MLS Systems and Gateways, Front, Back, and Middle

Jul 10, 2007  |  Michael Wurzer

There’s lots of discussion and confusion regarding NAR’s Gateway proposal. Is it an MLS or not? Whoops, there’s the first mistake. What’s an MLS? Clearly, the MLS is more than just the software that tracks the listing data, but the Gateway isn’t concerned with that, so I think it’s safe to qualify the dicsussion to MLS systems. But what’s an MLS system?

For this discussion, I think there are three major parts to an MLS system, the back, middle and front. The back-end is where the listing data and images are stored. The middle is where the business logic typically is stored. The front-end is what the end-user sees, typically through a browser of some sort.

Part of the confusion regarding whether the Gateway is going to replace the current MLS systems or not stems from the lack of distinction among these three core components to an MLS system. When the Gateway is mentioned as a replacement of the MLS, the whole enchilada is involved, front, back and middle. This expression of the vision entails a new web site that all agents would use for entering listings, CMAs, prospecting, and everything else MLS members do today with their MLS systems. People enamored with this vision see sites like Zillow and think it’s not far from what they might need for an MLS and so NAR, with all that member money, should be able to whip something up like it in no time, particularly since Move already has, Top Producer, etc., under their belt. Knocking off the MLS space should be a cinch. Well, that’s the theory anwyay, but I don’t think I’m going too far out on a limb to say that vision is long-term, at best. So, for now, we’ll give that the Warren Koeller chuckle and move on.

When proponents of the Gateway say that it won’t replace the MLS, they mean that it won’t replace the front-end, at least right away. In at least one vision, the back-end will be a RETS server capable of being accessed by any RETS-capable front-end. In this scenario, the local MLSs may continue to play a role in providing this front-end to their members, but the back-end will be one big data repository of all the listings, images, tax records and more.

Now here is where it gets interesting, and where I think the Gateway proposal may be flawed, fatally. First, when one software vendor creates the front-end and another creates the back-end (even a standardized back-end), you quickly end up with data in more than one place. The front-end vendor works with their clients to provide them with a competitive advantage, something new. Yet the repository is common for everyone. So, to create that advantage, they start adding data locally, accessible only to that front-end or the users of that front-end.

Here’s a simple example. Let’s say a repository is created and, as is common of big repositories, sets a limit on the number and size of photos that can be stored for each listing. Some clever software vendor will come along and create a system to allow users to store an unlimited number of high-resolution photos. Or perhaps the repository doesn’t allow for storage of listing flyers or large remarks fields or some key data field needed for a particular area. That need will get filled. But the result is that all of a sudden all of the data is no longer in the repository. Hmmmm. What was the point of the Gateway again?

I’ve written about this before, regarding Zillow, but the same analysis applies to the NAR’s Gateway proposal. The idea that all the data can be stored in a single repository is flawed. It will never happen, because there will always be some agent somewhere who wants to differentiate their service and they’ll start collecting data outside the repository. That’s what innovation and competition does, it differentiates. (This not to say that a national repository of core listing data isn’t useful. In fact, I’ve previously suggested that a distributed national repository could be very valuable and could foster competition instead of inhibit it.)

More important than content created by agents, however, is all the content related to properties created by others outside of the real estate sales industry. Every day, new information is being created and the creation of a repository won’t harness that creativity. Home owners, neighbors and more will create valuable information about specific properties. How do we harness that information?

Back to standards, this time highly focused. We need standards for identifying properties. With a reliable standard adopted widely, content created anywhere could be linked together simply by properly identifying the property. All the current systems (address, parcel number, latitude/longitude) have challenges. We need a reliable standard. Working to establish standards is something that large organizations like NAR can excel at where individual competition often cannot. Moreover, a reliable property identification system really is a necessary pre-cursor to a gateway of any kind. But, ironically, with a reliable property identification system, a gateway likely isn’t needed.

10 Responses to “MLS Systems and Gateways, Front, Back, and Middle”

  1. NJ Fritz says:

    Thanks Mike, for your comments and perspective.

  2. S L Gladman says:

    Great information, would love to hear the follow-up and comments.

  3. Robbie says:

    Given your last paragraph, are you advocating that the NAR (and other interested parties) get together with local governments, (and anybody who might have public geolocation data) to help draft a nationwide property identification standard or merely an industry wide standard?

  4. Jim Duncan says:

    One argument that I have not been able to defeat is this – the MLS’s primary purpose is to facilitate cooperation and compensation amongst Realtors; it is not designed to be the primary data source. If this type of thinking predominant around the country, then all this talk of a Gateway may be moot.

    I think that the discussions about the different parts of an MLS or Gateway are worthwhile, but we need to sort out the fundamental “what is an MLS” question first – is cooperation/compensation a necessary function? Should the Gateway have all properties? (YES) If it doesn’t have all properties, just those entered by Realtors, then sure, we as Realtors are shooting ourselves in our collective feet (again). By focusing only on “our” inventory, we are unnecessarily limiting the scope of what we intend to do.

    I’ll bet that logically, the logistics and technical aspects are doable. It’s changing the mindsets that may be nearly impossible.

  5. Robbie: Yes, exactly. Again, that kind of standard-making among a wide variety of parties is one area where large organizations sometimes can be effective (not often, but sometimes). Perhaps NAR could lobby the federal government to create a national property ID system that translates all the local parcel IDs that already exists, sort of like the domain system translates the IP addresses.

