mattsretechblog: matt cohen (Default)
2016-11-02 12:00 am
Entry tags:

Computer Vision and Improving Real Estate Search

Artificial Intelligence (AI) Creates New Opportunities

Last week I was on a panel at the RESO conference where we talked about software personalization
as an important trend. Back in 2008, I wrote some articles about improving prospecting and real
estate, one aspect of which was that we needed to get smarter about understanding consumer preferences so that consumers don’t have to page through so many listings or can at least see the most likely matches to their interests first. As I noted in that article, the tricky part was that “there are various qualitative aspects of property selection that we don’t currently track data for at the current time.” That’s where “computer vision”, a technology that is becoming both more robust and more common, could make a difference.

“Computer vision” is the ability of a computer to analyze photos and create data out of them. Imagine a consumer likes open floor plans, modern kitchens, wide driveways, high ceilings, mature trees, or lots of natural light – those are all things a consumer might mention when describing their dream home, but little of that information is reliably tracked by agents in most MLSs. With computer vision, that data could be extracted from the listing photos by a computer as keywords by which listings could be searched – without having to manually sort through many homes and many more photos.

I recently saw an example of computer vision demonstrated by RealScout (not a Clareity client) that I was impressed by. I don’t think any company has fully leveraged the potential of computer vision and created the “perfect” product with it, but this company had clearly made some real progress on applying computer vision to real estate search. During the demonstration they showed how they could automatically tag photos, so a consumer could, for instance, page through just the kitchen photos of multiple listings – even if the photos weren’t labeled “kitchen” by the MLS. The technology enables searches for normally unsearchable criteria, and to compare images for key features and rooms side-by-side and roomby-room, as illustrated below:



Images used with RealScout’s permission.

We’ve watched computer vision evolve over the past fifteen years or so – Google’s image search was launched in 2001 and has continued to get more sophisticated, and there are many other companies outside our industry that specialize in it. RealScout is not the only company using computer vision in real estate – during their demonstration they showed how Trulia has used it, and I’ve seen others explore this area over the past few years though not always release the resulting products. There are also quite a number of companies outside of our industry that license computer vision technology – of course, it would have to be optimized for real estate use. And there are some limitations to what the technology can do at this point, especially where listing photos are limited. No one wants their client to miss out on a home because the pictures didn’t highlight a certain feature. That said, this technology – if used artfully – can certainly augment existing listing search technologies and create a compelling user experience.

I have no doubts that our industry will continue to evaluate how to create great listing entry and search experiences using computer vision, and that the number of products – both existing and new –leveraging this technology will grow over time. This is certainly an area to keep an eye on.I have no doubts that our industry will continue to evaluate how to create a great listing search experience using computer vision, and that the number of products – both existing and new –leveraging this technology will grow over time. This is certainly an area to keep an eye on.
mattsretechblog: matt cohen (Default)
2014-01-21 12:00 am
Entry tags:

Beating the Drum in 2014

Professionally, I tend to operate in the background over long time-frames, and I’m happy with the improvements I’ve been able to make with that level of influence. But there are some issues on which I’ve been waiting too long for change, no matter how long I’ve been beating the drum. So, I’m going to push a bit more here in public, on the blog.

More Complete Data Standards

Alternate APIs to move real estate data have been springing up like wildfire, and while few developers have taken the bait and fragmented their development efforts, this has been a distraction from putting full focus on RETS. Thankfully, RETS is finally making good headway based on incredible volunteer efforts. The “data dictionary” is the holy grail of RETS: one set of fields and enumerations to code apps to. And RETS is now about more than just listings; it includes contacts, searches, and more. There’s more to do to improve the standard, but the industry at large is unaware of the progress that has been made, so it has not been demanding faster standards adoption.

Get involved! Join RESO and read up on progress at reso.org and push for standards adoption ASAP. Then we will see easier transitions between MLS systems. We will develop and bring new products into the marketplace more easily. We will be able to move visitor profiles and preferences from an MLS public website to a broker site and/or the MLS system itself. The promise of RETS will be fulfilled.

Client Collaboration

It’s time to expand online client collaboration past where we started and pretty much stopped innovating in 1998.

