During the CMLS conference a few weeks ago, the now perennial topics of raging regionals, syndication, parcel based MLS systems, and consumer facing MLS sites were still hot. I’ve covered these topics at length here on the FBS Blog for the last year or so but just started thinking about another, seemingly mundane issue that runs through all these issues, namely the best way to organize (and, therefore, search) MLS data geographically.
This issue impacts all the hot topics being faced by MLSs today: (1) regionalization and cross-MLS data sharing directly raise the questions of boundaries and geographies; (2) syndication of listings should be done in a consistent manner and the geographic data points are critical; (3) parcel maps are key to accurate geographic positioning; and (4) consumers want the easiest way to find properties in the areas in which they’re interested, and they want to see statistics and other data organized around those same areas, which requires solid and shared definitions.
Let me provide some examples to highlight the challenges. First, many MLSs have created their own “ares” or “regions” to make organizing and searching data easier. Here’s an example MLS area map from our customer in Santa Barbara.
MLS Defined Areas
The advantage of MLS defined areas is that they are easily learned by the agents and often represent the market fairly well. One of the disadvantages of MLS defined areas is that markets change, and maintaining the areas and the consistency of the data has historically been difficult. Also, as shown in the image below, the MLS defined areas raise the continual problem of the outer boundaries of the MLS. Lastly, there is some concern that MLSs defining boundaries or areas is a potential fair housing issue.
Outer Boundaries of MLS
Notice how when the map is zoomed out, the boundaries of the MLS become apparent and the issues of cross-MLS data sharing via the MLS defined areas raises all sorts of challenges.
To address some of these issues, many consumer facing web sites focus on city or zip code as the key geographic criteria, primarily because those are most familiar to consumers. The challenge with city and zip code is that they are just too broad and often bear no relationship to the actual real estate markets within the city or zip code as shown above in the first image where the zip code actually is split across two different MLS defined areas.
Zip Codes Are Too Big and Not Market Defined
To combat the problem of zip codes and cities being too big and the boundary limits and constantly changing nature of MLS defined areas, some MLSs have gone to a “grid” format that simply carves up the geography into squares or rectangles.
Map Grids Provide Consistency and Scalability But Are Not Descriptive of the Actual Market
These grids are very useful because they can easily be extended to new areas with consistency, but the problem is they do not describe the actual market, which has twists and turns every which direction and those twists and turns often mean tens or hundreds of thousands of dollars in median home prices. Accordingly, market analysis is not useful based on a grid system and the grid system isn’t easily used by consumers.
New Mapping Technology Allows Custom Areas
To address some of the limitations above, some MLS systems and consumer web sites allow users to draw directly on the map to define exactly the area in which they are interested. There are several limitations to this approach in most systems: (1) many agents and most consumers don’t want to or don’t know the detailed boundaries of the true market areas; (2) the areas that are drawn are not shared with others who may want to use them as well so users can learn from each other; and (3) they aren’t consistent enough to enable gathering of statistics and other data for agents and consumers.
Yet another approach has been to gather neighborhood boundaries. Two companies working on collecting this type of data are Maponics and Urban Mapping. In addition, Zillow has put forward some neighborhood files as open source files for contributions from many.
What I’ve been thinking about recently is how we might be able to create a massive win-win for MLSs, agents and consumers by enabling the real estate professionals to contribute neighborhood information directly into efforts like those linked above. The result could be a nationwide set of neighborhood boundaries that accurately define the market areas and allow for easier organization and searching of MLS data and presentation of market statistics.
What do you think is the best approach to organizing listing data geographically?
The ground between portals and IDX is being defined in large part by Alex Chang from Roost. As a result, I was psyched Alex agreed to be a part of our panel on MLS portals. I had several take aways from Alex’s talk:
After the panel, I had many participants say how much they enjoyed Alex’s perspective, which was just enough outside the MLS perspective to be fresh and invigorating.
