Imagine an office building. What do you see? Does it have a sleek, modern interior with plenty of natural light? Do you see polished concrete floors in an oversized lobby with industrial backlit artwork adorning the walls of the elevator banks? Can you picture the rows of long narrow desks to accommodate open seating? Are the chairs ergonomic? Don’t forget conference rooms, reception areas, restrooms, and kitchens robustly stocked with gourmet espresso. Can you see it? Do you have a clear picture in your mind of this pristine new office building?
Now imagine what’s missing. It’s the most important part, some might refer to it as the building’s heart, its lifeblood. Yes, you guessed it: people. People quite literally bring a building to life. Now, think of all the people who might work in or for this new office building: receptionists, security guards, doormen, maintenance providers, salespeople, managers, executives, administrative assistants, chefs, housekeeping attendants, property managers, accountants, and so on. These people do more than just earn money in this building. They engage and collaborate. They connect and forge relationships. They create, grow, and develop the world around them. And they generate data.
The people living and working within a building’s four walls not only generate data, but they also bring it to life. Without people occupying and using its spaces, a physical building provides significantly less usable information on its own. No people equates to no leasing information, no occupancy data, no utility usage, no foot traffic reads, no sales generated, no amenities needed, no services provided, no maintenance requests, no qualitative feedback given. All of which are important measurable metrics required to make a well-informed decision. Sure, the saying goes, “location, location, location,” but why not factor in all of the “human” information now at our disposal? Probably because the process of turning data into insights has traditionally been difficult and inefficient, but as technologies evolve, that can no longer be an excuse.
“With access to the engagement and interactions of people, buildings are critical to composing the human element within data.” This statement from Julie Miner, the real estate advisory services leader at CohnReznick, rings true for all kinds of commercial real estate assets, not just the aforementioned example of an office building. Consider multifamily units, hospitals, or hotels and the human activities that regularly occur in these places. Childbirth, marriage, death, and the infinite moments in between are all taking place within the four walls of commercial real estate, which equates to multitudes of extremely valuable data.
The insights generated from this data are literally shaping the future of our built world and how we interact with it (and with one another). For instance, having surgery in a smart hospital would be a very different experience than having surgery in an “analog” hospital. Aside from having better technology, capabilities, and most likely doctors (who want access to the best technologies), a smart hospital would be capable of anticipating your needs through technology, like wearable sensors to track your health in real time and automatically deploy medications or alert a nurse. Even within the realm of office space, people are already using data to measure air quality, wellness, and even life expectancy—all of which can be vastly improved using technology. Data indicates a rising demand for green, sustainable buildings. If you live in New York City, the law will soon require full transparency in this matter, and who wants to lease from or work in a low-grade building? All the more reason to embrace technology and innovations that can help get your D building to a B. Industry professionals should be able to understand how data is generated and consequently, how to analyze and use that data. Easier said than done.
The reason understanding data architecture within commercial real estate can be so challenging is because there are so many ways to categorize and catalogue information. No agreed upon lexicon exists. Yes, data must be collected, stored, and defined. And it has been, to an extent, for many years, but not very efficiently. Old analog real estate records, however meticulously kept, are sure to have gotten lost in the translation to digital. And any digital records that were manually input are bound to have flaws. Humans can be messy and organized. Our data follows suit. We are not always as efficient and rule driven as the machines we build. Our records are only as good as the people keeping them.
The technology we have today, particularly AI, works better when it’s supplied with more information to analyze. An asset with ten years of data will provide better insights than an asset with only two years of data, at least when it comes to machine learning. Humans have the advantage of context, but even that can be translated into quantifiable data and fed into an algorithm. How it comes out is dependent upon the platform’s ability to turn all of this information into a digestible format, and not all data management platforms (DMP’s) are created equally. Cherre is a data management platform that has a comprehensive approach by sourcing from public, purchased, and private data. The platform’s ability to compile these separate sources into one place provides a more thorough analysis while making insights more accessible. Finding the right platform for your organization will be a key factor in making data architecture transparent, meaning the people who need the information are able to easily access and understand it. That has not always been the case.
