Cazoodle Blog

Shoppers no longer need to visit so many websites to find the electronics of their dreams. Cazoodle Shopping Search is the first and only organic shopping search service that simultaneously provides comprehensive as well as precise product information. The first launch is for consumer electronic products including laptop and desktop computer, digital camera and camcorder, printer, television, mp3 player, etc.

While many shopping search services exist already, Cazoodle Shopping Search provides far better results. Most existing services, including Shopping.com, Price Grabber, Nextag, and Google Products, require merchants to submit data feed and are, thus, inherently limited in their coverage. For “canon sd1200” query, Cazoodle returns 610 offers–compared to 33 offers at Shopping.com. The offers for the same product are organized into one group for easy comparison.

If you are at DEMO, please stop by our station #43 in pavillion. Or, watch live video stream of our on-stage launch (Today, at 3:00 pm PT).

To read more about the launch, please see our press kit.

• Download the full press kit zip file. The folder contains
  • Press Release and Fact Sheet (pdf file).
  • Company Overview (pdf file).
  • Screenshots
    • Comprehensive Organic Search (jpg file).
    • Precise Result Filtering (jpg file).
    • Offers, Reviews, PriceHistory, Images, and Specs (jpg file).
    • Shopping Caddy - Firefox Plugin (jpg file).
    • Shopping Caddy - Save Money (jpg file).

Recently, I wrote about our dramatic increase in user traffic. We are now here at National Apartment Association Education Expo (booth #622), to showcase to property managers from all over United States on how they can maximize their online exposure using our service.

To begin with, let me recap why renters find our service unique and so useful. Our service is the first and only comprehensive organic search for apartments crawled organically from everywhere on the Web. The product is available for all locations in United States. For example, for “New York, NY” our website has over 100, 000 apartment choices. For “San Francisco, CA“, nearly 10, 000 choices. So if your property is listed anywhere on the Web, quite possibly our search results already show it.

Property managers all over United States have enjoyed benefits of our search engine. The snippet of each result on our site includes the hyper-link to the original source, and a phone number to reach property managers—both of which generate many new leads for property managers.

(Method 1) If your properties are not present on our search result, please let us know. We will be quick in adding your website to our crawlers (at no fee), so that your listings get exposure to our increasing user base.

(Method 2) We also provide functions for uploading properties in bulk using XML data feeds (again, at no fee). Some property managers who want better control of their listings find this quite useful—you can simply create XML data feeds hosted on your website, and our crawlers will update your listings every night. You can easily remove/add properties on your feed, and precisely control different specifications, including bedroom, bathroom, rent, images, availability, etc.

(Method 3) Finally, but most effective of all, property managers can also use our Featured Listings Program (with performance based fee). We reserve a section above organic search results for paid featured listings. Since your listings will be on top of search results, you will gain more exposure—which will translate into more leads. You will be charged based on the number of users who click on your advertised listing.

Good News! Featured Listings Program is being offered for free, too.

Free Trial Period: Unlimited Advertising, at No Charge, Until July 31st 2009.

To learn more: Stop by our booth here at NAA (Booth #622), or contact us by phone or email.

Govind Kabra

We haven’t posted any blog updates for quite a long time. But we have a good excuse. Since we presented Apartment Search at Web 2.0 Expo 2 months ago, our traffic is skyrocketing… and that kept us busy.

Our traffic rise.

In just 2 months, our daily traffic has increased by factor of over 30 times (based on Google Analytics and our own logs). For now, let us just focus on publicly available numbers from compete. For May it is up at 52K uniques, up from 15K in April and 9K in March—a growth of 350% in last 2 months. Our daily traffic (not captured in this graph here) has increased 30x in these 2 months.

The pain in searching for apartments.

Users today have to look through site after site after site. A recent study from apartments.com shows that over 90% of renters use multiple online sites for their apartment hunting. And they have a good reason for that—because we did not have a “Google” like comprehensive search engine for apartments.

We have experienced this pain first hand. As graduate students at UIUC, we were always looking for apartments. And boy, it was so frustrating. We were looking at big posting sites, newspaper classifieds, online user forums, and then, of course, the long tail of landlord sites. Each site gave only a few apartment listings. We could never get big picture of all the apartment choices around campus.

So when we started Cazoodle, we built our first product—Cazoodle Apartment Search— to solve our own pain. The service provides comprehensive organic search for apartments crawled organically from everywhere on the Web. The product is available for all locations in United States. For example, for “New York, NY” our website has over 100, 000 apartment choices. For “San Francisco, CA“, nearly 10, 000 choices.

