In this paper using both R and Neo4J we map out a strategy for real estate companies to make more money.

 

Unlike most other industries Real Estate companies cannot come out with new products or lower prices to increase sales.  No one can be coerced into selling their home. The number of homes that will sell in a given year is beyond the real estate companies control.  The only thing that real estate companies can do to make more money is to increase their share of the number of transactions that will occur.

 

In the same way that baseball players have Offense and Defense roles,  realtors have Listing and Selling roles. In Real Estate parlance Listing is a homeowner hiring you to sell their home,  Selling a home is finding the Buyer for that home.  Sometimes the agent who gets the Listing also finds the buyer but this is usually not the case and we will ignore such instances.

 

Not  all realtors sell the same number of houses ,  in fact it is very skewed. In the area that we will look at 24 agents out of hundreds listed ⅓ of all of the homes in 5yr period.  Real Estate managers are myopic about the marquee amount of sales these Star Agents make and in an attempt to increase their market share futility try to recruit these Star Agents.  I think there is a better approach.(A word about the data. We are using genuine MLS data covering 5 years worth of transactions . We have replaced all the identifying information with random information.)

Star Agents can rarely be induced to switch firms.  Unlike average agents who work on a 50/50 split with the company whereby they evenly split any commission collected with their company Star Agents are often on a 80/20 split.  That is about as much as a real estate company can offer and still manage to keep the lights on. Competing real estate companies have nothing more to offer to induce a Star Agent to switch to their  firms. But we will show that it turns out that each star agent has a coterie of Selling Agents who consistently sell their listings . These Selling agents individually do not have eye catching sales numbers but if  several of them are brought together , because they work at a lower commission rate, the real estate company can make more money than if it had recruited the Star Agent. The table shows that on $10million of sales the the real estate company of the Star Agent would have earned $40,000 while the real estate company of the Average Agent would have earned $100,000.

 

	Sales Amount	Commission Split	4% commission on Sales Amount	We Assume Half goes to a 2nd agent	Split Btwn Agent-Co
Star Agent	$10,000,000	80/20	$400,000	$200,000	160000 / 40000
Average Agent 1	$3,000,000	50/50	$120,000	$60,000	30000 / 30000
Average Agent 2	$3,000,000	50/50	$120,000	$60,000	30000 / 30000
Average Agent 3	$4,000,000	50/50	$160,000	$80,000	40000 / 40000

In this study we use actual data so all means of identification have been anonymized .  Unfortunately this anonymity costs us some readability. Here is a simple graph using fabricated data to help explicate what our graph illustrates.  In this example the green circles represent the top two listing agents and the red circles represent Selling Agents that sold the most of the Listing Agent’s listings.  The blue circles represent the listing and what it sold for. Real estate companies should focus their recruitment efforts on acquiring the coterie of the Star Agents.

 

Using Neo4J and switching  to actual data here are the top 5 Listing Agents

 

By clicking on the node and selecting the icon at the    6 o’clock position the sold Listings of this Listing agent appear.

 

Now do the same thing but to a blue node..  Notice the “legend” at the top. Green indicates the Listing Agent, Blue the Listing, and Red the Selling Agent.  As can be seen in the table above it is these Red nodes, the coterie of the star Listing Agents Selling Agents that real estate firms would profit from recruiting.

 

Here is how the analyses was done.

Our data looked like this:

ListingID	ListingAgent	ListingBroker	SellingAgent	SellingBroker	SoldPrice	Year	ListingAgentName	SellingAgentName
2728	702	BAC	36	AEI	585000	2006	20	775
1	47	GF	73	II	695000	2006	6	17
1019	259	A0A	115	II	330000	2006	731	979
712	125	A0B	306	II	465000	2006	20	775
965	226	AAF	348	AEB	470000	2006	20	775
1098	263	BAC	579	AGC	785000	2006	702	979
342	75	FH	600	II	1950000	2006	236	11
1837	405	B0F	608	AFC	677000	2006	2	8
6	447	AAE	657	AEA	530150	2006	316	15
2465	619	DG	824	AGC	3000000	2006	458	676

 

And can be found here: https://docs.google.com/spreadsheets/d/e/2PACX-1vQ3F_ZLfGeoQbW4IMXTXIiTmRMoLwYQeWcd3PguQwBGykjZVkJ3zeFRB-TK70sFomddf2e6_bMAYsxz/pub?gid=0&single=true&output=csv

 

The Data Processing and the  Tree Graph were done in R and the at code can be found here: http://rpubs.com/Joe11579/451108