Now, the way data is being used and collected is changing. Traditionally, data has lived in the world of IT and is primarily used by people in technical, or “back office” roles. And the people in non-tech roles haven’t complained much about that arrangement. But it doesn’t do them any good to have only a portion of the available information when making important decisions, like where to target future investments. Nor does it do the IT department any good to sit on loads of raw data that need to be managed and protected. Once raw data is turned into a digestible, organized format, then data architecture can begin transitioning into “front office” positions as well. A successful data strategy integration requires a collaborative effort throughout the entire organization, and in many cases, a separate data science department may be necessary.
Manny Bernabe, founder and CEO of bigplasma.ai, believes that having a separate data science department creates enough autonomy for those individuals to explore what platforms make the most sense for an organization, as one size fits all does not work for data strategies. In some cases, companies can implement purchased software to sufficiently cover their needs, but for others, building a custom platform to fit the unique needs of the business will be the best option. For most mid to large companies, bringing in a data analytics consulting firm will help them to determine the best course of action.
Bernabe discusses the importance of hiring trained data scientists and engineers who can then collaborate with IT departments, analysts, and consultants to optimize strategy. He also believes it’s possible to train current members of an organization in data science, but typically, you would want people with experience leading the initiative. (Meaning, don’t promote an analyst who isn’t fully versed in data science to run the show.) Depending on a company’s size, it may be important to hire or promote an executive level chief data officer (CDO) to lead the strategy. Hiring a data consulting firm can help you figure out if that’s necessary and guide you to the right resources.
Regardless of who is leading the data strategy, they need to keep an open dialogue with other C-suite executives to make sure that any technology being implemented or developed will ultimately add value. Business-minded individuals will be able to guide engineers and analysts to ensure the intended outcome of the strategy is doing at least one of the following: cutting costs through efficiency, enhancing the company’s service or product offerings, or creating new platforms that can be monetized. The strategy leader will also need to manage change by maintaining consistent data architecture for company-wide accessibility.
Humans can be messy and organized. Our data follows suit. We are not always as efficient and rule driven as the machines we build. Our records are only as good as the people keeping them.
Regardless of how a company chooses to implement their data strategy, the entire organization will need to adopt the necessary changes, particularly in terms of secure information sharing. Certain data is subject to privacy mandates. Companies need to be cautious in the kinds of data they collect as well as their methodologies for sharing and retaining it, meaning cybersecurity cannot be an afterthought. If anything, cybersecurity needs to play a larger role company-wide. IT departments will need to ensure everyone is up to par on privacy standards and best practices for information sharing protocols. Nearly all data strategies will call for a more holistic approach to the collection and sharing of information, once again emphasizing the need for collaboration. There is no “I” in data.
As an overarching concept, data architecture contains a few key points to denote. First, there are different kinds of data, and I don’t mean just quantitative versus qualitative. Although both of those do exist within the context of commercial real estate, they are not as black and white as you’d think. For example, how many vacancies an asset currently has is clearly quantitative, whereas the accumulative narrative history of a property might be qualitative, but could also include quantitative data within that history. A property’s accumulative narrative history would certainly include a purchase timeline with dates and costs being quantitative, but it could also include a property remodel or expansion. Perhaps the property was once a factory that has been repurposed into loft units for multi-family living. Both quantitative and qualitative data in commercial real estate are equally valuable, but that value fluctuates depending on the data’s intended purpose and the individual using it.
To further categorize data, you can assess it through the three kinds of lenses, as coined by CohnReznick. The first is data taxonomy. By definition, taxonomy is “the branch of science concerned with classification; systematics.” Data taxonomies help people understand where to find specific information. Vincent Dermody, managing director of CohnReznick Australia, breaks down data taxonomies into three subcategories for assets: descriptive data, product data, and insider control data. Descriptive data refers to the physical description of an asset, including square footage, the physical space (i.e. how many floors, how many units, physical amenities, etc.), building capacity, finishes and materials, and technical hardware. Product data refers to the monetization of the asset including lease terms, net operating income, net operating costs, valuations, forecasts, and any other information that is vital for determining an asset’s financial performance. Lastly, insider control refers to information that is not as easy to access. It includes data like foot traffic in retail spaces, the speed of how quickly a unit will be leased, and other pieces of information that comes from “reading between the lines” in respect to an asset’s operational performance.