Clearly, our recent traffic growth shows that users are badly in need of an organic search engine for apartments.

Some more interesting tidbits on our user demographics, based on Alexa.

Age-wise: Service is not so popular in 18-24, more popular in 25 - 34 (when they have job), less popular in 35 - 44 (when they own home), and again popular in 45 - 54 (when they retire and move to new location).

Service is more popular in male users (house-hunting seems to be job of males, I guess), and in families with no children (makes sense—people like to own house when raising family).

Finally, but most surprisingly, people use our service more at work than at home.  We also noticed in our logs—daily traffic on weekends is 20% lower than on weekdays. What does this tell us? No, I am not talking work ethics (we all know about that). What I am thinking is that employers need to recognize how difficult it is for their folks to find apartment. How about customized Apartment Search, say, “Cazoodle for IBM,” or “Cazoodle for Microsoft?” Sure, we can do that. Call me.

The emergence of specialized vertical search system would clearly be the next big revolution in search industry (dubbed as Web 3.0 era). A key barrier in realizing this vision is the technology to understand the semantics of semistructured Web-data. In this blog post, I would explain how we, at Cazoodle, addresses this challenge.

Before I explain our method, let me stress on two key requirements for integration at Web-scale: a) Scalability, and b) Accuracy. Several (failed) attempts are based on writing software programs (or wrappers) which would essentially reverse-engineer the HTML/javascript code. Clearly, this is super expensive, even if you offshore to India (upwards of $100 per website per year). On the other hand, automatic techniques (even after encoding domain knowledge) are not so accurate. It is noteworthy that some semi-automatic tools are available; however, they are not as adequate either. For example, Kapow RoboSuite, Connotate AgentSuite, and Fetch (primarily targeting enterprise applications) or Dapper (end-user web-based tool). The models they use for their “agents”/”robots” are either not expressive enough to match requirements of many sites (necessary for Web-scale) or not robust enough to changes in sources over time (necessary for accuracy).

Our solution is AgentIDE, an interactive agent development tool, developed based on our previous research work at University of Illinois Urbana Champaign. First, based on domain encoding, an initial source model is automatically generated. Then, our agent builders (human operators) can debug the agent and make necessary modifications. The key strength lies in the design of the IDE, and the learning approach—agent builders do not need any programming skills, and can be from diverse backgrounds such as students from music, psychology, finance, etc.

We model sources using a collection of expressive and persistent features that reflects how the sources are presented to the end users. Expressive features allow us to describe all kinds of source data precisely. And persistent features tend to be robust even when the source sites change in their structure. Furthermore, the set of features used are not fixed. Rather, these are statistically analyzed over both the source and the time dimensions, to identify what model is more accurate as well as robust to source changes. This allows us to control the cost of maintenence significantly.

With its unique combination of scalability as well as accuracy, our agent technology can enable truly Web-scale vertical search systems. Try our first product, Cazoodle Apartment Search, to judge for yourself—the system is available nationwide, and can find at-least 20-30 times more rental choices than the currently popular services. We plan to expand to new vertical domains soon, including online consumer electronics shopping, and finding local events. If you wish to use our technology to build similar services in other domains, just let us know, and we would be happy to collaborate.

Quoc Le
Lead Developer

Last week, we were at the Web 2.0 Expo to launch our Apartment Search product. The product is now available for all locations in United States, providing atleast a factor of 20-30 more choices than the currently popular services. The video on the right-side gives a good demo of our product, thanks to Charnoff Len, who created this video for his blog on time saving tutorials.

Although the economy is hard-hit, atmosphere at the Expo was full of excitement. All (three) of us were talking non-stop to the visitors on our booth.

Visitors generally gave rave comments… not only liked our Apartment Search product, but also eager to see Cazoodle advancing to new markets. Why shopping search limited to laptops? When will Cazoodle do job search? Used Cars? How about iPhone app for Apartment Search? Yes, we wish to do all. And also, looking for partners who would like to co-develop similar services in new domains.

Bloggers dubbed us as “Semantic Search,” “data-aware vertical search,” “intelligent vertical search” and the cutest of all—”nifty little search engine that could develop into something powerful.” Some asked about data quality, specially broken links. Data quality is our top-most priority. However, broken links are an issue for any search engine. We are confident that we can address this concern by maintaining our crawl more up-to-date.

A common question was—”How do you do it?” The secret sauce is in our technology for enabling Web-scale data extraction and integration. I guess, given the interest, although information is proprietary, some disclosure is in order. Watch out for our next blog.

