Matchmaking for real estate? How big data gives us the power to forge client connections

August 5, 2015 by Gina Thelemann in Articles 

Earlier this year, Keller Williams founder Gary Keller called predictive analytics one of the top four tech trends of 2014, and now, industry expert Brad Inman has echoed his sentiments. In his opening remarks at Inman Connect in San Francisco yesterday, Inman said that big data and algorithms are one of five major trends he’s keeping a close eye on in the industry.

Just one hour later, three matchmaking experts took the stage to discuss their work, and how connecting people romantically can relate to how agents pair with buyer and seller clients. Dr. Steve Carter, the Vice President of Matching for eHarmony, spoke about his role leading the data scientist and analysts who create and tweak eHarmony’s famous algorithms. His co-panelists, April Beyer and Lani Yadegar, emphasized the importance of chemistry, saying that even the best match on paper can not work if there’s no spark.

The parallels between digital Cupids like Dr. Carter and what we do at SmartZip are undeniable. Here are the three ways our business models overlap, from the data matching to the connection that is required for a personal or professional relationship to flourish.

1. The imperfect, but oh-so-useful data

Dr. Carter mentioned that eHarmony asks 500 questions to its users before it begins matching them with potential new partners. That may seem like a hefty data set, but think realistically about the difference between who people say they are, and who they actually are.

In some cases, it might be a superficial fib. I’ve had multiple friends tell me that their online matches are routinely two to three inches shorter than they claim on their profile once they meet in person. In other cases, the fib may be accidental. Certainly we’ve all met a Stage Five clinger who believes that he/she is perfectly independent but ready to settle down -- tomorrow.

When it comes to the big data of the real estate world, we are often working with digitized records that were transcribed by hand after decades toiling in file folders at county records offices. Sometimes, even that legacy data is unavailable in non-disclosure counties and states. Some states offer relatively accurate taxed-assessed values for homes annually, while others perform those only when the home is sold or the homeowner requests a new evaluation after an extensive remodel.

In short, eHarmony’s imperfect data may come from their direct users, whereas the real estate industry’s imperfect data comes from countless sources in various formats. In both cases, the data imperfections can be smoothed over once predictive models start churning out results that can be tested.

2. Leveraging (and weighing) that imperfect data

When eHarmony offers matches to the Stage 5 Clinger from above, they are of course not starting from scratch with his 500 questions. They are simultaneously using the data and results of millions of past eHarmony users to influence their current predictive models. As the data grows, so does a data company’s ability to put different weight on certain variables in certain settings.

Dr. Carter said that they are often only trying to be a little better because when you are working with such a large scale, even a 5% lift is going to make a massive difference in terms of success rates.

The same is true in real estate, and at SmartZip. The exclusive territories we sell are optimized to help agents identify the homes most likely to sell, and also to market to those top prospect homeowners. When we work with agents to find the best market area for them, we often try two to three different territories in the area. In this mapping, we look not only at the territory's turnover rate, but also how well we would have predicted specific listings if they’d purchased the territory a year earlier. This is a process known as back-testing the data, and we tweak our models until the past predictions are as close to perfect as we can get them. In many cases, it's also the difference of just a few percentage points.

In easier terms, consider this: After seeing four homes in one subdivision sell last year, you may think it's the highest turnover area near you. Our predictive model goes past your anecdotal evidence to say that you are better off going with the adjacent neighborhood that had nine listings last year. Again, by moving the dial just slightly, it can make a big difference when you’re trying to land elusive seller clients.

3. Sealing the deal with chemistry

Humans are thankfully not machines. Just as dating requires chemistry, so too does landing a real estate client. Buyers and sellers are savvy, and they want to work with someone they trust and genuinely like. eHarmony can set up a fantastic date with two like-minded people who are perfect on paper, and it could fizzle if one of them reaches for their phone mid-conversation or is 20 minutes late.

The same goes for real estate agents who are trying to land clients that have been identified through big data predictions. The predictions and marketing can bring you to their doorstep -- but only a genuine attempt to connect and provide a value will get you in the front door to sign the contract.

Want to check out the turnover and prediction strength in your area?

We love showing off our data and analytics. Reach out today for a no-pressure (and no obligation) territory check.

comments powered by Disqus

Back to all posts

Recent posts

SmartZip Talks Data & Property Management Marketing with FourandHalf

August 1, 2019 by Judy McHargue

Last month we had the privilege of talking with the great folks at...

[Infographic] Best Seasons to Sell A Home - 2019

June 12, 2019 by Judy McHargue

Ask most anyone which is the best season to sell a home, and likely...

8 Reasons SmartZip is an Agent's Coolest Marketing Solution for Summer Listings

June 1, 2019 by Judy McHargue

Summer selling season is officially upon us. There's still time to...

Check listing predictions in your territory

Get started