Extate seem poised for success
September 24, 2007 at 7:28 pm | In extate, property2.0, search, technology | 8 Comments
It’s been about 18 months since Byte Play Ltd, the company that owns Extate.co.uk was founded by Douglas De Jagger and Artemi Krymski, two post grad computer students from Imperial College.
Late last month, Byte Play announced that Extate had now received their first major round of funding from The Accelarator Group (TAG), the venture capital conglomerate who had previously backed Lastminute.com, Last.fm and Agent Provocateur. This past weekend Artemi and I had a chance to catch up over coffee and as expected, we ended up discussing some issues surrounding the funding plus and future plans for Extate, now that they have some heavyweight players backing them up.
TAG chose to back Extate over the other property vertical search engines – Artemi claims – because of a crucial difference in their business model over other vertical property search engines. Many bloggers assume Extate is basically the same as other vertical search mash-ups; throw a crawler out, grab some data, mash it up, get a funky logo and and voila, deal done !
But pay closer attention,and you will notice differences, some subtle, some not so subtle, perhaps the biggest being that Extate only produces search results directly from estate agents website and not via third party ‘property portals’ such as Findaproperty.com, Propertfinder.com, or Rightmove.
Nestoria, Oodle.co.uk and the popular OnoneMap.com offer a significant portion of their search results from portals, but Artemi feels that this deteriorates the quality of their search results. He mentioned the popular ‘flyboarding’ strategy that many agents use, when advertising via portals and believes that close to 30% of all listings on portals are actually ‘flyboards’:
“What that does, is when a customer calls in about a property he’s seen advertised, the agent says that the property has been sold and then offers them another one, or takes their details for their database.”
Artemi feels that the information on the agents website is always fresher and more up to date than what’s on the portals and that by crawling agents websites for data, the Extate engine is much more beneficial for consumers. What is also quite unique about Extate, and what probably makes them technically superior to their rivals is the ‘artificial intelligence’ (AI) technology which they employ in gathering data from websites.
The Extate AI analysis tool, not only crawls the web for property data, but is also able to extract useful information out of the sites it crawls, enhancing user experience with the ability to specify price and property location on a particular site. So for example , if a user is looking for a ‘2 bed house over 300 sq ft’ the Extate engine simply does not just use simple text-matching to match user queries to webpages. Instead the patented AI technology permits the Extate engine to extract information such as square footage details and other information from within the estate agent’s website. No two estate agency websites are designed the same and the search parameters vary from website to website.
The AI analysis ultimately answers the users query in a ‘natural language format’, something that no other property search engine in the world is currently doing, a process that Art describes as ‘technically challenging’, but clearly worth the effort if it translates to a greater user experience and increased consumer traffic.
A search on Extate takes the user straight to the source of the original listing, extracting data using artificial intelligence, and like pretty much all vertical property engines, they do it all for free; neither the agent nor the consumer pays for the service. So of course the challenge is to make money and the problem is that right now they’re not making much. At the moment people tend to look at Extate and other vertical search engines with a lot of curiosity, deciding the best of the bunch is probably the one that’s most “fun” but does anyone really take them seriously?
Rightmove, of course has a bunch of traffic and record profits as over 13,000 agents pay a minimum of £250/month to list properties on the portal. But can this model last with the new players like Extate offering such a viable alternative? In today’s downward spiraling property market ‘free’ is a very tempting word for an agent, but the problem of course is that Extate hardly has any traffic and for the moment, don’t seem much of a problem for the likes of Rightmove or Findaproperty, but:
“the portals realize we’re a better model for agents, they’re extremely happy we don’t have traffic but they realize thier approach probably will not last much longer. if not us, then maybe Google, who knows, but the information will be free”.
Things change quickly in the world of 2.0 internet, and things are changing quickly in today’s property market. Extate seems poised to take advantage of this atmosphere in that their technology is unique, and their business model transferable and expandable at relatively short notice.
8 Comments »
RSS feed for comments on this post. TrackBack URI
Leave a comment
-
Recent Comments
Blog at WordPress.com. | Theme: Pool by Borja Fernandez.
Entries and comments feeds.
Hi Harvey – Just a point of clarification. While the Extate guys have done a good job, their technology in this regard is far from unique. We at Properazzi have always used this technology to extract property information from crawled websites, and we do so in all European languages – whether the site is in English, French or whatever. I’m also aware of other property search sites that use the same technology.
