A Modest Attempt Of Predicting TheFuture
Do you want to know how to make more money with reports that help you make decisions for the future?
AirBnb has over 5 million hosts and has reported to have a total of 1.5 billion successful stays since founding in 2007.
Can you imagine 1.5 billion people standing somewhere?
Looking for a place to stay…
How would you decide what host to pair up with each guest?
Let’s just imagine that for a second.
In 2020, alone there were 77,414,123 successful stays in North America.
And there were only 2,655,700 listings.
Let’s put it into numbers we could figure for a second. If there were like 100,000 listings in Chicago and potentially 1,000,000 guests out there, with different dates and time frames, there would easily be 100,000,000,000 ways to make that happens.
If you don’t like numbers, that is 100 Billion (with a B) possible ways to just schedule short term rentals for a small area.
What Is It With All ThisData?
With rental listings, the amount of data you can collect is endless too…
And the list goes on and on.
With the ultimate goal being:
Pair the guest to the host with the highest probability of successfully booking a stay
The old way of sorting this data yourself to find these numbers used to work.
And in today’s new world of computers learning to learn, things have changed a bit…
You can now use an endless number of free machine learning models to simulate scenarios and help you get one step closer to predicting the future.
That’s what AirBnb does.
How To Win With Imperfect Models
With the help of Machine Learning models, AirBnb is able to use information on the guest, their stay, and other things along with the owners listing to connect guests to the best hosts and connect hosts with the best guests.
A win win!
Even simple ML models, can take in thousands of records from a spreadsheet, and draw relationships in the data that would take you hours in just the click of a button.
And the best part is? It doesn’t have to be perfect.
It just has to be right more often than you and a bit faster if we can (please, and thank you).
You can use these simple models in your business too. I’ve done it for property managers, brokers, and real estate investors too.
Recommended by LinkedIn
For example, let’s say you had a list of leads that all had separate times they wanted to move in with different kinds of information, how would you find them the right place?
An ML model I would love to work on is a scheduling system that will pair the lead with a property that they have the highest probability of staying in.
(While also respecting fair housing rules, you wouldn’t want to waste someone’s time if you have a place in Chicago and they’re looking for a place in a suburb)
“This Wouldn’t Work ForMe”
Okay you might feel a bit overwhelmed…
“Machine Learning”
“Predictive Analytics”
“Instantly Identify Relationships In Data”
“Use It In Your Business”
That’s a lot! All you need to get started is data.
If you’re a real estate investor, property manager, or even a broker, there’s hope.
If you have a spreadsheet, or even better, a ticketing system, you can easily export your leasing data onto nice spreadsheet (or csv technically) that you can use to train your machine.
Now for training the data, and we’ll just cover this part. If you’re technical, you can probably find something to do this, and if you’re not, then find a contractor who can do this for you.
Find columns you have that can be converted from categories or statuses to numbers. For example is you have a Class A property and it can go to Class D, then you would change each property class to numbers from 0 to 4.
And you’d do the same anywhere you can besides addresses.
While there are more complex ways to do data analysis with ML, here’s one you can try today.
With your newly formatted spreadsheet, get ChatGPT, and tell your model to draw relationships based on the data you give it.
And Voila!
It’s alive! Try asking it a question based on the data you have, maybe:
“What are the best apartments I can offer for leads looking to move in June 2024?”
And see what it says…
And remember, the machine can only do what you teach it to do!
References
