- Machine learning algorithms will shape our decisions and behaviors in important ways, from working leads, to making offers, to writing contracts, to deciding when and how you work as real estate agents.
- Uber's algorithms behind the supply-side management actively work to spur more drivers to the road when demand is high, and there are no people involved in the process.
- Algorithms in CRMs and lead routing software can determine which agents respond fastest and route more and higher quality leads to those agents.
- Algorithms can show agents which houses a particular buyer is most likely to buy, set buyer/agent communication intervals up to the most optimized cadence, and decide which agents should represent which buyers and sellers for the highest likelihood of profitability.
This year is already the year of the machine. With tech titans promising $1 billion for the Open AI project, artificial intelligence powering Google search results, and AI labs popping up at Facebook, Microsoft and more, there is no question that now more than ever, smart machine learning algorithms will play a leading role in the future of technology and in the real estate industry.
These machine algorithms won’t just make our lives easier, either. They will shape our decisions and behaviors in important ways, from working leads, to making offers, to writing contracts, to deciding when and how you work as real estate agents. Looking at everything algorithms will soon be responsible for, it’s not too farfetched to say that your next broker may be an algorithm.
Your next broker may be an algorithm.
Not convinced? Let’s look at how algorithms can impact agent behavior. As independent contractors, agents set the terms of how they work, including when and for how long. But how much free will do agents really have in the age of algorithms? These software programs are already eroding the notion of autonomy.
Uber uses machine learning algorithm to manage driver supply
Courtesy of Uber
Consider Uber, this time not for the well-trodden rider convenience reasons, but for the case study in how its algorithms manage the supply of drivers. Drivers are independent contractors (for now), and get to “choose” when and for how long they drive.
But for the company to be successful, Uber has to ensure that there is enough supply (cars and drivers) on the road to meet demand at any given time (to pick up passengers).
The algorithms behind the supply-side management actively work to spur more drivers to the road when demand is high. Through a series of calculated and persuasive techniques, the algorithms slowly apply more pressure to latent drivers to balance against the need. From sending inactive drivers push notifications about the spike in available fares, to the deployment of surge-price bonuses, where drivers earn more per ride, algorithms make the overwhelming case that as a driver, you should be on the road right now.
No people involved
It’s important to note that there are no people involved in this process. No manager looking at the rider wait times and doling out orders through dispatch to get more drivers on the road. It’s done by the algorithm, and that algorithm gets more effective over time. It learns what incentives get the most drivers on the streets, which drivers are most likely to respond to which types of requests, and it tweaks everything constantly to maximize performance — down to the wording of notification that most effectively gets the right response. These optimizations are fed back into the system to improve its performance next time.
Now, let’s take that same concept and look at real estate again. There are plenty of opportunities for similar dynamics.
Real estate application
Lead response is one prime opportunity. We all know that response time is critical to converting leads to clients. We hear it all the time. Now algorithms in CRMs and lead routing software can determine which agents respond fastest and route more and higher quality leads to those agents. Again, the algorithm is trained over time based on the individual behavior — no intervention by a manager necessary.
Respond faster to leads, get more leads. Now how autonomous are you, really? Play by the rules, or your personal production is negatively impacted. Take time off, don’t respond instantly, and you may see your lead quality and quantity diminish, without you ever really knowing and without anyone actively taking that action “against” you.
How about lead quality? Credit may soon be extended based on your Facebook profile. If you can lend money based on a Facebook profile, you can certainly score a prospective homebuyer or seller lead by the same information. Algorithms can route the high quality leads to the “best” agents and sort lower quality leads to newbies and those who don’t meet the software’s threshold to get the best ones.
Who or what defines ‘best’?
How would we define best? By reviews, perhaps? Much like Uber can suspend drivers for low passenger ratings or for not accepting enough fares, there’s no stopping an algorithm from taking similar action in real estate.
And beyond leads, algorithms can show agents which houses a particular buyer is most likely to buy, set buyer/agent communication intervals up to the most optimized cadence, and decide which agents should represent which buyers and sellers for the highest likelihood of profitability. They could tell you how much to write an offer for based on what is known about the seller and the listing agent.
The algorithm becomes the broker
When machine learning algorithms do all this, how independent is everyone? The algorithm becomes the de facto broker, even it it isn’t the one holding the insurance policy and license.
Which leads to this question: will machine learning algorithms need to be licensed? Will an agency oversee the settings of these machines, similar to the Nevada Gaming Commission overseeing slot machine performance? Or will the machines be the ones with the “free will” to do what they’re programmed to do, the best that they can?
It’s not a question of if, but when, for all of these innovations to come to the market. The question will be who defines the rules of the road, and who wins and loses.
Will agents be as autonomous as they are today? Will consumers get a better experience? Will brokers use them to maximize their profitability? Or will it be the software companies building the algorithms that are the real winners, getting smarter with every new piece of data flowing through their networked software?
The first of these innovations are already in the market. People are already trying to figure out the logic that makes them tick. These questions need to be answered sooner rather than later, as they will only become harder to untangle the deeper these algorithms entwine themselves in our everyday work.
