episode of Propmodo’s Metatrends podcast, hosted by Propmodo co-founder Franco Faraudo.

Read a transcript of their conversation below, or listen to the full episode at Propmodo.com.

Franco Faraudo:

My name is Franco, I’m the editor and co-founder of Propmodo, and welcome to an episode of our Metatrends podcast, where we look at the bigger picture when it comes to the things changing the property industry.

For years now, real estate companies have been told to make more data-driven decisions. On the surface, this is good advice. With all the data that is now readily available to property companies, it is silly not to use as much of it as possible to inform what are usually huge financial decisions. Other industries, like securities trading for example, have been transformed thanks to advanced software and analytic techniques. Most stock purchases used to be made by individuals on the trading floor based at least in part on their intuition. But now most investors and asset managers use algorithms to help them find value in the market where others haven’t. So, I wanted to learn more about how real estate organizations can use data to bring them into a new level of sophistication on par with other industries. 

Now, I know what you’re thinking. A podcast about organizing data, how thrilling. But my investigations into this topic quickly turned into a conversation about things that had nothing to do with data, things like philosophy, diversity, and the meaning of understanding. 

The person who helped me understand how these metaphysical ideas relate to real estate data architecture is Tama Huang. She is Chief Innovation Officer at CohnReznick, an international advisory, tax, and assurance firm with a long history in the real estate industry. I met Tama at a real estate conference, and I distinctly remember when I met her for a few reasons. First, she’s an Asian woman, which is rare in the white male-dominated world of real estate. Second, she had her two beautiful daughters with her, and as someone who has also brought his young kids to a few of our conferences, I can appreciate the mixing of business and family life. But when I talked to her, I realized that the most unique thing about her was actually her educational background. See, she didn’t go to school for business or computer science, as you would expect. Instead, she studied linguistics. Well, I’ll just let her tell this.

Tama Huang:

I’m a linguist by training and actually went to graduate school focused in the field of hermeneutics, and hermeneutics is the science of translation. And, you know, in history and in literature and our understanding of history and literature has historically been very much given to us through a male lens, because men were literate, men performed the translations. And it hasn’t been until we’ve had, you know – and I shouldn’t just say men, but, you know, really the Western canon – it hasn’t been since we’ve had the ability to have translations and interpretations of people with a more diverse background that we’ve come to appreciate and understand historical events and historical texts in a more contextualized way and in a more kind of rich and diverse way. And I think that those things are true here, too. You know, the big thing about data architecture, and how we use it and so on, is to make sure that we don’t come at it with a singular lens. And that singular lens should not be an IT lens, that singular lens should not be a finance or accounting lens. You know, really making sure that when we embark on initiatives of this nature that we bring a truly cross-functional skillset to bear, that we bring a diversity of experience to bear, and that we tackle it not forcing people to think or consume information in a certain way, and that certain way is usually in the way that the person who has built it or the team that has built it or the types of people that have built it have come at it.

Franco Faraudo:

See, I bet you didn’t expect to learn a word like “hermeneutics” while talking about data architecture. Well, if you think we’re getting philosophical now, just wait. The reason that Tama’s background in linguistics launched her into a career in commercial real estate technology is the way that it helped her define the nature of meaning itself.

Tama Huang:

There’s a French linguist, philosopher linguist, by the name of Ferdinand de Saussure. And what Ferdinand de Saussure said was, to get to understanding, to get to understanding, we need to take into consideration three things: the sign, the signifier, and the signified. Okay? So, the signified could be a thing. Let’s say it’s a tree. The sign is the fact that we call it a tree, right? You can call it whatever you want in any language you speak. You know, that’s just fine. But those two things alone are not going to give you understanding or meaning, because you need to understand who’s looking at the tree, who’s calling it a tree, and for what purpose, so the signifier becomes important.

Franco Faraudo:

Ferdinand de Saussure’s contribution to understanding is that we have to consider the person being communicated to, not just what is being communicated. What language do they speak? What connotations do they have with that word? Tama thinks that the people who build data architecture need to do something similar – to understand how data is being used in order to find a way to make meaning out of it. Much like language, data is only as good as the information that it delivers. 

Now, again, I know what you’re saying. “I came here to learn about real estate data, and what does a French philosopher have to do with it?” Well, first of all, Saussure was actually Swiss. Second, Tama’s point of contextualizing data around the person using it is at the heart of the biggest advancement in the history of data science. When the internet first became a thing, a number of upstart tech companies tried their best to organize the HTML free-for-all that was growing exponentially every day.

Tama Huang:

We ask questions of our data, and you know, the old way of asking questions was with the Yahoo way, right, we have these categories and you ask the question, and for some of us, we may remember Yahoo not as a news platform but truly as the original search engine, and you go down to search and you’re going, “No, no, no, that’s not what I meant.” And, you know, you retrace your steps, you go up that decision tree and come back down the different tree to get to the answer that you were looking for. So, you know, I liken that to finding a precise answer to an imprecise question. And it’s really hard because, you know, there are a thousand ways to get something wrong, and there’s only one way to get it right. And when we chase that one way to get it right, it’s very difficult. I believe that that is why, you know, kind of the traditional BI [business intelligence] projects and big data warehouse projects of the past have failed, is because we’re chasing that one right out of the million wrongs.

Franco Faraudo:

But all the coders and all the data scientists and all the MBAs that Yahoo brought in to help solve this problem couldn’t figure out the right matching algorithm. The one that would give a precise answer to an imprecise question. Then two grad students at Stanford discovered a way to rank a page’s usefulness, based on how many times it was linked on other sites. This allowed them to surface potential answers to queries of any kind, no matter how imprecise they may be. 

