- We need a simple magical solution that caters to agents and stays away from big words like machine learning and fancy algorithms.
- No algorithm can predict life cycle events in anyone's life, and it does not matter how much data is crunched on age, demographics, etc.
- If we can somehow let the consumer remember who the agent is all the time and not wander away -- we have achieved a significant milestone.
You can’t open any magazine or an article and not read about predictive something. Some call it predictive behavior, some call it predictive conversions, and there is other jargon floating around, too.
I admit that I’ve mucked around with a few myself. What does it all mean, and how is it relevant to my business as an agent?
Someone, somewhere invents an algorithm — the best thing since sliced bread to predict human action. Some algorithms want to predict which Starbucks will you drink coffee at today based on chats on Twitter.
Some want to predict what online purchases you will make, and some want to solve the problem of who will sell their home based on demographics, age and a zillion other parameters. And the list goes on and on.
If you look underneath it all, it comes to a massive over-usage — and sometimes-delusional conclusions derived from an age-old field of mathematics called statistics.
“There are three kinds of lies: lies, damned lies and statistics.” – Mark Twain
Now if you add to these big words such as “machine learning,” you have a situation on steroids. Statistics is a powerful branch of mathematics used to study weather, stock market and many other diverse phenomena. Used and analyzed sensibly with human reasoning, it can add value or otherwise spell delusion.
Statistics at work
You can drive on Las Vegas Boulevard and look at the sprawling casinos around, or you might drive to any downtown and see the large insurance buildings. Either way, you will see statistics at work.
Insurance companies — incidentally, like the casinos — know how to make statistics work in their favor. Can the agent do the same? You shouldn’t be surprised that some of the statistical techniques used have names, such as Monte Carlo Simulation.
The real estate consideration
Now let’s consider a real estate agent whose goal is to list homes and get seller leads.
But before that, let’s understand why do people sell homes. There are many reasons, but every seasoned agent knows that buying or selling a home is a life-cycle event — not statistical.
People sell homes primarily due to life-cycle changes — choosing early retirement, losing a job, getting a new job, getting married, divorced, having kids or the kid got in to Princeton — and many other things.
No algorithm can predict life-cycle events, and it doesn’t matter how much data is crunched.
Now let’s look at machine learning. Another term from the good old days of artificial intelligence of the 1980s has now sprung new life due to recent development in robotics. It’s meant to learn recursively from past experience. I’m always surprised that we expect computer programs to learn from mistakes but humans seldom do.
Any of these techniques — whether it’s machine learning or otherwise — require a massive amount of good and relevant data. It’s the old adage GIGO, which stands for garbage in, garbage out.
One single entity or brokerage doesn’t have all the data to analyze. It comes to assumptions and models and many other things that are even harder for a mathematician to comprehend, let alone an innocent agent.
The agent’s workload
One factor crucially missing here is that an average agent is heavily overworked, and most are paid entirely on commission. Many, if not all, are juggling several priorities with finances, careers, family and severely constrained with time. They don’t intend to learn one more thing and add it to their plate.
They all know that they need to follow up with clients, and that takes time and energy that they don’t have. If not, the clients can drift away with their own lives and forget who the agent is.
So where do we go?
We need a simple magical solution that caters to agents and stays away from big words like machine learning and fancy algorithms. It needs to offer meaningful automation, produce tangible results, induce no additional work for agents and have a baseline follow-up that doesn’t let the consumer drift away from the agent.
It’s not the number of leads you have but how many you keep in touch with that really matters.
We need a solution for how to keep in touch with the entire lead ecosystem of the agent (referrals, past clients, current prospects, website, portal leads and so on) — without causing additional work for an agent. The lead might mature in six days, six months or six years. Time will pass whatever the case.
If we can somehow help the consumer remember who the agent is all the time and not wander away — we have achieved a significant milestone. If so, when they are ready for their next purchase, they would call the agent. Agents love ringing phones, and theycan do the rest without the heavy dosage of machine learning.