When it comes to collaborating with buyers, sure, agents need to interact online as they do today around listings. But they also need a more robust search and better search result content, otherwise they’re pushing the client back to the advertising portals. I’m talking about neighborhood info, school info, public records info, and WalkScore. I’m talking about enhancements like lifestyle search. Relevant market trends and statistics for the client search areas (DOM, Inventory, List/Sell Prices, etc.) would also be compelling content. Messaging is also important; having proper alerting options, including email, text message, and phone is key to help ensure agents provide timely responses to questions. I’m all for the collaboration portal offering a way for agents to provide documents to their clients, but I think that these documents need to be more collaborative . For example, financial worksheets need to be interactive and provide alerts in both agent and client directions when a new version is created by “cloning” an old version for editing. Open houses visit planning, note-taking, and feedback should either be built in or deeply integrated. There’s so much more – this is just a starting point for how agents and buyers could be collaborating. And don’t forget that this needs to be mobile-device-friendly.

What do I look for when it comes to agent collaboration with sellers? Obviously, there is messaging, open house activity and feedback, seller-oriented market statistics and updates (i.e. new similar listings and price reductions), interactive and collaborative financial worksheets, and an agent activity log – the flip side of the buy-side functionality. There could be a reverse prospecting tool with “what if” capability – allowing the agent and seller to explore what would happen if improvements were made or the price were changed. There could be an interactive marketing plan and materials, including the listing’s location on advertising portals and metrics for advertising effectiveness inside and outside the MLS. One thing I definitely would love to see is the provision of CMAs, AVMs, and associated financial worksheets that allow for easy change and new versions over time as the market and comps change over the life of the listing.  Again, everything the seller needs to know, and all of the service the agent provides the seller, needs to be accessible from ONE mobile-friendly client collaboration portal.

Business Intelligence (BI)

Brokers and agents need to receive business intelligence from their websites and apps and learn as much about their site visitors as Amazon and others learn from their visitors. I don’t mean just general trends, or raw data about a specific visitor – but actionable intelligence. We should be able to tell, from how visitors use a site, where they are in the lead process and what kinds of listings they are really interested in. For starters. I’m just going to point the way here – not spell it out.

The “Cs” – Community, Curation, & Crowdsourcing

It has taken years, but finally MLSs and brokers are starting to add community information to their systems. MLSs and brokers still foolishly leave crime information out, forcing consumers to go elsewhere for it, but overall community information is much improved. Nonetheless, what has been missing is curation – information reflecting the professional’s knowledge both of the community and the client to create a great client experience. I’ve illustrated this before on my blog showing listings on a map of a neighborhood:
  •     where the best food shopping is
  •     hardware/home improvement
  •     the nearest bank (that the client uses)
  •     the restaurants that meet the client’s taste
  •     the businesses that match up to their hobbies (the judo dojo for Judy and the gym for Jim)

and so forth.

Smart organizations should be crowdsourcing from within (at least!) to create libraries of rich “user generated content” about their community that can be leveraged toward this kind of experience. Business intelligence can be derived from how the clients use this information.

Agent Ratings

Stop complaining about agent ratings! They are going to happen. The question is whether they will happen the way you want them to. So, if you are a broker, it’s time to make up your mind. Do you want the MLS/Association, where you will have control over the process and rules, to handle ratings, or do you want someplace else, where you will have to take what they give you, to do so?

Changing Role of Realtors

We’ve been talking for years about the Realtor role changing, moving away from the transactional role and evolving toward a longer term role in clients’ lives – a financial-planner-type role or something else. I think we need to fast-track that conversation about what we want Realtors to be in 10 years or in 50, and change the culture and use technology to support that role.

Bang!

So, if you are a software provider in our industry, it’s okay to listen but also it’s important to lead. And if you’re a customer of a software provider that is primarily “customer-driven”, be a smart customer and start pushing for what you need. The industry has a lot of challenges coming up over the next few years. These are just a few of the items I have my eye on that, if executed, could provide a positive benefit for the industry overall. I’ll certainly be bringing up some more at the MLS Executive Workshop in March. Let’s all keep pushing together to make the best possible future for ourselves!
mattsretechblog: matt cohen (Default)
2012-11-05 12:00 am

Real Estate Document Management in 2012

Clareity Consulting has always been of the opinion that online transaction management –especially document management – has been a critical technology for the real estate profession.

This paper examines the real estate document management space in 2012, how the solutions providers and their products are differentiated, ongoing challenges to success of the category, and opinions from providers on where document management is heading.

Download the full report:
Real Estate Document Management in 2012
mattsretechblog: matt cohen (Default)
2008-06-18 12:00 am
Entry tags:

Future of MLS Features (2008)

A Look Ahead – Far Ahead

Introduction
The purpose of this paper is to generate discussion on possible MLS system future features by providing a big picture view of the changing relationship of real estate professionals with each other and with consumers, the changing relationship of local and regional MLSs with each other, and to illustrate, at least at a high level, how these changes may be either enabled or reflected technically in the MLS system of the future.