Tama Huang:

Until Google. Google came onto the scene completely inverting that. Google purposefully created Cartesian unions. Cartesian unions are many to many. Complete explosion of answers. So that was an imprecise answer to a precise question.

Franco Faraudo:

Cartesian unions, often referred to as Cartesian squares or Cartesian products, are a way of organizing data into a grid. This allows every combination to be mapped based on coordinates. By doing this you can then examine the relationship between two pieces of data. The most common illustration of this uses a deck of playing cards. One way to bring order to the 52 cards is to put them in order from two to ace in each respective suit. If you were to do this on a table, you’ve just created what is basically a Cartesian square. 

I won’t spend too much time on what is called analytic geometry, because it doesn’t apply too much to where we’re going next, and, frankly, I don’t really understand it well enough to be able to explain it to anyone. One thing I will say is that, interestingly, it was invented by another well-known philosopher: René Descartes. I know, I know, I told you we would talk about how to organize real estate data, and so far, we’re still talking about French philosophers. Well, first of all, even though Descartes was born in France, he spent the last half of his life in the Netherlands, so he’s kind of also Dutch. Second, don’t worry, we’re there now. Remember, Tama advises real estate organizations all day long. So you better believe she has more than just a philosophical level of understanding of real estate data.

Here she explains a bit about why real estate data, in particular, is hard to organize.

Tama Huang:

If you thought of questions and answers as kind of your supply chain, right – so you have supply and demand, the demand is the question, and your supply is the data that surfaces up that answer – we could categorize it into a couple of things, right? So there’s the known and the known. That is to say, I have a known question, and a known answer. We’ve done a really good job in real estate at that. You know, there’s always, you know, tweaking we do in that little area. You know, what’s the commencement date? Well, did you mean the lease commencement date, event commencement date? Did you mean the move-in date? Did you mean – what do you mean by commencement date? I mean, that sounds like that should be a pretty easy answer, but it isn’t, right? What’s the square footage? Is it the, you know, occupied square footage? Is it the square footage I’m using in my CAM [common area maintenance] calculations? Is it the square footage that you’re using to, you know, pay your rent? What’s the square footage. Like, those are, like, you know, questions that we’re asking where you would think the answer is pretty straightforward, but depending on who the person is that’s doing the asking or what the purpose is for which you want that answer, right? Is it to know, you know, how much we need to clean, or how much paint we need to buy to paint, or, you know, things of that nature? Or is it to calculate, you know, how much you need to pay me for the CAMs that I’m dividing up across the tenants, or is it because, you know, that’s how we’re going to calculate how much rent you’re going to pay? You know, there are lots of things like that where there are nuances to this, and it depends on the purpose, it depends on what you want to use it for.

Franco Faraudo:

It isn’t just the data that is diverse, either. Any real estate transaction touches dozens of people in a wide variety of roles. Depending on what they do, they have very different needs for the data. So it is the job of the person organizing the data to match the two: the person, and the correct form of data for them. Here’s Tama again.

Tama Huang:

One of the challenges that we’ve always had is that we, you know, as leasing agents, to enter into an accounting system, and, you know, that’s just a personality mismatch, right, like eHarmony would never make that match for you. Or we’re asking accountants to be like, “It’s good enough, it’s good enough.” An accountant is always going to want to come down to the last penny, and no rounding, you know. So matching the requirements as to what it is that you’re using it for with the type of tool, with the level of granularity with the data, right? I mean, our beta initiatives succeed and fail not because we don’t have the technology. It is because we’re not using the technologies for the right purpose, and we’re not matching the purpose of why people want the data and use the technology with what it is that we provide for them.

Franco Faraudo:

If there exists this mismatch between data and the person using it, it becomes less useful. It can cause hours of duplicate work and double-checking, and makes it almost impossible to automate certain processes. It also reduces your ability to find patterns that we’re not even looking for in the first place.

Tama Huang:

There are things that are, you know, where we have unknown questions with known data, or known questions with unknown data. So when you think about it, it’s, you know, a lot of real estate technologies and a lot of real estate data is really about the rearview mirror. It is, how did we do last month, how much rent did we collect, how much did we spend? Basically, after closing the books, we look back. We don’t drive a car looking back. We look forward. We look at the windshield, not the rearview mirror. And that’s where the data becomes, you know, a little bit more unknown, and that’s also where the questions become a little bit more unknown, especially in times like today. Right, so that is where we’re looking at forecast data, at budget data. You know, where we are wanting to do scenario modeling. And, you know, the technologies and the data requirements that we then have as we look forward, as we look through that windshield, is very different. So we have different, you know, kind of data personas, if you will. The looking back folks, transaction level, they’ve got their sleeves rolled up, they’re just really digging in. Act and transact, and in the data and in the mechanics and in the plumbing, and really going through that. You know, as you come up and you’re starting to get a little bit more to the unknown questions, or the unknown answers, you know, now you’re starting to be an analyst. Right? So we now need to come up, consolidate and aggregate the data, so that we can do something with it.

Franco Faraudo:

In normal times, not seeing unknown unknowns coming can cause real estate companies to miss out on potentially profitable opportunities. But when unknown situations arise, like they have these last few months, as the coronavirus has pretty much put most of the world out of work and frozen the real estate market, it can be the difference between thriving and insolvency. There is no perfect way to organize data. Every organization has its own considerations. But after talking to Tama I realized the most important consideration isn’t the data itself. It’s matching it to the people using it, and having the right signifier at the right time.

This podcast is one of four in our Metatrends series about the technology trends that are shaping the real estate industry. Please check out the rest at Propmodo.com, and don’t forget to sign up for our daily email to get all of our thoughtfully researched and painstakingly written content delivered directly to your inbox every day.