Thanks,
Tom Dibaja
Properazzi – the property search engine
Comment by Tom Dibaja — September 25, 2007 #
Tom / Harvey
Can only reiterate Toms views, many vertical and non vertical sites (oodle for one) have the same technology.
Duncan
Comment by Anonymous — September 25, 2007 #
Artemi from Extate here.
Crawling and parsing technologies have been around for ages. Crawling a few large portals is extremely simple, an off the shelf crawler would do the trick, plus a few regular expressions to parse them, no matter what language.
We have “deep” crawling, ie we tackle sites which Google or yourselves can’t handle, eg:
http://www.halifaxhomefinder.co.uk/HxHFHome.asp
If you do, then please correct me.
We also extract attributes from natural language and go directly to source, hence no listings from portals on extate.
I think properazzi and oodle are excellent services that aggregate as much information as possible, and are extremely useful for international property buyers.
Extate’s model is not an alternative to that, instead we’d like to replace the portals that aggregate the data from the long tail. Compared to the portals we’re now 2nd only to RightMove in UK coverage.
If you’d like extate’s search results to be included on your sites, then we’d be glad to have that discussion.
Comment by ak102 — September 26, 2007 #
Tom / Duncan, thanks much for the clarification, Art, I don’t know if I quite understand when you say that extate wants to “replace the portals that aggregate the data from the long tail.” Are you saying that portals in general and Rightmove in particular are benifiting from a long tail model? Or are you saying that extate is taking advantage of their lack of long tail strategy?
One observation about long tail BTW, it doesnt always generate mass traffic for websites, even the longtail ‘godfather’ admits this:
http://tinyurl.com/2tot3a
Comment by renthusiast — October 2, 2007 #
Hello,
My name is Anton Malinovsky and I’m the CTO of Migoa Internet Services, I really appreciate the ammount of job you’ve done, but really do not understand the claim for being unique. Yes, you are a deep crawler, as long as you can pass the forms But we and I’m pretty sure a lot of other engines can do that The site you’ve posted has a post form, so what? all the parameters can be easily looped through. And the location parameter can be hadnled using a predefined list, it is easy. There are no dynamically generated url which can not be aquired from the page, or flash, or cookie control which prevents the giving direct links to your users. Sites with dynamic urls like this http://www.avendrealouer.fr/annonces-immobilieres/petites-annonces.aspx?param=Page;ANC;1P;1;2P;1;3P;1;4P;1;5P;1;TE;1;MA;1;LOC1;92;Piece2;1;Piece3;1;PKG;%20%3E%201;CHA;1;Piece1;1;AP;1;Piece4;1;Piece5;1 can be problematic for crawling, but also possible. You example is quite regular.
Comment by Anton — October 4, 2007 #
@Duncan
Does oodle crawl? I remember reading this:
http://real-estate-net.blogspot.com/2007/07/oodlecouk-now-claiming-more-properties.html
So got the idea that you don’t, pardon me if I’m wrong.
@renthusiast
Portals aggregate data from the long tail because thousands of small agents send them feeds. We want to do this automatically, and aggregating data from thousands of small data sources is technically much more difficult than a few large portals.
@Anton
Indeed site scraping has been around for decades, and crawling is a much easier problem than parsing.
What’s unique about our approach is that we don’t have an individual crawler & parser for every website we scrape. We can mine data from websites we’ve never seen before, like Google does. It’s not a necessity when you aggregate data from a few portals – as you can find the time to manually specify how to crawl and parse them, but becomes a must when its hundreds of sites that keep changing.
PS well done on muroa, looks nice!
We believe that agent’s should own their information, and users should see the original, rich and up-to-date details on agent’s websites, instead of being locked in by portals.
Comment by ak102 — October 5, 2007 #
Art,
it is possible Oodle have the technology and are not using it for some reason? (possible business/economic), or maybe i got it wrong in my original post. hopefully Duncan can expand in this forum or elsewhere
Comment by renthusiast — October 7, 2007 #
Yes, we do crawl, we also take feeds in all shapes and sizes from Xcel, CSV, XML etc. We have over 75,000 sources across the UK and USA and have spent over 2 years developing crawls specific to both each catetgory e.g. property, motors, to rent etc and geography e.g. USA, UK. At the end of the day its not how you get your listings is whether you can build a great consumer product on top
Comment by Anonymous — October 7, 2007 #