This paper is not focused on detailed description of what features are popular already today, for example:
•    Mapping bird’s eye or street-level views Big pictures in slideshows and flyers
•    Total MLS staff control over fields, reports and business rules Easy setup/management of RETS data feeds
•    Single Sign-On (SSO)
•    Public records data intermingled with MLS data in reports and improved statistics

This paper also does not focus on the usual incremental changes to current MLS features, but rather explores the future of MLS systems and their role further ahead.

Clareity always advises our clients during their MLS system selection process focus on the core features (‘the steak’) and not be overly sold on other features (‘the sizzle’). Too often, a largely volunteer based Task Force can be swayed by a single ‘sizzle’ feature and forget that most importantly the system must perform core functions such as listing input and search as efficiently and accurately as possible, and that the system must have high availability and fast performance. With some of the more popular MLS vendors currently having significant issues in these core areas, I want to make sure that this paper is not seen as a call to take your eyes off the system core. That said, the definition of core functionality has expanded somewhat in recent years and will continue to expand and change – and we can’t ignore that either.

By consulting for many MLS vendors over the last decade, Clareity has strongly contributed in the development of the product vision for today’s modern MLS system. Clareity was a strong proponent of features such as integrated contact management and CRM, functionality for assistants and teams, and coordinating all of the leading real estate software vendors on Single Sign-On (SSO) technology and information security improvements. Not every feature we’ve thought up or recommended has been adopted though. Some ideas, such as good uses for automated valuation models (AVMs), Clareity has advocated for many years, but it took Zillow and Zestimates® to serve the MLS and brokers a wake-up call. AVM’s are just now starting to be integrated properly, in just a few MLS systems, using high quality AVM tools from companies like First American and Cyberhomes.com.

What follows in this paper are some of the cool features from my MLS product development notebook. Hopefully some of these features will show up in your MLS system of the future. If you like one or more of these features, ask your vendor to for them (or build it yourself, home growers!).

Mapping: Not Just About Showing Property Location

Mapping has currently been used in MLS systems to show the location of properties, and occasionally through data layers and other interactivity, to show information about the property and its surrounding areas. However, mapping has a lot more promise than it has been used for currently.

Mapping can be a great tool for communicating agent knowledge about neighborhoods and communities. In some systems there is currently a way to turn on specific categories of “points of interest” (POI), but does it really help a gourmet seeking a home in a high-end community to show them every McDonalds and Burger King in a two-mile radius? Not at all – rather, if the agent shows the consumer that map, it demonstrates that the agent doesn’t understand their client. It definitely doesn’t show the client that the agent is the neighborhood expert and can help interpret the plethora of information available. So, one key feature for turning maps into a useful tool to build a bridge between agents and consumers is allowing the agent to customize the map, edit the content shown to the consumer and add user generated mapping content.

Illustrated below, an agent is showing the listings desired by the consumer alongside some specific restaurant and shopping options. You can see that in the Bistro detail shown, customized text and additional information has been entered by the agent, showing the client that they know the neighborhood, and have been to this restaurant before.

user generated map content

For example, if the prospective buyer had a child that studied karate, the agent could have added the nearby dojo to the map, along with the commentary “I think Suzy will really like the karate instructor at this dojo.” Or, if the buyers had children in elementary school, the agents could add rich, relevant and even personalized content about the local schools as well.

The key to successful user generated mapping content is for it to be very easy to add the content. It must be easy for agents to add new custom points of interest, pre-fill basic information from existing data sources, and create content libraries that they can leverage to create custom maps for consumers with a minimum of entry or re-entry. Getting these workflows right is critical to feature adoption.

Another area of mapping that could be greatly enhanced is to use mapping layers to show demographics. In many surveys Clareity has performed, agents seem very skittish about this – especially when it comes to showing crime maps. Some agents have legitimate reasons for skittishness – fear of being accused of steering or other violation of fair housing laws are valid concerns – but it’s up to real estate professionals to provide the consumer the information they want and need to make a buying decision. If consumers want it and the real estate practitioner won’t provide it, they’ll get it elsewhere and the value perception of the REALTOR® will continue to decrease. As former NAR president Billy Chee said to me back in 2002, “The consumer is the lion coming over the hill.”