We even got some Television coverage—Brian Shields from Kron 4, a local TV channel popular in San Francisco interviewed us in the following video.

… well, not launching; Just attending! To admire the other companies launching.

Although attendance was reduced to half (only 39 launch compared to 60-70 in past), the quality of companies was pretty good. Interestingly, no two companies had overlap in their product offerings. Don’t miss the video recordings on demo website of Smarty Card, Kutano, 7 Billion People. Skout was hilarious. Sarah Perez at ReadWriteWeb has good coverage of Semantic Search space (1/2, 2/2), most relevant to Cazoodle.

I found Demo quite unique in its style. Each person in the audience could be tagged into 1 of the 3 demographics: (1) a VC, (2) a journalist, or (3) a start-up in need of 1 and 2.

At Cazoodle, we were interested in 2 key questions:

• (Q1) Did the launching companies find Demo worth it?
• (Q2) How could Cazoodle prepare for such a major launch?

For Q1, I talked to several founders. The answer seemed to be mostly Yes, for two reasons:

1. The high investment (Demo costs 18K$) implied they wanted to “complete” their product before launch. This immediately meant they got more work done in same time frame than they thought they were capable of.
2. They all got significant press coverage, and potential VC leads… which is consistent with our observation above about Demo audience demographics.

For Q2, I pitched Cazoodle to all the 3 types of attendees—VCs, Journalists, and other companies, and got many comments.

1. Some encouraging… Greame Thickins interviewed me for “Cazoodle: Future Demo presenter?”. Another journalist: “That’s true semantic Web then?”. A start-up: “We would want to use your data extraction”
2. Others discouraging… VC1: “Anand already did this long time back at Junglee”
3. By far, most useful advice… “Define your focus. Enterprise vs. End User is most startup’s question. Cant do both. Work on one thing, perfect it. If one fails, then only move to another thing. Ideas come all the time—reject even the good ones.. if they are not your focus.”

Why don’t you also give us your comments? Here is a short 5 min video. Let us know what you think.

Govind Kabra

Kelsey Group, a leading provider of local and vertical market research, puts Cazoodle on its Entrepreneur watch list for our innovative products and technology.

Entrepreneur Watch: Cazoodle Crawls Vertical Listings

Cazoodle, a new listings-based service, has launched from the incubator at The University of Illinois at Champaign-Urbana (yes, Marc Andreeson’s former territory). The site currently crawls for apartment listings and shopping. Additional vertical categories such as events are anticipated, notes Professor Kevin Chang….

Kelsey Group is also host of premier industry conferences including ILM & Marketplaces. Are you planning to attend? We sure are! We will be in LA for Marketplaces 2009 next month. Let us know and we would love to reach out and show you our stuff.

Govind Kabra

In this first post of my 2 post series on How to Search for Apartments, I will focus on how to do so using our own—Cazoodle Apartment Search. The service, now available for 18 metropolitan areas, has seen steady growth in user base… many of these users write to us about their experiences, amazements, and frustrations. These user feedbacks have inspired us to draw concrete roadmap for our product development.

Today, we are proud to announce a major milestone, codenamed “Local Release”. We believe apartment finding is a process, comprising of several steps beyond just browsing through candidate apartments. And so in this release, we aimed to localize several of these steps right into our search console. The key-highlights of the release are:

• View Surroundings Beyond visualizing the location of apartment on map, now you can also inspect its surroundings using google street view. You can also evaluate its proximity to other services using map layover function to overlay near-by restaurants, bars, schools, parks, etc.
• Draw On Map If you have set preference on where to live, you can now use Draw On Map function to restrict search results to the (rectangular) region you interactively draw on map.
• Save Favorites We understand apartment hunting requires many search queries. To let you track and revisit the apartments you like, across multiple search queries, we allow you to add them to favorites.
• Share Apartment Less than a quarter of the apartments are 1 bedroom, meaning, more often than not, apartment searching is a collaborative effort between you, your friends, family, and potential room-mates. Now you can share apartments by clicking on the email icon.
• Apartment Guide Sure, you have a long list of questions. Which neighborhoods in New York are safer to live in? Your tenant rights in Miami? How to set up gas and electricity in Chicago? Places to visit after you moved into San Jose? To answer these, and many more, we created location specific apartment guides.

Beyond this localization of various steps into a single console, as our next milestone, the “Analytics” release (yes, a codename ) will provide various statistical information to facilitate business intelligence, and end users, alike. Further next, our “community” release, would provide functions for easier collaboration between various participants in our eco-system, including users, landlords, and advertisers.