In my last post on the public MLS site panel discussion, the panelists seemed to agree that IDX programs don’t conflict with a public MLS site. The more I think about this issue, though, the more I think they do conflict or involve a “fork in the road” that needs to be illuminated before choosing one or other or to build a new road that combines the two. Put simply:
As Brian Larson says during the video, there is a fork in the road here. Perhaps new roads will be built that create the best of both worlds, but, today, the difference between IDX and public facing sites remains distinct and choosing both creates the potential for conflicting policies. I think the difference between these policies is highlighted well by Duane Sauke’s refusal to weep for buy-side brokers who get to use his listings on IDX sites.
Some time ago, I wrote this in a post called Listings, Leads and Losers, Oh My!:
IDX was a great idea. Larry Cragun and Kris Berg have it right, the agents should be delivering the listing information to the consumer. Importantly, this argument should not be confused with those who think agents make a living simply by hording the listing information. I seriously doubt that ever worked as a long term business model, but, if it did, there is no question that such a model won’t survive today. No, what I’m suggesting (and what I think Larry and Kris are suggesting) is that the agent is on the front lines and there should be as little friction as possible between them and their potential clients. Isn’t that what the web revolution is all about, efficiency, knowledge, and power to individuals? That was the promise of IDX and I think it can still be brought to fruition.
Here’s how we can revive IDX. Brokers and agents should be allowed to put all the listing information on their web sites. The IDX data set should be complete (which should get the DOJ off the NAR’s back) and brokers should be assured that someone else isn’t going to sell leads from their data (this is called compromise). Once data standards are established, promising a strong and comprehensive data product from brokers and agents to consumers, NAR should create and heavily promote a branding and logo program for legitimate IDX sites so consumers can have confidence they are seeing everything. (No, I’m not crazy, this is a really good idea, if you pause to think about it. See, the NYSE and NASDAQ for examples.) Consumers should be able to shop till they drop, without being bothered, until they’re ready to engage. But, when they are ready to engage, they should be one click away from the agent who is going to serve them. Getting funneled through lead site after lead site before getting to the agent is not efficient and is not consumer-friendly.
This post was the genesis for my question of whether a public facing MLS site conflicts with IDX programs. Put another way, is branding IDX a better investment for members in an MLS than branding a public facing MLS site? Is there enough budget to do both well? What is the market impact of one versus the other? Which would members prefer? I didn’t put all of these questions to the panel, but I think the responses are valuable on the main point.
The first question I posed during our panel discussion on public facing MLS sites is a variant on the classic “level the playing field” issue and whether public facing MLS sites benefit big brokers, small brokers, both or neither. Victor Lund starts off the discussion suggesting that public facing MLS sites would benefit big brokers more, because, having more of the listings, they would get more of the leads or traffic. Duane Sauke, the broker on the panel, said that, if it is true that big brokers would benefit more, then they’ll still not want an MLS web site because they’ll want to own the space entirely.
Brian Larson then follows that up with his mantra throughout the panel discussion, which can be paraphrased as “show me the data.” Lastly, Brian suggests that MLS web sites have the possibility of dampening competition by “treating everyone equal” and puts forward Zip Realty as a new web competitor that has succeeded but may not have if it was competing against a public facing MLS site. Combining Brian’s observations, I wonder what the data shows about Zip Realty in Houston?
What do you think? Do public facing MLS sites benefit big or small brokers more? Or neither? Or both proportionally? Following Alex Chang’s suggestion that the answer depends on how the site is constructed, how does IDX fall into the mix?
P.S. Isn’t the ablity to deep link into a Viddler video awesome?
A week or so ago, we held our annual FBS Summit and hosted a panel of speakers on the pros and cons of MLSs creating and promoting consumer facing listing sites. We recorded the sessions and now have the videos uploaded for public viewing. We had to break the nearly two hours of video into many small parts and the easiest way to show them all is with the links below. Over the next week or two, I’ll be writing some posts with additional commentary on some of the videos and embedding them at that time but if you want the full experience just click through the following links.