Mapping also has great power to display complex information in a way that’s very easy for people to interpret. One of my favorite visualizations is the ‘weather map‘ or ‘heat map’. Consumers can readily obtain heat maps from Trulia, Zillow, CyberHomes, and others, but not from their agents. Why is this? While some MLS systems already have heat maps to show days on market or price per square foot, it’s easy to imagine other heat maps with even more useful information. The example below shows what areas are ‘hot’ or ‘cold’ for investors by showing appreciation over time. Such maps could also show vacancy and absorption or even percentage differences between initial asking price and final asking price or sale price, or even show shading representing the percentage of properties in foreclosure.

heat map showing investment desirability

I’ve shown ‘heat’ two ways on the map above – with colored icons and with color shading. It’s probably only necessary to use one method or the other. Icons will certainly be technically easier to implement than shading, though at a wider zoom level area shading may make more sense.

Bridging the Gap between Internet and Installed Software

Why make the consumer open up a web browser and go to a web site to see their latest prospect matches? Why even expect they would check their email? Why not ‘push’ the results right to their computer desktop and get the information right in front of them when they start up their PC?   This is both convenience to the consumer and value-add to the agent.

The illustration below shows two Widgets that I created back in 2005 – one designed for the consumer showing the results of a prospect search, the other for the real estate professional, showing listing activity in their market area, along with what emails, inquiries, and tasks would await them when they logged into the MLS.

desktop widgets

Toolkits by companies such as Google and Yahoo!, as well as widget capabilities built into Windows and Mac OS, make widget creation fairly easy. Coldwell Banker added a very simple widget to their web site last year, but I’m imagining much more sophisticated widgets, especially for professionals. Recently, I’ve begun to see capabilities developed to allow even more bridging between the Internet and the desktop – where the widget can store some data locally and provide some functionality even if the user has gone offline. As this technology evolves, I expect that the opportunities opened up by its use will continue to grow.

Integration of Broker System Features
At some point I expect, or at least hope, that MLSs will have much deeper integration with broker back-office systems and/or build in more broker features. There would be significant broker data management and workflow advantages to building features into the MLS such as:

•    Lead Generation / Management tools
•    Marketing tools
•    Competitive analysis for Recruit/Retention
•    Content syndication tools (listing distribution to other web) Productivity / profitability measurement tool

To dig a bit deeper in one of these areas, an agent productivity / profitability measurement tool may include such elements as:

•    Income and Expense Tracking
•    List/sell/total production graph and chart
•    Drilldown by month / week / day / date range Drilldown by enterprise / office / team / agent / listings Productivity modeler (Actual / What If)

The “what if” modeler may allow for adjustable components such as commission splits, selling office commissions, desk cost coverage %, closed to list ratio, average marketing time, transactions to list ratio, and more. The system would then be able to show total $, GCI, agent $, company $, market $, desk $, net $, and $ change (from previous and base scenarios).

These types of features have been in various different broker tools – but really depend on the MLS for the data to properly implement them. Again, either the key will be deeper integrations with existing products or building these types of tools right into the MLS.

Features to Better Support Agents
Most MLS features are focused on the agent, but there’s still more that can be added to the MLS for them, including:

•    Listing presentation or other marketing pieces as robust as the CMA w/ MLS sales statistics and showing data integrated
•    Buyer’s agent presentation
•    Easy mail merge marketing pieces w/ tax data
•    A chart/report showing housing value increase or decrease within specific search criteria – to detect price trends within a specific neighborhood – and the ability to set alerts if sale price conditions start to occur for a specific search.

As MLSs continue to regionalize and engage in data shares, creating a better system for agents to find each other and provide referrals will be increasingly important. I believe that more advanced roster search functionality will be important if an agent in one area needs to be able to find the agent in another area to best serve their client.  Being able to see who is the expert in the types of properties desired by the client and who is most experienced and ‘best’ at facilitating buying or selling those properties via statistical analysis is key. Potentially there would even be qualitative agent ratings, open to other real estate professionals or even the consumer. Like eBay ratings, there would be a way to address disputes. There are already a number of web sites providing mechanisms for agent ratings – why wouldn’t “organized real estate” want this mechanism to be someplace where we could manage the rules around it and have it integrated with other agent information and statistics?  Consumers will have access to several agent rating services – this is inevitable – because everything is being ranked on the Internet.

Integration of Appraiser Data
Will appraisers ever be brought into the fold? Every few years this comes up and new appraiser platforms such as Zaio are developed – though usually they have not succeeded in the long term. Why separate appraisal systems from the MLS system – is there not synergy? Shouldn’t data standards such as RETS be worked on together with appraisers? How will they be incented to participate in a common data platform, so that everyone benefits?

