Content Matters: Vertical Search

October 24, 2006

Google Launches Personal Search

Google has thrown its (very large) hat into the personal search engine space with the launch of Google Co-Op Custom Search Engines.
The Custom Search Engines (CSE) are similar to those provided by Rollyo, Yahoo Search Builder and Eurekster. You create a CSE by entering a series of URLs that you wish Google to index. Google generates code for you to place on your website or blog and, voila, instant vertical search.

Unlike Rollyo, which limits you to 25 sites, Google indicates that there will be no limit to the number of sites you can include in a CSE. And while the search engine results will include Google Adsense ads, you can personalize the presentation of the results page. And, of course, you can't launch a product today without a social software spin, so Google allows you to enable users to contribute sites to your CSE. You also have the choice of restricting the CSE to only those sites you entered, or to include all web results while giving preferred weightings to those sites you have selected.

Google also includes the Google Marker function, a toolbar plugin which allows you to add content to your CSE as easily as you might tag a page in delicious.

As a quick test, I created a content industry CSE in about 5 minutes, including content from various blogs and websites covering the content and technology space. ~~I'll integrate it into the blog later, but for now~~ Update: I've added the CSE to the blog, just above the Rollyo box, so you can test both out and see which results you prefer or you can test it out here.

For more on Google Co-op Custom Search Engines, take a look at posts by John Battelle, Matt Cutts, Steve Rubel and Michael Arrington.

Barry Graubart on October 24, 2006 in Search, Technology, Vertical Search, Web 2.0 | Permalink | Comments (3) | TrackBack (0)

Technorati Tags: Google Co-op Custom Search Engine, personal search, Rollyo, vertical search

October 12, 2006

InfoCommerce Wrap Up

The second day of InfoCommerce seemed more lively and more relevant (at least to me).

Andy Goodenough led with the keynote, talking about how they'd built Highline Media to a $50M business in just over three years through a combination of acquisition and organic growth. It was interesting to note that Highline's data division, largely a byproduct of its editorial focus, will exceed $10M in revenues in 2006.

My vertical search panel followed, with presentations from Shannon Holman of ALM, Han Huang of FindGuru and DeSilva & Phillips Partner Jeff Dearth. We covered a lot of ground in an hour, showing examples of successful vertical search as a core (ad supported) product, as a value-add to subscription products and as a means of driving premium content sales.

One of the most interesting panels of the Conference was the Inferential Data panel, with presentations from Planalytics, MedMeme and Kelley Blue Book. Scott Bernhardt, SVP of Sales & Marketing for Planalytics walked us through a fascinating presentation of their weather-based business intelligence applications. Basically, they overlay weather data on a set of proprietary models to generate forecasts for retailers. For example, we had a cold and wet September, which translates to an improved market for rainwear, sweaters, soup and cough medicine and a down market for lawn & garden products and barbecue charcoal. In other words, they projected a good month for Kohl's (who later reported a 16% boost year-over-year in same store sales) and a poor one for Home Depot (who are yet to report their September numbers). I'd not seen Planalytics before and was very impressed with their offering.

Renny Ponvert, who is stepping down as President of Hemscott following its acquisition by i-Deal, led an interesting presentation on "Who's Winning and Why". Renny, a veteran of the financial information business, got everyone's attention by describing the defensive position many publishers take: "Insecurity, masked by cannibalization concerns creates a low-growth, slow death spiral", exemplified by the newspaper industry. On the positive side, Renny identified three tactics that content provider can use to their benefit: Visualization (examples were NY Times bubble charts and CNET Heat Maps), the use of crawl technology to scale your content business and the utilization of vertical search to reduce noise and improve the depth of your site.

InfoCommerce is a somewhat quirky conference. Of the content industry events, it probably has the largest percentage of old media including a handful of ostrich-like directory publishers. That said, there are always some interesting models from companies that otherwise don't show up on the radar. This year, for me, those included Planalytics, Healthline Networks and Health Market Science. Russell Perkins, who runs the event in conjunction with Roxanne Christensen and Megan St. John, has successfully transitioned the focus from traditional directories to now include vertical search, the inferential data companies mentioned above, b2b lead generation and other data-driven content businesses.

On Wednesday evening, InfoCommerce held a dinner where they presented their "Model of Excellence Awards" to about eight products. Alacra was honored to receive an award for its Alacra Compliance application.

If you haven't been to an InfoCommerce event recently, I'd recommend you put it on your calendar for next fall.

