Thanks to Pandora we no longer have to search for new music. Pandora searches for us. We set up a play-list on a genre, song or band and new tracks present themselves based on common attributes. When we like or dislike each new track, Pandora learns our tastes. And continues to refine the tracks it adds to our play-list. Extend that matching logic across the entire user base and the app becomes exponentially smarter. More valuable.
It makes other ways of searching for music feel archaic.
Like searching for real estate listings is today.
There’s no question that designers and UI experts have made real estate as pretty as can be. And developers have made it simple and quick. Everything from walkability scores to education data and local blogs have been mixed into the experience. But in the end, searching for homes is still arduous. Users are forced to manually weed through their options.
One of the most ironic facets of real estate search as it is today is the set of fields commonly provided to users at the onset of the process (location, beds, bath, price). For these things are not always what becomes important to them in the final leg of their decision making. How often did your seller not buy the 4 bed, 3 bath, $650,000 home and instead choose the 3 bed, 2 bath $500,000 place because it had a better view, a bigger back yard, a different school district, or a finished basement to die for?
Real estate search today might be pretty, mashed and fast. But it’s not smart. Or intuitive. Or built to do anything other than to provide us a fabulously detailed patch of content we have crawl through.
But imagine a Pandora for real estate. A system that allowed users to tailor a search by liking and disliking the listing results the system provides them based on their broad parameters. By crunching our responses along with those collected by all who came before us, search could be considerably more productive.
The nuances of how this could be created are not lost on me. But if Amazon can do it with books and Pandora can do it with songs, I’m convinced real estate could do it with listings. Especially if we could find a way to suck in the user generated content – the things people are thinking, tweeting and facebooking about properties. Yes, a legal minefield – but one we should attempt to sweep.
This sort of thing would be possible if we placed more energy into what could be rather than protecting what was. Or worse, indulging the real estate molehills this industry loves to turn into mountains.
A map here, an API there – big deal. This is what I’m talking about:
What are your thoughts?
Sign up for the 1000watt Spotlight e-newsletter
and keep up with the ideas, apps and people that are changing real estate.
20 Tools to Bring your Real Estate Business to the Cloud
A mist of difference
Freaking...
It’s got to be frustrating for buyers given the clumsiness of current tools. I know it’s frustrating for me sometimes, as an agent! This is definitely a step in the right direction.
oh i know where this came from.
Hi guys,
I think what you’re describing here is contextual search, something we’re seeing a lot of online and in mobile, but has yet to really take hold in real estate. Filtering methods such as geolocation, past browsing, and most importantly those you know, personally or professionally, are the new drop downs and checkboxes, which I think is long overdue. Your friends, as in real life, are your filters. Having these contextual / lifestyle filters dynamically combined with the ‘nuts and bolts’ search parameters will potentially lead to a smarter ‘suggestion’ engine, one that actually matches your search in a way that’s far beyond a simple algorithm’s result (for example, NetFlix does this and it never really returns films I want to watch) – is something we’d all find very valuable, not just for real estate (I’d love restaurant suggestions based on this).
I think mobile advertising is getting close to what I’m talking about, but applying these ideas to RE search, and moving away from a ‘select your options’ manner of finding anything sounds very exciting.
Whoever develops this is going to be a friggin millionaire. Better patent it before I do, Marc.
That’d be cool. Sounds like the algorithm part of that deal would be real expensive to develop and keep razor sharp. But, man-o-man.
I also wonder why we don’t have some magic that would embed some sorta web-tracker-code into the individual MLS property entry – one code per address. Seems we should be be able to enter web-tracker property code into the Google Search Box and it’d serve up every single place/web url that the property address/data is displayed on the web. Handy tool for agents to track and insure accuracy as well as reports to the seller.
Keep Thinking BIG.
Cheers.
Here’s what I question: the delivery of next results.
I have no patience, so sitting there in front of my machine doing A/B tests with each new result would drive me insane.
How would you propose delivering next results?
It must start with the data fields the MLS requires, and the details the Listing Broker offers up about the property. Example, here in NJ – the square footage of a home is still left open and not a required field, crazy. If the source doesn’t or isn’t required to enter the searchable data how do you build a system to allow for tailored search. Maybe this is where NAR’s RPR system can help the consumer get to tailored searching.