Stay Cazoodled… a lot more to come!

Govind Kabra

Cazoodle currently has four offices in different geographic locations. The trade-off of hiring good people in different places is the communication overhead across timezones. As a result, collaborating efficiently becomes more important for a startup like us. Thanks to the Internet age and the open source world, there are many tools available for various purpose of collaboration.

We use TWiki(http://twiki.org/) for general documentation, meeting summaries, and daily journals. TWiki is a “structured wiki”(http://en.wikipedia.org/wiki/Structured_wiki), which supports numerous structured customizations and programming to support daily business functions. The E-mail notification feature keeps everyone in the same page. And, the regular expression-enabled search feature allows us to find documents efficiently. The rich library of TWiki plugins also makes life easier for finding various additional functions. However, at the current point we do not find any calendar plug-in that fits our need, so we use VCalendar(http://www.ultraapps.com/app_overview.php?app_id=19) for the company-wide calendar. 

Like many open source projects and companies, we use Trac(http://trac.edgewall.org/) for software project management. The project-specific documentation is also kept in the Trac wiki. The features of Trac almost fits everything we need except the code review, which we plan to use Review Board(http://www.review-board.org/) in the future. We also implemented tools to generate reports from the Trac database. If you are going to run a software project, we highly recommend you to try Trac.

For forums, phpBB(http://www.phpbb.com/) just works. We use it as the company forum for discussing all sorts of issues– in a sense each forum thread is an asynchronous meeting, where all of us at different timezones can talk when they are awake. WordPress(http://wordpress.org/), as you see here, is the tool we used for blogging. And the micro-blogging service provided by Yammer(http://www.yammer.com/) makes people more likely working in the same location.

For real-time communication, Skype(http://www.skype.com/) or its alternatives(http://www.voip-news.com/feature/10-sky … es-091707/) is your friend. TightVNC(http://www.tightvnc.com/) is the right tool for screen sharing and we plan to try Gobby(http://gobby.0×539.de/trac/) for collaborate editing.

What’s your favorite collaboration tool? Share with us.

York, Developer team

When we are browsing the web, we are probably looking for something or some information, either for fun, school, or work. We might google first and get to the right page if we are lucky, or we have to google several times and go to different pages to get the whole picture. Wouldn’t it be nice if the browser knows what we might need and helps us put together as much stuff as possible, and saves us some time of clicking?

Nowadays, browsers are getting smarter with more built-in features, or you can install a lot of third-party plugins or extensions. The most common example may be the automatic dictionary or glossary lookup. Like MashLogic, whenever you mouse over to a term, a popup window will try to tell you what it’s about by gathering information from many other sites. BlueOrganizer can add a smart link automatically next to the link to a book or music you might be interested in, and tell you its popularity, where to get it, or even the reviews from your friends. Deckkr lets you stay at the same page and still be able to navigate other sites. There are still many other nice and cool tools which make browsers more fun and smart, but what exactly should be a killer application for smart browsing? This is up for debate and should be determined according to different purposes.

What can we expect more from a browser? Most search engines can provide us a page with multiple results. A better vertical search engine can give us more and better organized results from different sources. Still, the browser has no idea about the differences across pages and domains. It’s only a window to the WWW. To make the browser smart, data awareness is the key component. It has to be able to digest the content and recognize special objects and patterns. Taking apartment search as an example, when a user is browsing a page with apartment listings (say, a landlord site), the browser can show the user other “similar” apartments in the proximity of those on the current page. Or, with the specific knowledge about apartment domain, such as rent, bedroom, location, etc, it can store the results we are interested in a better way, not only via traditional URL bookmarking, but rather like a shopping cart that knows about the structure of the data objects being remembered, even with comparison and analysis. With such “data awareness”, the browser can become a vertical search engine for apartments (or any data of interest).

Smart browsing can be applied to all kinds of domains, not just the apartment search. Using the same framework and backend system, with a little modification, we should be able to deploy it to other areas, like regular shopping activities. Shopping for an apartment can be easily replaced as shopping for a book, a car or some services. The main difference is still about the domain-specific knowledge. Apartment domain is our current focus, and there are still rooms for improvement. Enabling smart browsing is one of the areas we are trying to achieve. We want to make the apartment search as easy as just one click away, and eventually, make search a totally different experience!

We see the possibilities of immersing search with browsing, although it might still be unclear how smart browsing will materialize itself. Everyone might have different ideas on how to make browsers smart, and what the most wanted features should be. As we are still shaping our products towards this direction– People, what do you think? What’s your desire?

Yuping Tseng.