No two real estate projects are the same. A challenge for software, an opportunity for AI.

An interesting fact about the real estate market is that no two projects are the same, and there are a near infinite number of ways they can be different. This fact, and its second order effects, is crucial in understanding real estate markets.

‍

In this essay we’ll cover the following topics:

• The factors that make each real estate project unique
• The challenge traditional computer models have with modeling this nuance
• How AI unlocks new use cases for better Real Estate tools

‍

Every single real estate asset is a One of One special

Each real estate asset is unique - this is obvious to anyone working in real estate, and intuitively most people understand this from experience living or visiting different homes. Your first apartment is different from your second apartment which is different from your friend’s across the street.

‍

There are many factors that contribute to the uniqueness, which range in scale from large design choices to tiny details and from physical reality to legal aether. Large, physical differences are easy to spot - things like the colors of the walls, the placement of the windows or the architectural style. A closer look reveals differences in the details - the quality of the windows, the design of the light fixtures, or the source of the marble in the countertops. 

‍

Separate from the tangible, physical differences, there are numerous other variables that make real estate assets unique. At a property level, things like zoning restrictions and planning commissions are important to consider. Many properties will have easements for utilities or special access that need to be considered when understanding the value of a property. 

‍

For real estate developers and investors, the legal structure and financial waterfall is also critical to understand the value of a property. Projects are typically funded with various investment types that could include equity, mezzanine debt, or bank loans. The terms, fees and promote structures in these investment documents combine in unique ways that can lead to dramatic differences in profits. 

‍

The uniqueness of each asset leads to second order effects that uniquely characterize real estate markets. 

The details matter in real estate because buyers care

Real estate assets are capital intensive. For individuals, a home purchase is often their single largest investment and renters spend around 30% of their income on rent. Because of the high cost, the differences in even small details of a project are significant. People aren’t willing to spend such a significant portion of their wealth or income on a property if it isn’t exactly right.

‍

This is intuitively understandable for residential properties, but the same is true of commercial and industrial properties. Buyer’s aren’t interested in a property that is “close enough,” they need properties that meet their specific requirements.

‍

For example, a restaurant owner needs a kitchen space that works for their exact cuisine type. The owner of a pickleball business needs to find a space with roofs high enough to accommodate indoor play. Office tenants need space that fits their company - the right mix of open space and conference rooms, the right shared amenities, and an accessible commute for their employees. 

‍

All of these details are individually important, and the combination and variety of these details create an exponential number of possibilities. 

‍

Historically, this has presented a huge challenge for real estate software. For starters, it is difficult to actually digitize real world assets so that computers can work with them. This means that actually capturing all of this detail typically requires human input for each variable. This makes software adoption for real estate challenging. But the bigger issue is that the details truly matter for real estate, which is different from many other “real world” industries that are being digitized. 

‍

This has to do with the permanence of real estate. Real estate is bought, sold and leased over long time horizons - in practice this means that the small details have more time to be annoying or to be great. If you call an Uber, it doesn’t matter too much if you don’t like the color of the car, because the ride is usually over in 15 minutes anyway. If you stay at an AirBnB, it doesn’t matter if you don’t love the placement of the washer and dryer because you’ll usually be back in your own house a few days later. Those details do matter to real estate developers, because buyers care and buyers ultimately set the price. 

Real estate projects each have their own market

Perhaps the most interesting thing about understanding real estate markets is understanding that they can’t be generalized. The market price for each specific property is set by the specific buyer that is willing to pay the most for the specific combination of details. This means that in creating a new project, developers can’t just pull the nearest comps and call it a day. They need to do the work to understand the details and specifics that local buyers are interested in, and then deliver the project with attention to those details. 

‍

Right now, this is a skill that only humans can do, and this is why real estate developers become more successful with more experience. They train themselves to understand those details and deliver the right product. Tools like Excel or Autocad are utilized, but oftentimes the great deals are made by real estate developers trusting their experience and executing their vision. 

‍

This is another reason that software has struggled to gain adoption in the real estate industry. Traditional software is not capable of gaining experience. Developers can use software to crunch numbers and run analysis, but at the end of the day most software products are simply taking inputs and transforming them into outputs. Traditional software can’t understand a real estate development project in the same way that a human can. 

Opportunity for AI based tools

These phenomena present a challenge to traditional real estate software, but an opportunity for AI based real estate tools. 

‍

The basic promise of Artificial Intelligence is computers that are smart. As humans, we know that computers are good at certain things like doing math and storing data. We also have all had experiences where computers have been frustratingly stupid - excel formulas thinking numbers are dates or having forms reset because an address was formatted incorrectly. Computers are good at computational tasks, but the inputs have to be super exact. The exactness limits how effective they can be, when the detail matters as is the case for real estate. 

‍

What’s changing with AI based tools is that computers are now able to ingest “fuzzy” or unstructured data. It can parse through complicated documents and understand the meaning in a way similar to how humans understand it. Computer vision models can even take pictures as inputs and accurately determine what information is being depicted and what is relevant. 

‍

AI models allow Real Estate software tools to be armed with the same context that humans can access. Computers can now understand and work with the exact details that are present in a project - the details don’t have to be formatted into neat and tidy excel tables. Software tools can effectively interact with the messiness of real life real estate projects, and understand each project’s individuality. 

This is going to create exciting opportunities for real estate developers to elevate their ways of working. We’re building the tools for this at Deco Base. Please reach out if you are interested in connecting.