The seemingly endless supply of digital information has given rise to the era of big data. Big data platforms have developed technologies and statistical methodologies that can make sense of huge amounts of information by collating, organizing, and eventually extracting intel from massive swarms of raw, unstructured digital content. These platforms are able to identify correlations from tables so big that even a team of the best mathematicians in the world would struggle to find. The ability to use larger, more complex, and less structured data has helped to create a lot of new insights in almost every industry, and the property industry is no exception. With massive stores of information, automated analytics tools, often powered by AI or machine learning solutions, can help make sense of where opportunities exist in the massive real estate market.
One of the most important aspects of modern real estate involves accurately pricing homes and properties based on a series of factors. The appraisal process has traditionally only used a handful of comparable properties to come to a conclusion about the value of an asset. But big data platforms are able to draw from a significantly larger database of home and property profiles, using algorithms to generate more accurate property appraisal values. The idea isn’t necessarily to replace the well-trained professionals that already do the work, but instead assist them.
Josh Panknin describes it best. He’s the Director of Real Estate Technology Initiatives in the engineering school at Columbia University, and Visiting Assistant Professor at New York University’s Schack Institute of Real Estate. “I think the potential for some of these advanced technologies is immense if the industry can develop better data collection and aggregation methods, but the reality is that we don’t currently have the data needed to run these algorithms at scale,” he said. “What I think is much more viable and much more valuable given the current lack of quality data available in real estate is to use AI, machine learning, etc. in semi-automated functions that assist human experts in making decisions rather than expecting purely automated decision-making. Fully automated and accurate is currently unrealistic for most real estate functions.”
The world is now digitally oriented, and just about everything is done online or through some combination of real and virtual experiences. Marketing professionals must take this into account by including online campaigns, which help to develop more successful strategies.
One of the more promising aspects of digital and data-oriented marketing is the option to target specific demographics. Property companies can better market to potential real estate buyers and sellers using online data and personalized profiles. The promotional content is tailored specifically for each user, drawing them closer to the message. An advert or marketing blurb is so much more potent when it draws from the viewer’s personal life. Even better, data-driven marketing solutions can meet audiences precisely where they are, meaning any platform, any channel, right where they are most engaged.
As Panknin says, “data-driven decision making has completely transformed other industries and real estate has taken notice. Making sense of the vast amounts of information that exists and pulling nuanced insights out of that information is beyond the ability of people to do effectively, so the purpose of using data in the real estate industry is to better understand the drivers, relationships, and patterns and structures that drive outcomes. But other industries accomplish better analysis because they have a lot of very high quality, detailed data. Real estate does not.”
Another aspect of property development, or managing properties, in general, involves maintaining existing structures and residences. Most real estate companies have more than one property they are responsible for, which can make the day-to-day of caring for said properties more challenging.
AI and machine learning tools can make the entire process simpler, by identifying which properties need support, and even by predicting emergencies. For instance, one warehouse in Oakland, California, could have been saved from a fire by merely looking at failed inspectional data.
For the potential solution, information feeds into a central system, all of which can be used to inform and direct an AI controller. The technology then alerts the necessary contacts, detailing what’s going on and where. It’s a more automated approach to property maintenance and development.
Abandoned properties are another concern, not just because they’re taking up valuable land but also because they pose a danger to the surrounding communities. Data-driven algorithms can be used to manage and protect companies from related problems through acquired insights. Abandoned property audits, for example, are becoming more and more common. AI can help companies find suitable meaning for these properties, or at the least, help predict when they are going to be targeted by audits.
Beyond that, the technology can help locate potential opportunities for ideal development zones. Finding new properties or development zones for various projects is no small feat. It’s not just about the location, but also about the surrounding community, as well. When choosing a suitable place for a new location, decision-makers must consider the local populace. What are their buying patterns? What businesses will thrive in that particular area, or will it be difficult to bring in new clientele? What competition exists nearby, and how will that affect the performance of the business?
Algorithms are useless without data. But in commercial real estate, the data is terrible.
