SES New York 2010, March 22-26
Subscribe to SearchDay, our free daily e-mail summarizing the day's Search Marketing News.
Recent Comments

Back to Main

May 23, 2005


Could a New Algorithm Increase Profits for Search Ad Providers?

Phys.org has posted a news release from Georgia Tech University in Atlanta about a new computer algorithm, "that could further increase profits for search engine advertising."

From the news release:

Upon examining the Google ad model, Vijay Vazirani, together with his two Georgia Tech Ph.D. students, Aranyak Mehta and Amin Saberi, and Umesh Vazirani, a professor of computer science at Berkeley, realized that always giving the top spot to the highest bidder is not the best strategy for Google. The top bidders might rapidly exhaust their budgets and get dropped from the auction, thus reducing the competition for that keyword.

Google?s profits will be higher, the researchers reasoned, if it somehow weighs both bids and remaining budget when ranking ads. They found a mathematical formula that finds the optimal trade-off between bids and remaining budget, maximizing what the advertisers are spending.

The research team has filed a provisional patent for their work to ensure that the research remains in the public domain. The team continues to explore other applications for their algorithm such as in engineering and for solving other computer science problems.

Btw, the research itself was motivated by a question asked of one of the researchers by Google's former Director of Research, Monika Henzinger, while he was interviewing for a job at the Googleplex last year.

If you want to learn more about the research, here's the full text of the paper: "Adwords and Generalized Online Matching'' NOTE: The paper is in Postscript format. You can use this site to convert to PDF. Also, this excellent overview article: Computer Scientists Optimize Innovative Ad Auction (PDF), is more than worthy of your attention.

The researchers have not yet published their work, but Saberi, now at Stanford, gave a talk on the algorithm that some Google researchers attended. He was subsequently invited to give another talk on the result at Google. [Monika] Henzinger, who originally posed the problem to Saberi, has since taken an academic post at Ecole Polytechnique Fédérale de Lausanne but continues to consult for Google. She requested a video of Saberi?s talk.

Posted by Gary Price on May 23, 2005, 6:46 PM | Permalink


Yahoo Announces a Challenge for Webmasters, Bloggers, and Web Developers

This morning Yahoo launched the YQ Challenge to find the "most innovative" implementations of their YQ contextual search platform.

Yahoo launched YQ in beta earlier this year.

What criteria will Yahoo use to judge the contest?

* Relevance: How effectively have you set context so that YQ produces results with maximum relevance?
* Prominence: How prominent is the placement of YQ on your site?
* Placement: How creative and intuitive are the placements of YQ icons on your site?
* Usefulness: To what degree does YQ enhance the overall user experience on your site?

The winner of the YQ Challenge receives $5000. Second place receives $2500. $1000 for third.

This FAQ has more info. The deadline for submitting entries is June 19th.

Posted by Gary Price on May 23, 2005, 3:30 PM | Permalink


Some Publishers Not Happy With Google's Library Digitization Program

When I first learned about Google's plan to digitize the full text holdings of several large libraries one of the first things that came to mind was the many copyright issues that would come from the publishing community. Well, here they come.

In March, the Harvard Crimson ran a story about copyright issues and the Google Library program

Today, two more articles about concerns coming from the publishing community over the program have been published.

The Chronicle of Higher Education and Business Week have articles about a recent letter sent by The Association of American University Presses (about 125 scholarly publishers) saying that the Google Library program, "appears to be built on a fundamental violation of the copyright act." You can find the full text of the letter here. AAUP is requesting that Google respond to the letter by June 20th.

What seems to be of most concern to the publishers is the millions of books that Google plans to scan from library collections which are still in copyright.

Let's review what Google Print/Google Library consists of.

+ Google is working with publishers to digitize new material that is shared directly from the publishers. Material will be full text searchable but each publisher determines how much of the book you can view during a visit.

Material will not be printable (yes, you could do screen caps for each page). All of this is very similar to what Amazon.com is doing with their Search Inside the Book program. All book entries will have direct links to online book merchants allowing users to purchase the full text.

+ Google Library Program
Launched in December, Google's library program plans to digitize every book, both copywritten and public domain material held in several major library collections, and make is searchable via Google Print.

Several of the libraries that have joined the program are testing the program before committing to a full digitization. Regardless, if all goes as planned, scanning this massive amount of material will take many years.