Matt,
I don’t think it’s gonna be possible. How could an engine interpret your selection?
Type of building, location, sq. footage?
It would probably be just like your Netflix experience, I believe.
@Christian
agreed. If it acted exactly like Pandora, that would suck. But it’s not what I have in mind.
@James
I give this idea freely. Would love to see an MLS take a stab it.
@Kevin
anything is possible
Hey Kevin,
I think the idea of contextual search is to remove the process of ‘selecting’ so heavily integrated into the checkbox and drop down search pages of current RE sites. With contextual search (at least as an idea), whatever you’re doing online is contributing to your ‘selection’ – who you talk to, what you buy, what shows you watch etc. as it creates the ‘lifestyle’ content which Marc is describing.
I dont believe that ‘square footage’ is an initial search filter, it’s more something you use to refine results, perhaps. I think neighborhood, style, view, amenities etc are far more important – subjective lifestyle stuff. I think if it’s possible to create a system whereby you integrate dynamic and accurate lifestyle data with location-based real estate information, filtered by social media and financial content, then you have some small steps towards intuitive (and very powerful) search engine capability for matching people with properties.
Remember that the ‘lifestyle matching’ functionality of things like NetFlix, Amazon and even something like Pandora or Apple iTunes Genius is really still in its infancy, and it’s far easier to create a page of checkboxes by way of ‘selection input’ than make something that learns and understands the user on a deeper level (using data gathered from across the web). Privacy issues? Sure, but I think if my social graph was able to make truly valuable recommendations for me, I’d do it.
You have to start somewhere, but I guess my point is that moving away from ‘selection’ is the way forward here.
Love this idea! I think it’s feasible, too. Regardless, though, the concept makes me think about other ways to deliver information with “stickability”.
Love this too (but you knew I would…). Don’t see many reasons it could not be done. The thoughtful comments are some of the solutions. If the right players collaborate (MLS are you listening?), anything is possible. Someone will (or already is building it…), or is going to make something like this happen. Think of Marc’s concept above, combine with the iPad, mobile devices, hyper-local… you get the idea.
It’s funny that you mention this now.
I was on a 7 hours drive yesterday and I was thinking of a very similar system. You seem to be a little further ahead than me in your reasoning but I had to watch out for traffic and such…
I make websites for real estate agents in Montreal and while things are a little different here, it’s something I would really like to implement.
Folks
Think of this application and viewing pane in the context of an email alert system.
I feel you on this, and have been thinkin about something like this for awhile. But how the home data is entered into the MLS is so key. I would be so nice to find a small MLS that was willling to go through the legwork of get the right standards set up… we’ll see
[...] A smarter real estate search – Imagine if searching for a house was like listening to music on Pandora. I am just not sure [...]
I think this concept has merit and it is something that Coldwell Banker has taken into consideration when we released the beta version of our new website. We added a search tool that allows you to find properties based on your general likes/dislikes by simply rating images thumbs up or thumbs down. The tool is called BlueScape.
The more you rate the better your results become but it’s also tied to your preferences from previous visits. You can also narrow the results to a specific area if you choose but the initial idea was for more of a dreaming type of search.
It’s not a perfect system by any stretch but we feel it’s a building block for a quasi-housing genome type search tool that will only improve as we get more info from our site’s visitors. Here’s a link to check it out: http://bit.ly/cDWJaE
[...] From 1000 Watt Consulting – A smarter real estate search [...]
I know how to do this. I’ve done this before in a previous life. I created a decision engine based on IBM Websphere that was used in many large IBM retail clients and for doing predictive selling in a call center environment. Writing the application and developing the rules is easy.
The hard part is getting everyone on the same page…all the MLS boards, all the agents, to come up with a database scheme that will allow you to have fields like “View Rating”, “Convenience Rating”, etc., in other words a rating system that can be boiled down to a scoring system the application can use.
The closest thing I have found to this in real-estate, is an IDX search from a company called BestHomePro (http://www.besthomepro.com/BHP2_IDX_Agents.cfm). It really is not close but it does start down the path with the “standard” real-estate search fields.
It can be done. Someone raise a little capital and I’ll get right on it.
I’m constantly impressed by the selections Pandora chooses for me. Sometimes it’s mind blowing that it chooses the songs it does. I think to myself “man, I haven’t heard this song in like 10 years, but I freakin’ love it!”