RETS Implementations
Continued improvements in the ease of setting up listing syndication and even accepting listing input from broker systems will be possible as RETS continues to evolve. I think these are core MLS functions, and will change the role of the MLS system as diagramed below. A lot more detail on this subject is available in a separate paper, available by contacting Matt Cohen here.

MLS of Today - and MLS of the Near Future

MLSs will also need to work to address the security of listing data either being syndicated or even exported directly from the MLS. Because of that latter element of the problem, use of secondary products will always leave a significant issue unattended – unless the solution is ‘baked into’ the MLS. None of the MLS systems on the market today have established effective controls for solving this issue, though Clareity Consulting attempted to get the ball rolling by sharing plans for such as a system with all the major MLS vendors back in 2004, in a document titled, “Protecting Against Illegitimate Use of Data by Legitimate Users: Processes of Data Licensing, Delivery, and Use Monitoring”.

The core of the system, diagrammed below, is to include a process for data use licensing, request and delivery, and verification – all built right into the administrative user’s view of the MLS. MLSs could get a handle on where the data should be via the licensing process, data and images would be individually watermarked (yes, I know that data watermarking is a tall order), and methods of efficient compliance management put in place.

API Security

I’ve got to admit that I’m not sure the perceived cost/benefit model will ever make it likely that such a system would be built – but I’d like to see this issue addressed. Once weaknesses in MLS user authentication and protections against hackers are put in place, this area is the largest security challenge for any MLS.

Social Networking
Real estate is, by its nature, a social business – so another area where both standards and deeper integrations may come into play is in social networking. Various major social networking sites have explored development of a common programming interface (API) for social applications across multiple applications – for example the OpenSocial standard.  If MLS functionality expands its capabilities toward social networking, it certainly would be interesting to see how the MLS could interact with other applications through such interfaces, opening up whole new possibilities of how real estate professionals interact with their colleagues and clients.

The Original NAR “Future of MLS PAG” Vision
Originally, the NAR “Future of MLS PAG” vision was to have a central back end data repository, allowing for front-end interface of choice, provided at the local brokerage, MLS or association, vendor, and franchise levels, along with a baseline front end available through the NAR. Diagrammed below, this wasn’t a bad idea, though the MLS PAG has since evolved its vision toward something that has seemingly little to do with MLS.

Universal Listings Database with Front End of Choice

I still think the original vision made a lot of sense, especially at the natural market region level, then being linked together into larger areas. Of course, most MLS systems are not currently architected to use separate back-end databases, but I expect this will change in the future.

Lastly, to facilitate the regional data share process, or even to make it possible for brokerages/agents to have their own custom data shares beyond a single region, MLSs will need to make it easier to automate creation of data mash-ups from different MLSs as much as possible. I imagine a data mapping expert system that facilitates inclusion of multiple data sources, automatically mapping data to a common set and “wizarding” corrections and additional mappings. Of course, the system would still need to reflect the data mapping into reports, statistics, and other parts of the system.

Conclusion
Clareity is constantly researching new ideas for MLSs.  Our expert consultants are regularly engaged in the product management and development process with leading MLS vendors and home-grown systems. Through end-user surveys, interaction with MLS executives and staff (80+ of top 100 MLSs have been clients of Clareity), our annual Workshop and attendance at MLS system sales demos, Clareity is constantly taking the pulse of the industry, in terms of what features are desired in an MLS system. But Clareity goes beyond this research and is always looking ahead.

One of my favorite product-development related quotes is from Henry Ford, great automotive pioneer, who said, “If I had asked people what they wanted they would have said faster horses.” There’s a lesson in that quote for MLSs that say, “We’re member driven,” and for MLS vendors too focused on the mantra, “We’re customer driven.” It’s important to listen, but it’s also important to innovate and lead.

Those who wish to keep the functionality of the MLS more limited may insist that the role of the MLS should be constrained to only those functions needed for the facilitation of cooperation and compensation between brokers. That is, of course, the core of the MLS, but it should also be recognized that the MLS is the core business platform for agents as well, and that the MLS may need to continue to expand to support their needs in a multitude of ways.

What has been described above may be of interest, perhaps may inspire, but it’s up to you. We in this industry often passively ask ourselves and our peers, “What is the Future of MLS?”   I think we need to take a more active, thoughtful role. To reference a quote attributed to Allan Kay of Apple Computer, “The best way to predict the future is to invent it!” MLS vendors and regional MLS operators can create the future of MLS, both supporting and driving the way local and regional MLSs interact with each other and the rest of the industry, and enabling REALTORS® to interact with consumers in new ways, preserving and enhancing their value as well as the ongoing value of the MLS system itself.