Question 1 – Do MLS sites favor big brokers, small brokers, both or neither?
Question 2 – Is an MLS site in conflict with IDX?
Question 3 – If an MLS has a public facing site, what are the best ways to drive traffic to brokers? (Pass through versus destination.)
Question 4 – Are leads from a listing search good leads given dual agency? How do consumers choose an agent?
Question 5 – Is promoting an MLS brand in conflict with the brokers’ efforts to promote their own brands? Is advertising by the MLS in conflict with the advertising of the brokers?
Question 6 – Should an MLS spend member dollars on advertising a portal?
Question 7 – How would the branding efforts of a site like Roost compare to what an MLS could or should do?
Question 8 – Can an MLS promote the REALTOR brand and the MLS web site at the same time?
Question 9 How can MLSs help train agents to respond better to consumers? Is that the role of the MLS?
Whoa, check it out, our little niche hit the New York Times today. They call it M.L.S., which I find funny, but most likely is correct. Anyway, on to the article. I wonder if Bob Hale knew they would start the article like this:
The triple threat of a weak market, legal pressure and increasing competition has compelled real estate professionals to offer their information more freely online, putting cracks in a walled garden of data that stood strong while the industry enjoyed its breakaway growth. It also presages an end to the days when sellers must list their homes with a broker so buyers can see them.
I also find this quote interesting:
Tom Hurdlebrink, chief executive of Northwest M.L.S., said his service’s shift [in allowing Redfin and others to display FSBO's and non-MLS foreclosures intermingled with MLS listings] was meant to “create a balance of giving consumers what they want while promoting the best interest of our broker members.â€
Bob Hale concludes:
But Mr. Hale, of Houston’s M.L.S., suspects that resistance will wane. “Their attitude has been, ‘Just because the consumer wants it doesn’t mean we have to give it to them,’ †he said. “It’s the sure way to your demise.â€
Hmmm, at some levels, this seems like a “dead if you do, dead if you don’t” conundrum, but, at the least, it poses a very good prelude to the discussion we’ll be having at our FBS Summit June 12 regarding public-facing MLS sites. We’re assembling a panel with Brian Larson from Larson/Sobotka and Marilyn Wilson from WAV, who each have written papers recently on the issue of public-facing MLS sites and have somewhat different views. We’re also in the process of getting at least one MLS executive and one broker with differnt views. We’ll start with some presentations and a panel moderated by me and we’re going to then follow with a speed Q&A session that I think will be really interesting.
For example, Trulia and Realtor.com are mentioned in the article as well. R.com is against FSBO’s for data quality reasons and Trulia is against them because they offend brokers. Who has the better model here? Does it all come back to Google in the end? Are we in a battle for links and link love? Because here’s the deal: Links require public-facing sites. Does the MLS have a role to play in that battle? Or not?
The Gateway/TREC/TBD discussions are heating up again, sort of. Re-visiting the topic seems timely with the latest iteration of the NAR PAG report recently “released” and the trimester RETS meetings in Philadelphia next week.
My take is this: The NAR may well have the cart before the horse. So far, the NAR’s PAG seems to have focused on defining a “system” for aggregation but the TREC documents avoid the fundamental question of the terms of use on which brokers and MLSs will agree to share data. Here’s a quote from the document:
Then, further down, the document says:
These two statements seem inconsistent, because an MLS or broker submitting data to TREC will be required to license that data to all other TREC subscribers and such a license is a transfer of intellectual property rights. This apparent conflict is just one of many key questions, however, regarding the terms of use for the data submitted to TREC.
The document also stresses that the data will not be publicly accessible. Yet, at the same time, the “Statement of Inevitability” at the beginning of the document lists the proliferation of “[c]onsumer-focused real estate websites” as one of the reasons TREC is necessary — to keep Realtors “at the center of the real estate transaction.” At this level, not only does the concept feel like the cart before the horse but also like trying to put the horse back in the barn, just as they’ve started to run free.