Building that experience into Real Estate search would be equally mind blowing. The process, the algorithms the testing… All challenges for sure, but like Marc said – anything is possible.
@David Marine
Interesting tool but that’s not really what I had in mind. Pandora doesn’t ask you to rate a hundred songs before it starts playing music, it just does and learns from what you listen to.
I find that it’s important to start showing properties as soon as possible. The idea is to learn from every property view to help suggest other properties, not replace the current search criterias with pictures of birds.
I find the pictures on BlueScape to be a little too abstract or vague and the system doesn’t narrow down based on my choices.
I voted down on a landscape picture and voted up on a cab in downtown NYC. I would expect the system to understand I’m looking for something in a city but judging by the next photos, it didn’t seem like my message got through.
The user doesn’t really feel like he’s making progress.
Like has been mentioned above it’s all about the data. I could do some really awesome stuff, if there were such thing as uniformity in the data from the MLS. Think of what we could accomplish with a strict set of data! Alas, we’re stuck in a world where some agents decide that one picture is “good enough” or “there’s really nothing I can add in the description…so why put one?”
It will take one of two things happening to really see this blow up.
1. A Local Board/MLS says, “we’ve had enough, we’re going to become industry leaders. If you want to be a part of our awesome MLS, you’re going to have to provide a really high standard of data for each of your listings.”
2. A brokerage with a lot of online weight, one of the national franchises, or someone like Big T or Big Zlo will have to develop this awesome system and way of helping people find homes. It will be so awesome that agents/mls’s will want to have their listings show up and will have to raise the bar on the data they provide as to not get left out.
I hope to see it happen one day. There IS a better way to search real estate. It’s so simple that most people don’t realize what it is (well, those with the proper funding
), but we’ll get there one day.
@David,
At the time Bluescape felt more gimmicky than it was practical. Still sorta does. My interaction with it is more game like than it is practical and intuitive. Not dissing it as I admire the attempts but the system doesn’t seem to learn and/or possess a consciousness. Pandora on the other hand does. The question is, how can we make this happen for real in our business?
I love Pandora! I also would love similar for real estate searches. However I often get frustrated by the lack of quality data and unless critical data was required I would be wondering if I was indeed seeing everything – houses matching my narrow criteria that were available – as many right now slip through the cracks unless I leave some narrower parameters out.
@Marc/@Laurent-
I appreciate your feedback on BlueScape and as I mentioned before it’s still a work in progress. The site is still in beta but I think the direction we’re heading in is what your conveying here in your post. It’s not a perfect match yet but neither was Pandora when it first launched. Kept trying to tell me I liked The Mamas and the Papas (not that there’s anything wrong with that).
I think the big question about a Pandora like tool is how to balance the fun of searching with the biggest purchase of your life. Would a tool like that help me narrow down my search or is it more of a game for the interested shopper as Marc remarked?
Another interesting post. By the way, I read Rework and it is filled with great relevant info.
Aloha,
Keahi
@David
Totally understand and I agree it’s clear we are both heading in the right direction and as you point out, Pandora has been at developing this for a decade. I suppose my sense of urgency comes from the reality that IDX has been around for about that long as has social media and for the most part, real estate search is hardly any different from its offline counterpart – the MLS book. Though it’s in electronic form, we still have to flip through listings, page after page, looking at a lack of photos, marginal descriptions and hardly any real decision support content.
As for the big answer to the question of balance, I think the first step is having people in our industry get their head wrapped around this notion of just how big and important buying a home is and demand that their MLS’s, their vendors and everyone who touches their search platform rise to that occasion and built something better. Now.
Not to overly simplify the matter but with all this negatively that surrounds RPR, when boiled down to a nutshell those in opposition are essentially preventing better data to be made available to the people who need it most – consumer through the delivery of their agents.
There mere reality is real estate, MLS’s, etc., have a supply of amazing data and access to a million agents with street knowledge. Merging this data with technology to make it engaging and beneficial is the easy part. Convincing every power that be that doing it is best thing this industry can do – there were impossibly wafts into the equation.
[...] Davison envisions an ideal system for real estate searches – wouldn’t this be grand? He imagines “a Pandora for real estate. A system that [...]
[...] few weeks ago, Marc envisioned “Pandora for real estate” – a property search app that could surface the perfect [...]