    Jim: If you’re talking about the offer of compensation being an obstacle, I don’t see that as a show stopper. There are regionals that have and are sprouting up all over the place that pretty quickly agree to extend the offer of compensation to the entire regional membership. Extending that nationwide isn’t impossible. Or, the system could be designed to only make that field viewable by those members in the local MLS. Or, as you and Mr. Swann (and others) advocate, perhaps the entire offer of compensation will go away (not likely, as evidenced by the increased use of the offer as the market has slowed across the country). Any of these three possibilities and others for the offer of compensation could co-exist with a national MLS system (ala Gateway).

    If, however, you are focusing on the cooperation aspect of an MLS, I definitely think you’re right — there will be many who find little value in cooperating (compromising) with people in completely different markets from their local market. I’ve seen many times with regionals (and it will only be worse with a national MLS) where members are subject to rules that seem arbitrary to them because they were designed for conditions in other markets than theirs and so don’t make sense in their market. There are dozens and dozens of rules that vary by locality. We serve 100 MLSs across the country and, over the years, we have added hundreds of business rules that can be configured by each of our clients to their own needs.

    Here’s just one example to give you the flavor of this: We have several MLS clients on the coast of Oregon, where many of the properties are expensive second homes to people living elsewhere who are often hard to contact. As a result, these clients need a different rule for how long a listing can be expired before it can be extended (renewed) than many of our other clients. This is just one example of many, where the local needs give rise to local rules. To put this back in context of my post, this is the middle tier or business logic part of the MLS system.

    I’ve written before how getting agreement on the rules of cooperation is critical to the function of an MLS. This is especially true given that the MLS membership is made of competitors. Adding to that the local needs and the very real differences among markets across the country, and it becomes very difficult to create agreement on the business logic or middle tier on a national level. The result: those put upon by the large organization that doesn’t see their market individually decide they are better off on their own and break out of the regional.

    Basically, the bottom line is that you are right, the “bit of representational democracy” I discuss in the post linked above is much, much more difficult at a national level than a local one. Where I think I might disagree with you, however, is in your implication that this is somehow unique to an outdated mindset of “Realtors”. While there undoubtedly are many local rules that are different for no good reason, there are just as many local rules supported by very real local and sound local needs to ensure cooperation among competitors in their local market. If the Gateway middle tier doesn’t handle these local needs, it will fail. (I was hoping to save the entire “middle tier” thing for a second post, but I’m glad you brought it up. Perhaps I can elaborate on the concepts further in a new post anyway.) 🙂

    So, yes, there are issues suggesting the gateway logic is flawed at all three tiers of the MLS system, which, importantly, only mirror the business use cases established by the MLS. I believe many in this discussion conflate the MLS system and the MLS. They either think the MLS data (back-end) is the MLS or the MLS rules of cooperation (middle tier) are the MLS or access to the data (front-end) is the MLS, but, in reality, none of those are the MLS. The MLS is made up of people who agree to work together. The success of the gateway, therefore, will hinge on whether the system models what the people need. That’s really my whole point. A system needs to model the needs of the participants. At all three levels, I question whether the NAR’s gateway proposal does that.

  6. Robbie says:

    Any opinion on data quality? I think street addresses could work (it works OK for the postal service), if there was some way to encourage members from make the copious errors that find their way into the MLS database.

  7. Addresses are fraught with problems for allowing computers (as opposed to people) to identify properties concretely (which is why you see the disparate or lack of data referenced in your post), and most MLS systems do a lot already to try to translate addresses into parcel numbers and lat/long in order to identify the property with more certainty for functions like history (tying multiple listings for the same property together). The reality, though, is that geo-coding addresses is far from 100% reliable, for many reasons, including the most basic, which is that the street networks are always dated to a degree and don’t have maps for many newer developments. So, even if you can present a map for the user to position the property (which we do in our system), it’s not always going to be accurate because the map may well be blank (no streets) for the development being targeted. That’s just the nature of the massive data collection process surrounding street networks. The inability for street networks to be completely up to date is one of the reasons we’re focused so hard on getting parcel databases in our system, because geo-coding against those is so much easier and more reliable. But even parcel databases have their challenges, especially as you start going across the country. What’s really needed is a nationwide standard for parcel (and, within parcels, units) identification. Think of this like a primary key in a database, it has to be unique and clearly defined, and addresses are not. The zip+4 goes a long way to making addresses more reliable, but those do not get to the unit level. This is a complicated issue with potential for high positive impact and that’s why I think NAR would do well to focus resources on it.

  8. Robbie says:

    My frustration as an IDX vendor is agents do things to make a computer’s life more difficult (bogus zip codes, misspelled street names, no house numbers for addresses etc.). It bothers me people in the industry want a great data integration story, but can’t be bothered to enter complete and correct information for their listing and keep their data “clean”. I’d prefer to get the easy things fixed before tackling more complicated issues.

    I agree that a “Universal Parcel ID” is great idea, but I think for it to work it would end up being like assigning IPs, domain names, or phone numbers. For it to be universal, somebody would have to agree to play the role of NANPA or ICANN, and I have no idea who or what would be the best steward of the primary keys.

    To avoid the out of date map issue (blank streets) w/ parcels, you’d probably have to a hosted solution, which would probably add more cost than value.

  9. […] The article concludes, “Yet another movement for change comes from efforts by the nation’s county recorders to agree on a uniform way to store property records online, which could severely curtail the need for title insurers.”  Whoa, that perked me up.  Some time ago, I wrote that the NAR should focus on working with county governments to establish a standard propert….  […]

  10. […] In fact, the only access to the data will be through one of the MLS systems; there will not be front-end provided by CARETS, it is only a repository of the […]