Big data technologies, especially with the help of machine learning algorithms, can assist in identifying potential areas for development alongside detailing requirements, challenges, and more. They can be programmed to integrate these factors or elements into the selection process, weeding out potential areas that aren’t a good fit.
The challenge is finding the appropriate data to feed into the system. “Every algorithm needs examples, and we call them data,” explains Timothy H. Savage, Ph.D., Clinical Assistant Professor of Real Estate at NYU’s Schack Institute of Real Estate. “Algorithms are useless without data. But in commercial real estate, the data is terrible,” and to deliver accurate recommendations the information absolutely must be correct and up-to-date. “The most accurate rent series I know is quarterly, and it has lived for about 40 years. Four times 40 is 160 observations on prices. The best data in CRE is granular and proprietary, but there is no central repository of CRE data.”
In other words, data used to calculate property development and gentrification must be combined with real world progress. “For development, we need land that is undeveloped, and in turn is developed. We also need to observe local and macroeconomic conditions,” he said. “This question will be important for evaluating whether Opportunity Zones have any positive impact that would have occurred absent law.” All of that must be integrated into potential algorithms to help inform the results.
“We need to focus on the basic economics of the problem. Given fixed supply, as the wall of capital floods the real estate market, returns will be driven down, and the best in class will focus on beating an index,” Savage added. “Algorithms will then be deployed for capital deployment. But they also need data for precision.”
Data precision can even help generate more relevant recommendations for customers. Some of the biggest websites in the real estate industry make money by assisting buyers and sellers alike, doing many things. One of the more prevalent ways is to make a series of recommendations. They might recommend a relevant realtor or sales company. They might help for-sale-by-owner parties to facilitate a transaction. They might even assist in more direct services like appraisals, property repairs, and much more.
To become more effective, these companies have to find new ways to engage and reach customers. The best way to do this, especially with the help of AI and machine learning is to deliver more targeted property recommendations. By analyzing details from a buyer’s profile, these platforms could help recommend and connect the relevant parties through targeted communications. Imagine connecting a buyer and seller who are nearly identical in terms of interests, living habits and needs? Not only would it help buyers find suitable homes faster, but it would also help the entire industry move along at a more brisk pace.
Finding compatible buyers and sellers isn’t the only concern, however. It’s also necessary to choose a fair and reasonable price that matches the interests of both sides of the equation. It’s no secret that trying to find a lucrative yet reasonable price for a home is tough. You don’t want to price a home too low because then you’re not getting the real value, which doesn’t bode well for future purchases, especially for sellers that want an upgrade. You don’t want to price the home too high either, because then it will take an incredibly long time to sell if it does at all.
Finding a fair price for a home is not the same process as generating an appraisal either. The sale price is hardly ever the same as the chosen appraisal value, and the two are often unrelated in the long term. Buyers rarely, if ever, are willing to pay the true value of a home. Sellers do try to get as close as possible, but it’s not always so.
AI tools can utilize pricing algorithms that fine-tune more alluring property values for the seller based on the market, local area, and actual property values. In other words, the system can spit out a valid and reasonable price that meets the needs of the seller, but also reflects the nearby community and its financial status.
At a glance, it’s quite easy to pick up on the overarching pattern of most data-driven applications in the industry. Big data technologies, including AI and machine learning, can be used to develop and apply analytics algorithms for both predicting and reacting to various situations. Also, the algorithms can help better target and engage audiences to improve the sale of properties, residential and commercial.
In real estate and property industries, technology can serve many purposes. It can help generate more accurate property appraisals, fair and well-reasoned home and property prices for sellers, and enhanced buyer recommendations based on a customer’s personality, interests, and behavior. Other uses include researching real estate and development opportunities, automating building or property maintenance, and creating targeted marketing campaigns to boost successful customer buy-ins.
Understanding a building is an incredibly complicated task. Not only do you need to understand the hard data points (its NOI, rent roll, appraised values, etc.) there is also a lot of less easily quantifiable information that needs to be considered. How do people react to it? What is its most valuable usage for both the ownership/management and the end users themselves?