What I think is most noteworthy is that library materials that are not out of copyright will be full text searchable online but not full text viewable. The searcher will only see a few sentences of text around the search term along with bibliographic info and links to purchase the actual book. In some cases, you'll also see a link to access the book from a library. In this day and age, will people be willing to wait to get the book(s) via interlibrary loan (if a local library doesn't have it) and more importantly, can libraries afford a large increase in the number of inter-library requests? Of course, also be in Google's future plans to offer the downloadable full text of any book online.

If the library content is in the public domain then the full text will be viewable online. The dates that Google is using to determine public domain material vary. From the Google Library FAQ:

If you're in the US, we've taken a very conservative stance and only books pre-1923 will be considered public domain. If you're not in the US, only books pre-1900s can be considered public domain because of differing copyright laws internationally.

So, that's the gist of it.

The Association of University Publishers says that simply scanning the copywritten material might be a copyright violation.

"Copyright means the right to make copies, period," said Peter Givler, the university-press group's executive director, in an interview. "Copyright law can seem pretty byzantine and technical and elaborate and complicated," said Mr. Givler, who wrote the letter, "but at its simplest, that's what it is. It's the right to make copies..."It's just a gigantic claim on its surface," Mr. Givler said in the interview. "There are just a lot of questions that need to be answered."

Google has also heard from UK's, Publishers Association,
It sent a letter to Google in February that touches on many of the same points that the AAUP letter discusses. The Chronicle of Higher Education reports that, "Google's answers thus far have not been reassuring." Business Week notes that John Wiley & Sons and Random House have also contacted Google about the library program.

Finally, comments from Lawrence Lessig at Stanford,

"For registered works it can be up to $150,000 per infringement," says Lessig. "I don't think any judge would do that because Google seems to be operating in good faith...but there's a huge exposure."

As I've pointed out many times, Google Print is hardly the only service making full text available. This post links to several of these other services. I also included a few in my original article about Google Library.

Posted by Gary Price on May 23, 2005, 12:55 PM | Permalink


MSN Virtual Earth To Take On Google Earth

Microsoft sends news today that founder Bill Gates has announced a MSN Virtual Earth service is to debut in the summer. The service is promised to provide:

  • Satellite images with 45-degree-angle views of buildings and neighborhoods
     
  • Satellite images with street map overlays
     
  • Ability to add local data layers, such as showing local businesses or restaurants

The service will allow users to choose from a number of different data types plus allow people to contribute their own information. The announcement came today at the D3 conference happening this week.

Here are some screenshots from the service to come. You can click on the first to make it larger:

050523-msn1.jpg

 050523-msn2.jpg

Last week, Google announced that its Keyhole software allowing satellite views of the Earth will be renamed Google Earth. Better images will also be available to those getting satellite views via Google Maps. In addition, Google Maps has also been getting enhanced by others adding on data from other sources. More on this in is covered in our Google Factory Tour Recap post.

So, the planned Microsoft service will help MSN compete against some Google gains in this area, not to mention moves others like Yahoo have been making with maps. Some past reading you may find interesting on these topics:

Posted by Danny Sullivan on May 23, 2005, 12:52 PM | Permalink


Personalization for the Mobile Web Searcher

TheFeature's Peggy Ann Salz conducts an interview with search engine personalization expert Barry Smyth about using personalization to improve the online experience of the mobile searcher by using informal communities.

TheFeature: Some argue that strong brands like Google and Yahoo! may cost mobile operators portal traffic since users may simply go to the search engine to locate content. How can your technology put operators on more even footing?

Smyth: The technology doesn't have to reside in the search engine, so operators can provide it as an additional layer on top of the search engine. This approach allows operators to capture important information from the user communities and deliver better search results. But it could also go beyond better search to a whole list of value-adds.

You can learn more about Smyth's work on his home page and from Changing Worlds, a company that's developing personalization technology. Barry Smyth is CTO. Thanks to G.L. for the links.

Posted by Gary Price on May 23, 2005, 12:00 PM | Permalink


ODP Forum Closes Status Check Service; Time For The ODP To Close?