This brings us back to the core issue: What are the terms of use on which brokers and MLSs will allow others to use their data? TREC tries answering this question only in the most narrow terms (broker/agent to broker/agent, in private, behind a walled garden) but the web insists on asking a much bigger question involving not just brokers and agents but consumers and third-party aggregators and so many others. As the links above show, brokers are already answering these questions for themselves by sending listing data here and there but not yet everywhere, which creates the very real likelihood that TREC will be too late to the party.
Of course, NAR is hamstrung in fostering the necessary dialog about usage of listing data on the web by the DOJ litigation. Also, in spite of the insistence that it isn’t any such thing, TREC very well may be the first gingerly step in the direction of addressing the question of how listing data should be used on the web. Rather than being the cart before the horse, an actual implementation of a national aggregation may be exactly what is needed — and all NAR can do for now — to incubate the discussion. For this, I applaud the work of the PAG.
Like the NAR PAG, I believe this is a defining moment for real estate data on the web and I encourage the NAR to consider how it can re-visit the big issue of sharing listing data on the web. That’s what the consumers want, not having to go through an agent to access the data. For example, one of the best uses of TREC data would be for an agent or broker to create a publicly accessible web site to engage their customers. Yet, that seems off the table. I’m pretty certain the NAR PAG would have loved to create a more sweeping plan involving both public and private access, but felt it wasn’t possible. The result is the proposal focuses on implementing a closed system instead of creating a foundation on which others can build systems.
What I think would be useful is for NAR to foster a discussion among brokers, agents and MLSs regarding the Open Web and what that means for real estate. This same discussion is occurring right now with regard to the web as a whole, and Brad Neuberg recently suggested: “If we take the long term view, how can we give the web an open enough infrastructure to evolve over time and meet each generations needs, while maintaining its structure enough to actually mean something and stay true to its promise, similar to the U.S. Constitution?” He emphasizes that this isn’t so much about specific technology but rather the general philosophy: “if we define the Open Web in terms of [specific] technologies, then we risk losing sight of what makes the web special and being able to have the intellectual nimbleness to evolve the infrastructure of the web. . . . We will be fighting yesterdays battle while allowing new, proprietary technologies to take over if we focus on technologies rather than philosophy.”
This is where I think NAR can provide leadership, by fostering discussions around how aggregated real estate data can be made most valuable in an ever changing world. IDX has been and remains one of the best tools available to agents and brokers for engage with customers on the web today. Is it time to revisit the IDX policies of old? Are the same questions and controversies that arose over VOWs in the DOJ litigation still a concern? Or is it now possible to redefine IDX in a way to make it even more useful? These questions are left unanswered by TREC, for good reason no doubt, but I think they remain the core questions.
In Part I, I discussed how (and if) consumers’ search for listings is related to their selection of an agent. Questions were raised about whether leads from listings are converting to customers, especially from an MLS listing portal like HAR.com or any site that promotes the listing agent, given the challenge of single-agent dual agency and (or any listing agent). What does the consumer want? What is their natural decision-making path to selecting an agent?
According to the NAR 2007 Survey of Home Buyers and Sellers, “Forty-one percent of sellers found their agent as a result of a referral, while 23 percent used the agent in a previous home purchase. Similarly, 43 percent of buyers relied on referrals to find an agent, while 17 percent of repeat buyers used an agent from a previous transaction.” In a year earlier survey, “7 percent [of buyers] found an agent on the Internet.” These stats would need to have changed dramatically to avoid the conclusion that consumers don’t choose agents based from listing searches or from the Internet at all. Rather, as then NAR President Pat Coombs said, “Real estate is very much a face-to-face people business”.