Barry Graubart on October 12, 2006 in Content Business, Events, Vertical Search | Permalink | Comments (0) | TrackBack (0)

Technorati Tags: Alacra, Healthline, Highline, InfoCommerce, Planalytics, vertical search

October 04, 2006

Vertical Search at InfoCommerce

One of the hot buzzwords of 2006 has been vertical search.

I've been asked moderate the Vertical Horizons panel at the upcoming InfoCommerce Conference, October 10-12 in Philadelphia. The panelists will include:

Shannon Holman, Director, Content Strategy & Development for American Lawyer Media
Jeffrey Dearth, Partner, DeSilva & Phillips
Han Huang, Director, Product Management for FindGuru (formerly ReedLink), Advantage Business Media

The focus of the panel will be on various business models for vertical search and why content companies are particularly well-suited to deliver vertical search. We'll also explore what it takes to deliver a compelling vertical search solution.

I'd appreciate your and ideas on vertical search. What issues would you like to see the panel address? What markets do you think are currently underserved by search? Where do you think the opportunities might lie?

I hope to see you in Philadelphia.

Barry Graubart on October 04, 2006 in Content Business, Search, Vertical Search | Permalink | Comments (2) | TrackBack (0)

Technorati Tags: ALM, content, DeSilva & Phillips, FindGuru, InfoCommerce, vertical search

October 03, 2006

Alacra Launches Google-Powered Vertical Search Engine

Alacra today announced the deployment of Google Search Appliances as the underlying technology powering the Alacra Compliance Web.

The Alacra Compliance Web is a vertical search engine used by financial institutions to accelerate the account vetting process. The Alacra Compliance Web is an integral part of Alacra Compliance, which helps financial institutions employ a consistent and documented process for KYC and Enhanced Due Diligence efforts to meet Patriot Act, BSA and FSA requirements.

The Compliance Web is an index of more than 500 global regulatory organizations, hand-selected and classified by Alacra Content Analysts. Sites include stock exchanges, government regulators, central banks, state insurance sites and more. Using Google Search Appliances lets Alacra leverage Google’s strengths in quality and ranking and multilingual capabilities.

One unique feature of the vertical search engine is that it highlights in yellow any "adverse" terms, such as convicted, indict or subpoena. These highlights help the user quickly identify any critical information among the search results.

Alacra has released, on a limited time basis, a searchable free trial version of the Alacra Compliance Web. You can take it for a test drive here.

Typical questions that an Alacra Compliance Web Search might answer:
* Is Neoware listed on a U.S. stock exchange?
* Is Segoes Securities allowed to conduct business in the Cayman Islands?
* Has GLG Partners been the subject of any regulatory disciplinary actions?

For more on the Alacra Compliance Web, take a look at the FAQ.

Barry Graubart on October 03, 2006 in Alacra, Content Business, Search, Vertical Search | Permalink | Comments (0) | TrackBack (0)

Technorati Tags: Alacra, Alacra Compliance, Google, KYC, Vertical Search

September 12, 2006

Alacra Compliance Wins Model of Excellence Award

The InfoCommerce Group has chosen Alacra Compliance as winner of its Model of Excellence Award. The annual Model of Excellence Awards recognize innovative companies which effectively combine world-class software with high quality data.

Alacra Compliance was selected a winner in the “Mixed Models: Process Integration” category, which focuses on workflow-based applications which leverage content combined with software.

According to InfoCommerce CEO Russell Perkins, “Alacra Compliance is an innovative workflow application that helps financial institutions get their arms around rapidly changing regulatory requirements”.

Since its launch earlier this year, Alacra Compliance has been well-received by AML and Compliance officers, seeking ways to streamline their processes to comply with Anti-Money Laundering, Know-Your-Customer and Enhanced Due Diligence requirements. Alacra Compliance enables federated searching across premium content, public records web sites and the Alacra Compliance Web, a vertical search engine, delivering a single set of results geared specifically towards research needed for Patriot Act, BSA and FSA compliance.

The Model of Excellence Award will be presented to Alacra at the annual InfoCommerce Conference, to be held October 10-12 in Philadelphia.

The Model of Excellence Profile provides greater details on the award and the underlying product.

Barry Graubart on September 12, 2006 in Alacra, Content Business, Vertical Search | Permalink | Comments (0) | TrackBack (0)

Technorati Tags: Alacra, AML, content, InfoCommerce, KYC, Model of Excellence Award

June 20, 2006

Craigslist Adds More Cities, International Coverage

According to the Wall Street Journal, Craigslist has just added 100 new cities, including 20 outside the United States, bringing its total coverage to 300 cities.