Been wondering what's going on with your Open Directory submission? The Resource Zone site has long offered a way to ask ODP editors to do a site status check, a great safety valve for frustrated site owners. But now there's bad news. Via Search Engine Roundtable, news that this service has now been closed: Discontinuation of site status checks. In short, providing this service was deemed taking too much time away from the forum's main mission of discussing the ODP.

Here's a thought. Maybe it's time to dump the ODP editing idea all together, in terms of what's allowed in. Wikipedia seems to get by with letting anyone contribute to its pages. How about letting anyone add any link to any ODP category, while the editors serve as Wikipedia editors seem to do, as high-end police force to keep spam out.

I know, I know -- a flood of spam would probably make that fail. OK, maybe the ODP could be restructured where anyone could tag pages based on the ODP categories, with the editors service as ontology police. That's been one weakness of tagging. A community-upward approach leads to no consistency in tags.

OK, OK -- more spam will still probably come of that. So here's a third idea. Maybe it's just time for the ODP to close, as one of its own founders suggested earlier this year.

Back in 1998, people were frustrated by the inability to get listed in Yahoo. And if you weren't in Yahoo back then, you are largely invisible to web searchers. NewHoo -- later renamed The Open Directory -- was born to provide a solution to the invisibility problem.

It was a good idea then. Today, crawlers rather than directories dominate the web. Google dropped its version of the Open Directory from its home page long ago, and Yahoo continues to bury its own directory (Yahoo Directory Makes Changes & Further Directory Decline for more).

The entire site status service at Resource Zone was established because of -- ironically given why the ODP was born -- the inability for some people to get listed with the ODP. If anything, that problem has gotten worse, given that Resource Zone feels its being overwhelmed by submission requests. Either some radical change is needed, or perhaps the ODP's mission needs to change.

Want to comment or discuss? We have a current thread related to this topic, Editorial: The Lords of Ye 'Olde ODP.

And looking for some Open Directory alternatives? Here's a past article from Gary with some.

Posted by Danny Sullivan on May 23, 2005, 11:36 AM | Permalink


BitTorrent Plans to Launch Search Engine & Carry Ask Jeeves Ads

While several earch tools exist to find BitTorrent material (movies, music, software and other files), the service itself has not offered its own search tool. That's about to change as BitTorrent launches its own advertising-supported search engine in the next two weeks.

Sponsored links will come from Ask Jeeves (apparently the Premier Listings that Ask sells directly, rather than the paid listing that it carries from Google).

BitTorrent speeds internet file transfers by shifting the bandwidth burden off the publisher, and distributing it among users downloading the file: Everyone downloading a file over BitTorrent is unobtrusively uploading it to other users at the same time so that large, popular files actually move at a faster rate than obscure ones.

The new search engine takes that dynamic into account. It resembles Google in operation, with a simple interface and results ranked by an automated process. But unlike a general web search, the BitTorrent web crawler interacts with each torrent behind the scenes to determine the number of nodes downloading and uploading through it. That lets the search engine order its results by the throughput of each torrent.

More in the Wired News article: Next for BitTorrent: Search. It also discusses the many copyright issues that BitTorrent might face as the service becomes more well-known and material becomes even easier to access.

"I think the search engine itself shouldn't be illegal, but I think [founder Bram](Cohen) will find himself inundated with notices of infringing material," says [Stanford Law School Professor Mark] Lemley. "He may find over time that his full-time job is turning off links." Moreover, being right might not be enough to keep Cohen and BitTorrent clear of the working end of a lawsuit. "I would be very surprised if he didn't get sued, because they've gone after a number of people who have much less connection to infringement," says Lemley.


Postscript: Prospect of Search Ads on P2P Site Rattle SEMs from ClickZ has a few more details from Ask and comments from search marketers on source exclusion.

Posted by Gary Price on May 23, 2005, 11:34 AM | Permalink


Search Forums Roundup: May 23, 2005

Today's SearchDay, Search Engine Forums Spotlight, features our weekly links to this week's hot topics from search engine forums across the web: Google Gets Personalized Home Page Feature - Proposed Search Engine Standard For Titles & Descriptions - Traditional Marketing As Part Of Your SEM Campaign - Steve Ballmer: 'One-hit wonder' Google could disappear in five years - Search Engine Optimization and Gresham's Law, and more.

Posted by Chris Sherman on May 23, 2005, 10:43 AM | Permalink

See More Posts From:

This Week | This Month