At the same time, the web undoubtedly is providing new ways for people to meet each other and the value of a relationship is often being tested by data. More sites are exposing consumer reviews, pricing, and other data to the consumer to help them make decisions about which agent to choose. There are sites like HomeGain, Homethinking, Agent Scoreboard, Agentopolis, and Incredible Agent that provide a variety of agent search capabilities. Some sites, like HomeGain, provide ways to compare agents on criteria like commission rates, years of experience, and consumer feedback. Others, like Homethinking, focus on productivity statistics, like homes sold and price ranges.
With little question, sites that provide more information to consumers are a good thing. Consumers clearly are looking for short-cuts to decide which agent to choose and, as noted above, the current short-cut is the personal referral, born of trust, whether deserved or not. The efficiency of the personal referral appears hard to beat. With just a few words, backed by personal trust, your friend or relative is able to communicate a wealth of complex information into a decision.
As web 2.0 companies slice and dice the data, however, I wonder if that efficiency can be matched? Is there an algorithm that will truly help consumers find the right agent? The 2006 NAR survey referenced above found “the most important factors in choosing an agent for buyers are honesty and integrity, followed by the agent’s reputation. Other important qualities buyers value in an agent include knowledge of the purchase process and responsiveness. For sellers, the most important factor in choosing an agent is reputation, followed by honesty and trustworthiness.” Given this, one would think sites providing consumer feedback about their experience with an agent would be very valuable, but does the wisdom of crowds math apply when the number of referrals an agent may get on-line in any given time-period is pretty low? Will one or more sites gain enough critical mass that this data can be aggregated in a meaningful way?
Perhaps more importantly, is it possible to synthesize this data into a “score” or “rank” in order to provide the consumer with a recommendation? This is what I’ll refer to as the tyranny of the ordered list. Whatever the method is for the agent search, the output is an ordered list, with someone coming out an top, just like we see with Google search results. The search algorithm is designed to bring the “best” match to the top. Even leaving aside the fact that many of the sites linked above have revenue models that create conflicts of interest to place certain agents near the top of the search results (or at least to the side like Google AdWords), the reality is that matching a specific consumer’s needs to a specific agent’s qualifications remains ridiculously complex.
Are these ranking sites really helping consumers? They are providing more data but is the data useful? The power of a ranked list is daunting, because it provides an easy short-cut. Why look at agent two or three when there is a number one? Yet is that ranking really anything more than arbitrary given the complex factors involved? Does the ordered or ranked list cut off due diligence when it really should just begin?
Perhaps the natural path for agent selection on-line is through social networking. From general sites like Facebook, LinkedIn, and MySpace to real estate specific sites like Trulia, Zillow, PropertyQube, ActiveRain and many, many others, the opportunities to meet people on-line is growing at a rapid rate. To this end, however, I think many are seeing panaceas where none exist. Just the other day, Dustin linked to a post from Curbed about a consumer being freaked out by their agent trying to befriend them on Facebook.
In this regard, the social networking tools Trulia and Zillow have provided surrounding listing content seem like a promising way for agents to build trust among consumers, but that brings us right back to the tyranny of the ordered list. Jonathon Dalton and Jay Thompson have been posting for some time about the ranking Trulia provides (or provided?) based on the highest number of answers, which resulted in a bunch of agents providing lots of answers of questionable quality in areas they knew little or nothing about.
In the end, modeling the consumer selection of an agent on-line is tricky business at best, and the personal referral is likely to dominate for some time to come, and, in many ways, I think this is a good thing. I consider, for example, brokers like Jay Thompson who just went independent and is building a great brand on-line through his blog and other sites the cream of the crop as to how an agent can communicate their value proposition to consumers. That value proposition will be very difficult, if not impossible, to measure or rank, but the web makes it possible for consumers to connect anyway.
I was talking on the phone the other day with Alex Chang from Roost and, through a broad-ranging discussion, we touched on the question of how consumers select an agent. I’m very interested in honing in on how listing search is relevant to agent selection. I mentioned that last week when I was at the Clareity MLS Workshop, someone flashed a statistic that said 68% (or something lke that) of agent selections were by referral from someone the consumer knew (parent, sibling, friend, co-worker, etc.).