According to Staci at PaidContent, the expansion includes 72 US cities plus 28 international cities from Beirut to Helsinki.

In the article, WSJ editor Brian Carney interviews Craigslist CEO Jim Buckmaster over lunch and bloody mary's. They hit on most of the standard topics (why do you leave a half billion dollars on the table?), but provide some interesting insights to Buckmaster's thinking.

Carney quotes Buckmaster "I do think that the Internet is a spectacular tool for any information business -- newsgathering and other journalistic enterprises are essentially in the information business. Another aspect to it that gets reported on is drawing the lines within the Internet itself with respect to content generators and various kinds of aggregation and search tools.

Where does the revenue end up in those kinds of scenarios over time? I think you'll see the lines will move from side to side in terms of where the revenue lands among the various players in the information economy, which is still very young."

Craigslist continues to be a very interesting experiment in how a content technology company can deliver fantastic value to its users while taking a low cost, minimalist approach to its entire business. While those outside the classified business may think Craigslist is not their problem, we can expect others to try to emulate Craigslist and freemium models in all aspects of the content space.

Barry Graubart on June 20, 2006 in Content Business, Vertical Search | Permalink | Comments (0) | TrackBack (0)

June 16, 2006

Google launches vertical search for Government market

Google today announced the (re)launch of Google U.S. Government Search, an updated version of their vertical search of federal government websites.

The Google Government search has actually been around for at least five years, though it was not heavily promoted.

The updated site indexes content from federal, state and local government sites - basically it's the subset of the Google index from .gov or .mil domains, along with some hand-picked sites from .com, .us and .edu domains.

Users can submit searches as they would on Google, but results will be targeted to those from government sites. For example, looking for information on John Wood, who serves as Chief of Staff at the Department of Homeland Security, I first used the standard Google search. It returned 289 million pages, led by John Wood Community College and John Wood Water Heaters. The DHS Chief of Staff was nowhere to be found among the first few results pages. The same search on Google Government yielded 3.5 million pages, with the DHS John Wood right at the top.

Google's relaunch of this site was timed to tout the hiring of Mike Bradshaw as head of federal sales. Clearly, Google is looking to sell its Google appliance more heavily in the federal markets and hopes that as users find Google useful for searching their extranets, they will want to use Google for their internal file access as well.

For publishers serving vertical markets, this should serve as a reminder of how compelling vertical search can be. Leveraging their domain expertise and the skills of database editors, content providers are uniquely situated to create useful vertical search solutions to drive traffic to their sites for advertising or to drive transactions.

While Google's U.S. Government Search helps improve search results, a publisher focused on the government space could take it even further, providing search by level (Federal, State or local), department or agency or even individual states or localities.

Update: An interesting comment from Outsell's Chuck Richard, who notes that "the SERPs are already populated with AdWords ads from information companies such as BidNet, Onvia, GovernmentBids, BidLink, and many more. This drives home once again that what on the surface looks competitive to information companies in the government space is being used by them to steer traffic back to their professional-quality, finely matched content." While these AdWords bids were purchased for Google (and not specifically for Google U.S. Government Search), it's clear that they will benefit from the positioning.

Barry Graubart on June 16, 2006 in Content Business, Search, Vertical Search | Permalink | Comments (0) | TrackBack (0)

June 05, 2006

Topix.net enters free classifieds market

Topix.net appears to be offering free local classifieds, in the mode of Craigslist.

When browsing a local city section on Topix.net, a new column of classified ads is displayed in the right column, along with a link to "post your own free ad". The classifieds can be posted into categories including housing, jobs, for sale, services, events and local shops. For example, the topix page for San Francisco includes 2 jobs, 1 for sale and 5 services ads. You can see from the two-level deep taxonomy the types of ads they are seeking.

The service seems to have been rolled out very quietly in early May. As such, there are just a handful of listings so far (8 for San Francisco, 11 for San Jose and 22 for NYC). It seems that the Topix team may be test-marketing the concept before an official launch.

As Steve Rubel points out, it's interesting that Topix, a joint venture between the three large newspaper chains, Gannett, Tribune and Knight Ridder, would go into the free classifieds business. As I've posted previously, Craigslist is putting huge pressure on the traditional classifieds business. Perhaps these three are willing to try to cannibalize their own classified business rather than allowing Craigslist to do it to them. If so, kudos to them for the effort, although they'll have to put the weight of their local brands behind the effort to close the huge gap with Craigslist.