Think of a taxonomy as a kind of filtering system. It would allow me to search all of the assets within a portfolio for a specific piece of descriptive data (for example, filtering by an asset’s number of units). Then, I could further narrow my search by layering product data filters, and so on, until I found all the assets meeting my desired conditions. I could then compare their performances and make judgements based on distinguishing factors or anomalies. When the right systems (or DMP’s) are in place, data taxonomies make life easier.
The second lens with which data can be assessed is an ontology, or “a set of concepts and categories in a subject area or domain that shows their properties and the relations between them.” Tama Huang, the chief innovation officer at CohnReznick, explained the concept using an example of the yellow pages versus the white pages. Remember (if you’re old enough to recall this ancient construct) the distinctive thud these giant paper data sets used to make as you set them on the counter. Google does not yet exist, and you need to find a telephone number to order pizza. Where do you start? Without knowing the name of the restaurant, the white pages are of little use. So you look under “R” in the yellow pages for “restaurants,” but it’s hard to determine which restaurants offer pizza. You turn to “P” for “pizza,” and there it is: “Joe’s Pizzeria!” In order to find information, we have to know how that information is categorized and catalogued in relation to its subject area. For data ontologies to work, the industry must agree on classification terms and how they’re catalogued in relation to one another.
The last lens through which data architecture can be understood is epistemology. Epistemology is the theory of knowledge, especially with regard to its methods, validity, and scope. It requires investigation into what distinguishes justified belief from opinion. True or false? Fact or fiction? Theory or hypothesis? In real estate, epistemology can be defined as the distinctive methods that real estate practitioners have agreed to use in the acquisition and sharing of property data. But we know from experience that humans are not always the best record keepers. One misplaced decimal can create substantial chaos in a commercial real estate database. That’s why verification and automation (when applicable) are so important. Overtime, accurate records create an accumulative narrative history for different kinds of assets and for the markets in which they reside.
Huang provided an example of a client’s request to assign each asset its own unique identifier. If a portfolio is global, tax identification numbers are an insufficient way to categorize them. To be able to both quickly identify and understand an asset, a timeline was created, including multiple data points like geographic location, original purchase date, original purchase price, if and when the asset was refinanced, and so on. Each point of data correlated to a hash or timestamp. These hashes were then compiled together to act as a data-rich historical record for the asset while also uniquely identifying it.
These property identifiers were used to build a catalog of the company’s entire portfolio of assets. This process highlighted the relationships between the assets’ data points. By comparing the properties’ attributes, they found important correlations and useful insights. Huang believes that in order for data utilization to continue progressing, business consultants and technologists (or front and back of office, as mentioned earlier) must be willing to collaborate and share information. A binary, technological approach can no longer be used when it comes to data architecture. She believes “an ontological construct is the way of the future,” which makes sense, considering it mimics how humans think about the world—in a series of contextual relationships.
If technology and commercial real estate were two different colored circles in a venn diagram, data architecture would be the shaded area in the middle where the two converge—their relationship creates vast pools of meaning. In order to decipher the meaning, both “front and back office” positions must be prepared to wade through darker hues where the colors are more richly saturated as they coalesce. Here, the path is not always clear.
People design and construct buildings. Buildings are bought and sold by people. People live and work in buildings. At real estate’s core, people and buildings are inextricably tied to one another. Just as we couldn’t imagine a new office building without considering the people inside of it, we must also make sure to include humans in any conversations around data architecture. Both buildings and information systems are built for a function, to help humans achieve more than they could without them. Whether an office building is able to do that depends on the way that people are able to work inside of it. Likewise, in order for a company’s data to be well architected, it must be built on a schema that is understandable and efficient for the people using it.