Upon reflection, that statistic wasn’t too surprising, because that’s how so many decisions are made. When I moved to Fargo from Minneapolis ten years ago, I selected an agent in Fargo from a referral from my brother. Think about this: I was coming to work for an MLS software vendor used by all the agents in Fargo, such that I had available to me all the data I could ever want on the productivity of every agent and yet my decision was made based on a referral from my brother. Why is that? We’re all looking for short-cuts. There is so much information available to us today that we need short-cuts, a way to synthesize the data and tell us what to do. If the 68% statistic is accurate, apparently those we already know and trust are just such a short-cut.
The last few days, I’ve been discussing with Greg Swann, Mike Farmer and others the value of a “seal of approval” as a short-cut for consumers. Greg’s theory is that we need an Underwriter’s Laboratory for real estate agents. Brian Boero from 1000Watt Blog agrees and adds the possibility that agent review sites like Homethinking, Agent Scoreboard or Incredible Agents might help fill the gap, too. The commonality of these endeavors (let’s call them “agent search” or “agent recommendation systems”) is to provide the consumer a short-cut to deciding which agent to choose, as well as providing the agents a path (network) to the consumers.
In many ways, agent recommendation systems are the essence of the web movement, exposing more information to individuals and giving them more control. The questions I want to pose in this post, however, are two: (1) is agent recommendation related to listing search; and (2) is agent recommendation likely to be ineffective or maybe even too effective. I’m going to discuss the first question in this post and the second question in a later post.
Let’s go back to the beginning of the post, where it was posited that a large number agent selections are made by recommendation from people we know. In light of this, listing search would seem to be quite separated from agent search. If a consumer is looking for listings, have they already found an agent or are they wanting to find listings first and then find the agent? Perhaps more important, is there a path from listing search to agent selection? In many ways, the two seem in conflict, because the listing search is going to identify listing agents representing the seller and not the buyer. (VAR Buzz has been conducting a great discussion on single-agent dual agency, showing the inherent conflicts in that practice.)
This raises a significant question about MLS listing portals that I’ve been pondering a lot lately. During the Clareity conference last week, I was once again enamored with Bob Hale’s presentation regarding HAR.com. Chris McKeever from CRT was there, too, and he posted some details regarding the presentation. Clearly, HAR.com is driving a lot of traffic back to listing broker sites. The question I have, however, is what is happening to that traffic? How does a listing agent convert a buyer inquiry on their listing into a client? Presuming most are not practicing single-agent dual agency, the most obvious answer would be that they refer the inquiry off to another member of their firm. Is that what the consumer wants? I’d love to see more data about how leads from MLS portals convert to customers.
Listing search does seem related to agent search on IDX sites, where the site owner is promoting mostly listings that are not their own. Yet, the question remains, is a consumer looking for an agent when they are looking for listings? In other words, are they going to pick an agent from their IDX site? This question is raised most prominently by Roost’s model of a national portal rotating IDX sites. When a consumer goes to Roost, they are directed into an IDX portal of a particular broker and inquiries on specific listings are directed to that sponsoring broker’s web site. So, the question becomes, will the consumer value the tools being provided by that broker enough to select them as their agent? Rephrased again, will the click-throughs convert to leads and the leads to customers? That question likely will be answered in the near-term as Roost and other IDX vendors collect data on these metrics, and I look forward to learning more.
Unfortunately, I have more questions than answers about how or whether listing search relates to agent selection. The path certainly doesn’t seem direct and yet listing content remains the gold everyone seeks. The question is what is the most natural path for the consumer to take from listings to agent selection, or are the two really distinct?
Tomorrow, I’ll turn to the second topic I raised above, namely whether agent search or recommendation sites are a more natural path for consumers to select an agent. One tentative title I have for the post is The Tyranny of Ordered Lists.