One interesting differentiator is that Topix is bringing the power of social software to these classifieds. They provide the ability for users to add comments to the classifieds, or to flag an ad as being offensive or inappropriate.

Topix.net and Craigslist are both on the 50 Content Companies that Matter list. It will be interesting to see how these two innovators position themselves in the classified advertising market.

Barry Graubart on June 05, 2006 in Content Business, Vertical Search | Permalink | Comments (0) | TrackBack (0)

May 11, 2006

Google Co-op - Update

Danny Sullivan has an in-depth overview and analysis of Google Co-op at his Search Engine Watch blog.

Danny describes it as Rollyo-like vertical search, combined with tagging (or, as Google calls it, labels). Users can "label" pages into various topics. Unlike end-user tagging with delicious, Google labels are geared towards publishers (or other vertical market developers) wishing to tag or categorize large numbers of pages into a predefined taxonomy. This process will be similar to what content providers have done for years using manual tagging or automated categorization software. The taxonomies can be hierarchical, with labels and sublabels.

It will be interesting to see how this pans out. One reason that categorization has succeeded for large content providers but has often failed inside the enterprise is that it takes a fairly rigorous editorial process to apply consistent tags to content. Google has always emphasized the machine's ability to make information navigable through algorithms. This approach, bringing editorial decision making into the process, is sure to lead to some compelling discussion.

Barry Graubart on May 11, 2006 in Content Business, Search, Vertical Search | Permalink | Comments (0) | TrackBack (0)

The 50 Content Companies that Matter: Topix

There are many news aggregators in the market today. Most of them add value to the news content by tagging and organizing it for their markets. While most have adopted some level of automation, they still rely heavily on editorial expertise to ensure accuracy.

Topix.net is a news aggregator which relies solely on automated indexing to create a comprehensive taxonomy to drive relevancy. Founded in 2002 with a mission of helping users quickly and easily find targeted news on the Internet, Topix.net today categorizes news from more than 10,000 sources into over 300,000 specific topics in near real-time.

The founders at Topix.net came out of the Netscape-driven Open Directory Project, an editorially-driven project to categorize sites on the Internet. According to co-founder and CEO Rich Skrenta, their experience with tends of thousands of volunteer editors made him “decide to avoid editors in Topix.net and stick to algorithms”.

Early on, one of their key issues was resolving how to serve relevant contextual advertising on a news site. While Google AdSense typically serves relevant advertising to static web pages, the results for news were less relevant and often embarrassing. In one of the more notorious examples, the New York Post published an article about a murder where the victim’s body parts were found in a suitcase. Adjacent to the article was an AdSense ad touting luggage (for those "do-it yourself-ers", I guess).

To combat that, the technology team at Topix.net developed a set of algorithms and a knowledgebase that could better understand what a news story was about and serve up more relevant ads. The knowledgebase is focused on key topical areas, such as geographical map data, lists of movies, CDs and entertainers, and more. The Topix.net algorithms leverage the information in the knowledgebase to classify each news article to the correct topic.

Today, Topix.net gets more than 3 million unique visitors per month. They estimate that to be about half of their readership, as 50% comes through the sites of partners AOL, Yahoo, CitySearch, Ask.com and others. Topix revenue is largely driven through serving AdSense ads on their pages.

Last year, 75% of Topix.net was purchased by Gannett, Tribune and Knight-Ridder (now McClatchy). That investment assigned a healthy valuation of $64M to Topix.net; not bad for a company with estimated revenues of $1M.

Just last week, Topix.net announced the launch of its publisher platform, a revenue-share model targeted towards small-to-medium news providers. They are also extending the invitation to bloggers a la Pluck's Blogburst.

Topix.net has done an amazing job of leveraging technology and domain expertise to create a fully automated system for categorizing news and serving relevant ads. Publishers often view editorial’s role in categorization as simply checking or verifying automated results. Topix.net has demonstrated that a more effective use of this domain knowledge is to use it in the development of a knowledge base and set of algorithms that can automate the labor-intensive tasks. And for that, they are clearly one of the 50 Content Companies that Matter.

Barry Graubart on May 11, 2006 in 50 Content Companies That Matter, Content Business, Search, Vertical Search | Permalink | Comments (1) | TrackBack (0)

Content Matters

Occasional ruminations on the convergence of content and technology.

About Content Matters

About

Email me

50 Content Companies that Matter